Previous Chapter

5. Machine Learning

About the Author

Index

Symbols

%automagic, Shell-Related Magic Commands
%cpaste, Pasting Code Blocks: %paste and %cpaste
%debug, Debugging: When Reading Tracebacks Is Not Enough
%history, Related Magic Commands
%lprun, Line-by-Line Profiling with %lprun
%lsmagic, Help on Magic Functions: ?, %magic, and %lsmagic
%magic, Help on Magic Functions: ?, %magic, and %lsmagic
%matplotlib, Plotting from an IPython shell
%memit, Profiling Memory Use: %memit and %mprun
%mode, Controlling Exceptions: %xmode-Controlling Exceptions: %xmode
%mprun, Profiling Memory Use: %memit and %mprun
%paste, Pasting Code Blocks: %paste and %cpaste
%prun, Profiling Full Scripts: %prun
%run, Running External Code: %run
%time, Timing Code Snippets: %timeit and %time-Timing Code Snippets: %timeit and %time
%timeit, Timing Code Execution: %timeit, Timing Code Snippets: %timeit and %time-Timing Code Snippets: %timeit and %time
& (ampersand), Boolean Arrays as Masks
* (asterisk), Beyond tab completion: Wildcard matching
: (colon), Array Slicing: Accessing Subarrays
? (question mark), Accessing Documentation with ?
?? (double question mark), Accessing Source Code with ??
_ (underscore) shortcut, Underscore Shortcuts and Previous Outputs
| (operator), Boolean Arrays as Masks

A

absolute value function, Absolute value
aggregate() method, Aggregation
aggregates
- computed directly from object, Aggregates
- multidimensional, Multidimensional aggregates
- summarizing set of values with, Example: What Is the Average Height of US Presidents?
aggregation (NumPy), Aggregations: Min, Max, and Everything in Between-Example: What Is the Average Height of US Presidents?
- minimum and maximum, Minimum and Maximum
- multidimensional aggregates, Multidimensional aggregates
- presidents average height example, Example: What Is the Average Height of US Presidents?
- summing the values in an array, Summing the Values in an Array
- various functions, Other aggregation functions
aggregation (Pandas), Aggregation and Grouping-Grouping example
- groupby() operation, GroupBy: Split, Apply, Combine-Grouping example
- MultiIndex, Data Aggregations on Multi-Indices
- Planets dataset for, Planets Data
- simple aggregation, Simple Aggregation in Pandas-Simple Aggregation in Pandas
Akaike information criterion (AIC), How many components?, Example: GMM for Generating New Data
Albers equal-area projection, Conic projections
algorithmic efficiency
- big-O notation, Example: k-Nearest Neighbors
- dataset size and, Example: Binning Data
ampersand (&), Boolean Arrays as Masks
Anaconda, Installation Considerations
and keyword, Boolean Arrays as Masks
annotation of plots, Text and Annotation-Arrows and Annotation
- arrows, Arrows and Annotation-Arrows and Annotation
- holidays/US births example, Example: Effect of Holidays on US Births
- transforms and text position, Transforms and Text Position-Transforms and Text Position
APIs (see Estimator API)
append() method, Pandas vs. Python, The append() method
apply() method, The apply() method
arithmetic operators, Array arithmetic
arrays
- accessing single rows/columns, Accessing array rows and columns
- arithmetic operators, Array arithmetic
- attributes, NumPy Array Attributes
- basics, The Basics of NumPy Arrays
- Boolean, Working with Boolean Arrays-Boolean operators
- broadcasting, Computation on Arrays: Broadcasting-Plotting a two-dimensional function
- centering, Centering an array
- computation on, Computation on NumPy Arrays: Universal Functions-Ufuncs: Learning More
- concatenation, Concatenation of arrays, Recall: Concatenation of NumPy Arrays
- creating copies, Creating copies of arrays
- creating from Python lists, Creating Arrays from Python Lists
- creating from scratch, Creating Arrays from Scratch
- data as, Introduction to NumPy
- DataFrame object as, DataFrame as a generalized NumPy array
- DataFrame object constructed from, From a two-dimensional NumPy array
- fixed-type, Fixed-Type Arrays in Python
- Index object as immutable array, Index as immutable array
- Index object vs., Index as ordered set
- indexing: accessing single elements, Array Indexing: Accessing Single Elements
- reshaping, Reshaping of Arrays
- Series object vs., Series as generalized NumPy array
- slicing, Array Slicing: Accessing Subarrays
- slicing multidimensional subarrays, Multidimensional subarrays
- slicing one-dimensional subarrays, Array Slicing: Accessing Subarrays
- sorting, Sorting Arrays-RecordArrays: Structured Arrays with a Twist
- specifying output to, Specifying output
- splitting, Splitting of arrays
- standard data types, NumPy Standard Data Types
- structured, Structured Data: NumPy’s Structured Arrays-RecordArrays: Structured Arrays with a Twist
- subarrays as no-copy views, Subarrays as no-copy views
- summing values in, Summing the Values in an Array
- universal functions, Computation on NumPy Arrays: Universal Functions-Ufuncs: Learning More
arrows, Arrows and Annotation-Arrows and Annotation
asfreq() method, Resampling and converting frequencies-Resampling and converting frequencies
asterisk (*), Beyond tab completion: Wildcard matching
automagic function, Shell-Related Magic Commands
axes limits, Adjusting the Plot: Axes Limits-Adjusting the Plot: Axes Limits

B

bagging, Ensembles of Estimators: Random Forests
bandwidth (see kernel bandwidth)
bar (|) operator, Boolean Arrays as Masks
bar plots, Bar plots
Basemap toolkit
- geographic data with, Geographic Data with Basemap
  - (see also geographic data)
- installation, Geographic Data with Basemap
basis function regression, Derived Features, Basis Function Regression-Gaussian basis functions
- Gaussian basis functions, Gaussian basis functions-Gaussian basis functions
- polynomial basis functions, Polynomial basis functions
Bayesian classification, Bayesian Classification, Example: Not-So-Naive Bayes-Using our custom estimator
- (see also naive Bayes classification)
Bayesian information criterion (BIC), How many components?
Bayesian Methods for Hackers stylesheet, Bayesian Methods for Hackers style
Bayess theorem, Bayesian Classification
bias–variance trade-off
- kernel bandwidth and, Selecting the bandwidth via cross-validation
- model selection and, The bias–variance trade-off-The bias–variance trade-off
bicycle traffic prediction
- linear regression, Example: Predicting Bicycle Traffic
- time series, Example: Visualizing Seattle Bicycle Counts-High-Performance Pandas: eval() and query()
big-O notation, Example: k-Nearest Neighbors
binary ufuncs, Exploring NumPy’s UFuncs
binnings, plt.hexbin: Hexagonal binnings
bitwise logic operators, Boolean operators
bogosort, Sorting Arrays
Bokeh, Other Python Graphics Libraries
Boolean arrays
- Boolean operators and, Boolean operators
- counting entries in, Counting entries
- working with, Working with Boolean Arrays-Boolean operators
Boolean masks, Comparisons, Masks, and Boolean Logic-Boolean Arrays as Masks
- Boolean arrays as, Boolean Arrays as Masks-Boolean Arrays as Masks
- rainfall statistics, Example: Counting Rainy Days
- working with Boolean arrays, Working with Boolean Arrays-Boolean operators
Boolean operators, Boolean operators
broadcasting, Computation on Arrays: Broadcasting-Plotting a two-dimensional function
- adding two-dimensional array to one-dimensional array, Broadcasting example 1
- basics, Introducing Broadcasting-Introducing Broadcasting
- centering an array, Centering an array
- defined, Outer products, Computation on Arrays: Broadcasting
- in practice, Broadcasting in Practice
- plotting two-dimensional function, Plotting a two-dimensional function
- rules, Rules of Broadcasting-Broadcasting example 3
- two compatible arrays, Broadcasting example 2
- two incompatible arrays, Broadcasting example 3

C

categorical data, Categorical Features
class labels (for data point), Classification: Predicting discrete labels
classification task
- defined, Categories of Machine Learning
- machine learning, Classification: Predicting discrete labels-Classification: Predicting discrete labels
clustering, Categories of Machine Learning
- basics, Clustering: Inferring labels on unlabeled data-Clustering: Inferring labels on unlabeled data
- GMMs, Unsupervised learning: Iris clustering, In Depth: Gaussian Mixture Models-Example: GMM for Generating New Data
- k-means, Clustering: Inferring labels on unlabeled data, In Depth: k-Means Clustering-Example 2: k-means for color compression
code
- magic commands for determining execution time, Timing Code Execution: %timeit
- magic commands for pasting blocks, Pasting Code Blocks: %paste and %cpaste
- magic commands for running external, Running External Code: %run
- profiling and timing, Profiling and Timing Code-Profiling Memory Use: %memit and %mprun
- timing of snippets, Timing Code Snippets: %timeit and %time-Timing Code Snippets: %timeit and %time
coefficient of determination, The bias–variance trade-off
colon (:), Array Slicing: Accessing Subarrays
color compression, Example 2: k-means for color compression-Example 2: k-means for color compression
colorbars
- colormap selection, Customizing Colorbars-Choosing the colormap
- customizing, Customizing Colorbars-Example: Handwritten Digits
- discrete, Discrete colorbars
- handwritten digit example, Example: Handwritten Digits-Example: Handwritten Digits
colormap, Customizing Colorbars-Choosing the colormap
column(s)
- accessing single, Accessing array rows and columns
- indexing, Column indexing
- MultiIndex for, MultiIndex for columns
- sorting arrays along, Sorting along rows or columns
- suffixes keyword and overlapping names, Overlapping Column Names: The suffixes Keyword
column-wise operations, DataFrame.eval() for Column-Wise Operations-Local variables in DataFrame.eval()
command history shortcuts, Command History Shortcuts
comparison operators, Comparison Operators as ufuncs-Comparison Operators as ufuncs
concatenation
- datasets, Combining Datasets: Concat and Append-Concatenation with joins
- of arrays, Concatenation of arrays, Recall: Concatenation of NumPy Arrays
- with pd.concat(), Simple Concatenation with pd.concat-Concatenation with joins
confusion matrix, Classification on digits
conic projections, Conic projections
contour plots, Density and Contour Plots-Visualizing a Three-Dimensional Function
- density and, Density and Contour Plots-Visualizing a Three-Dimensional Function
- three-dimensional function, Visualizing a Three-Dimensional Function-Visualizing a Three-Dimensional Function
- three-dimensional plot, Three-Dimensional Contour Plots
Conway, Drew, What Is Data Science?
cross-validation, Model validation via cross-validation-Validation curves in Scikit-Learn
cubehelix colormap, Choosing the colormap
cylindrical projections, Cylindrical projections

D

data
- as arrays, Introduction to NumPy
- missing (see missing data)
data representation (Scikit-Learn package), Data Representation in Scikit-Learn-Target array
- data as table, Data as table
- features matrix, Features matrix
- target array, Target array-Target array
data science, defining, What Is Data Science?
data types, Understanding Data Types in Python
- fixed-type arrays, Fixed-Type Arrays in Python
- integers, A Python Integer Is More Than Just an Integer
- lists in, A Python List Is More Than Just a List-NumPy Standard Data Types
- NumPy, NumPy Standard Data Types
DataFrame object (Pandas), The Pandas DataFrame Object-From a NumPy structured array
- as dictionary, DataFrame as a dictionary-DataFrame as a dictionary
- as generalized NumPy array, DataFrame as a generalized NumPy array
- as specialized dictionary, DataFrame as specialized dictionary
- as two-dimensional array, DataFrame as two-dimensional array-DataFrame as two-dimensional array
- constructing, Constructing DataFrame objects
- data selection in, Data Selection in DataFrame
- defined, Data Manipulation with Pandas
- index alignment in, Index alignment in DataFrame
- masking, Additional indexing conventions
- multiply indexed, Multiply indexed DataFrames
- operations between Series object and, Ufuncs: Operations Between DataFrame and Series
- slicing, Additional indexing conventions
DataFrame.eval() method, DataFrame.eval() for Column-Wise Operations-Local variables in DataFrame.eval()
- assignment in, Assignment in DataFrame.eval()
- local variables in, Local variables in DataFrame.eval()
DataFrame.query() method, DataFrame.query() Method
datasets
- appending, The append() method
- combining (Panda), Combining Datasets: Concat and Append-Example: US States Data
- concatenation, Combining Datasets: Concat and Append-Concatenation with joins
- merging/joining, Combining Datasets: Merge and Join-Example: US States Data
datetime module, Native Python dates and times: datetime and dateutil
datetime64 dtype, Typed arrays of times: NumPy’s datetime64
dateutil module, Native Python dates and times: datetime and dateutil
debugging, Debugging: When Reading Tracebacks Is Not Enough-Partial list of debugging commands
decision trees, Motivating Random Forests: Decision Trees-Decision trees and overfitting
- (see also random forests)
- creating, Creating a decision tree-Creating a decision tree
- overfitting, Decision trees and overfitting
deep learning, Caveats and Improvements
density estimator
- GMM, GMM as Density Estimation-How many components?
- histogram as, Motivating KDE: Histograms
- KDE (see kernel density estimation (KDE))
describe() method, Dispatch methods
development, IPython
- profiling and timing code, Profiling and Timing Code-Profiling Memory Use: %memit and %mprun
- profiling full scripts, Profiling Full Scripts: %prun
- timing of code snippets, Timing Code Snippets: %timeit and %time-Timing Code Snippets: %timeit and %time
dictionary(-ies)
- DataFrame as specialization of, DataFrame as specialized dictionary
- DataFrame object constructed from list of, From a list of dicts
- Pandas Series object vs., Series as specialized dictionary
digits, recognition of (see optical character recognition)
dimensionality reduction, Example: Handwritten Digits
- machine learning, Dimensionality reduction: Inferring structure of unlabeled data-Dimensionality reduction: Inferring structure of unlabeled data
- PCA and, In Depth: Principal Component Analysis
discriminative classification, Motivating Support Vector Machines-Motivating Support Vector Machines
documentation, accessing
- IPython, Help and Documentation in IPython-Beyond tab completion: Wildcard matching, Installing and Using Pandas
- Pandas, Installing and Using Pandas
double question mark (??), Accessing Source Code with ??
dropna() method, Dropping null values
dynamic typing, Understanding Data Types in Python

E

eigenfaces, Example: Eigenfaces-Example: Eigenfaces
ensemble estimator/method, In-Depth: Decision Trees and Random Forests
- (see also random forests)
ensemble learner, Motivating Random Forests: Decision Trees
equidistant cylindrical projection, Cylindrical projections
errors, visualizing
- basic errorbars, Basic Errorbars
- continuous quantities, Continuous Errors
- Matplotlib, Visualizing Errors-Continuous Errors
Estimator API, Scikit-Learn’s Estimator API-Summary
- basics, Basics of the API
- Iris classification example, Supervised learning example: Iris classification
- Iris clustering example, Unsupervised learning: Iris clustering
- Iris dimensionality example, Unsupervised learning example: Iris dimensionality
- simple linear regression example, Supervised learning example: Simple linear regression-Unsupervised learning: Iris clustering
eval() function, pandas.eval() for Efficient Operations-Other operations
- DataFrame.eval() method and, DataFrame.eval() for Column-Wise Operations-Local variables in DataFrame.eval()
- pd.eval() function and, pandas.eval() for Efficient Operations-Other operations
- when to use, Performance: When to Use These Functions
exceptions, controlling, Controlling Exceptions: %xmode-Controlling Exceptions: %xmode
expectation-maximization (E-M) algorithm
- caveats, Caveats of expectation–maximization-Caveats of expectation–maximization
- GMM as generalization of, Generalizing E–M: Gaussian Mixture Models-Choosing the covariance type
- k-means clustering and, k-Means Algorithm: Expectation–Maximization-Example 2: k-means for color compression
exponentials, Exponents and logarithms
external code, magic commands for running, Running External Code: %run

F

face recognition
faceted histograms, Faceted histograms
factor plots, Factor plots
fancy indexing, Fancy Indexing-Example: Binning Data
- basics, Exploring Fancy Indexing
- binning data, Example: Binning Data
- combined with other indexing schemes, Combined Indexing
- modifying values with, Modifying Values with Fancy Indexing
- selection of random points, Example: Selecting Random Points
feature engineering, Feature Engineering-Feature Pipelines
- categorical features, Categorical Features
- derived features, Derived Features-Derived Features
- image features, Image Features
- imputation of missing data, Imputation of Missing Data
- processing pipeline, Feature Pipelines
- text features, Text Features
feature, data point, Classification: Predicting discrete labels
features matrix, Features matrix
fillna() method, Filling null values
filter() method, Filtering
FiveThirtyEight stylesheet, FiveThirtyEight style
fixed-type arrays, Fixed-Type Arrays in Python

G

Gaussian basis functions, Gaussian basis functions-Gaussian basis functions
Gaussian mixture models (GMMs), In Depth: Gaussian Mixture Models-Example: GMM for Generating New Data
- choosing covariance type, Choosing the covariance type
- clustering with, Unsupervised learning: Iris clustering
- density estimation algorithm, GMM as Density Estimation-How many components?
- E–M generalization, Generalizing E–M: Gaussian Mixture Models-Choosing the covariance type
- handwritten data generation example, Example: GMM for Generating New Data-Example: GMM for Generating New Data
- k-means weaknesses addressed by, Motivating GMM: Weaknesses of k-Means-Motivating GMM: Weaknesses of k-Means
- KDE and, In-Depth: Kernel Density Estimation
Gaussian naive Bayes classification, Supervised learning example: Iris classification, Classification on digits, Gaussian Naive Bayes-Gaussian Naive Bayes, HOG in Action: A Simple Face Detector
Gaussian process regression (GPR), Continuous Errors
generative models, Bayesian Classification
geographic data, Geographic Data with Basemap
- Basemap toolkit for, Geographic Data with Basemap
- California city population example, Example: California Cities
- drawing a map background, Drawing a Map Background-Drawing a Map Background
- map projections, Map Projections-Other projections
- plotting data on maps, Plotting Data on Maps
- surface temperature data example, Example: Surface Temperature Data
get() operation, Vectorized item access and slicing
get_dummies() method, Indicator variables
ggplot stylesheet, ggplot
graphics libraries, Other Python Graphics Libraries
GroupBy aggregation, Pivot Tables
GroupBy object, The GroupBy object-Dispatch methods
- aggregate() method, Aggregation
- apply() method, The apply() method
- column indexing, Column indexing
- dispatch methods, Dispatch methods
- filter() method, Filtering
- iteration over groups, Iteration over groups
- transform() method, Transformation
groupby() operation (Pandas), GroupBy: Split, Apply, Combine-Grouping example
- GroupBy object and, The GroupBy object-Dispatch methods
- grouping example, Grouping example
- pivot tables vs., Pivot Tables by Hand
- split key specification, Specifying the split key
- split-apply-combine example, Split, apply, combine-Split, apply, combine

H

handwritten digits, recognition of (see optical character recognition)
hard negative mining, Caveats and Improvements
help
- IPython, Help and Documentation in IPython-Beyond tab completion: Wildcard matching
- magic functions, Help on Magic Functions: ?, %magic, and %lsmagic
help() function, Accessing Documentation with ?
hexagonal binnings, plt.hexbin: Hexagonal binnings
hierarchical indexing
- in one-dimensional Series, A Multiply Indexed Series-Data Aggregations on Multi-Indices
- MultiIndex, Hierarchical Indexing-Data Aggregations on Multi-Indices, The better way: Pandas MultiIndex-MultiIndex as extra dimension
  - (see also MultiIndex type)
- rearranging multi-indices, Rearranging Multi-Indices-Index setting and resetting
- unstack() method, MultiIndex as extra dimension
- with Python tuples as keys, The bad way
Histogram of Oriented Gradients (HOG)
- caveats and improvements, Caveats and Improvements-Caveats and Improvements
- features, HOG Features
- for face detection pipeline, Application: A Face Detection Pipeline-Caveats and Improvements
- simple face detector, HOG in Action: A Simple Face Detector-HOG in Action: A Simple Face Detector
histograms, Histograms, Binnings, and Density-Kernel density estimation
- binning data to create, Example: Binning Data
- faceted, Faceted histograms
- KDE and, Kernel density estimation, Motivating KDE: Histograms-Motivating KDE: Histograms
- manual customization, Plot Customization by Hand-Plot Customization by Hand
- plt.hexbin() function, plt.hexbin: Hexagonal binnings
- plt.hist2d() function, plt.hist2d: Two-dimensional histogram
- Seaborn, Histograms, KDE, and densities-Histograms, KDE, and densities
- simple, Histograms, Binnings, and Density-Histograms, Binnings, and Density
- two-dimensional, Two-Dimensional Histograms and Binnings-Kernel density estimation
holdout sets, Model validation the right way: Holdout sets
Hunter, John, Visualization with Matplotlib
hyperparameters, Supervised learning example: Simple linear regression
- (see also model validation)

I

iloc attribute (Pandas), Indexers: loc, iloc, and ix
images, encoding for machine learning analysis, Image Features
immutable array, Index object as, Index as immutable array
importing, tab completion for, Tab completion when importing
In objects, IPython, IPython’s In and Out Objects
index alignment
- in DataFrame, Index alignment in DataFrame
- in Series, Index alignment in Series
Index object (Pandas), The Pandas Index Object-Index as ordered set
- as immutable array, Index as immutable array
- as ordered set, Index as ordered set
indexing
- fancy, Fancy Indexing-Example: Binning Data
  - (see also fancy indexing)
- hierarchical (see hierarchical indexing)
- NumPy arrays: accessing single elements, Array Indexing: Accessing Single Elements
- Pandas, Data Indexing and Selection
IndexSlice object, Multiply indexed DataFrames
indicator variables, Indicator variables
inner join, Specifying Set Arithmetic for Joins
input/output history, IPython, Input and Output History-Related Magic Commands
- In and Out objects, IPython’s In and Out Objects
- related magic commands, Related Magic Commands
- suppressing output, Suppressing Output
- underscore shortcuts and previous outputs, Underscore Shortcuts and Previous Outputs
installation, Python, Installation Considerations
integers, Python, A Python Integer Is More Than Just an Integer
IPython, IPython: Beyond Normal Python
- accessing documentation with ?, Accessing Documentation with ?
- accessing source code with ??, Accessing Source Code with ??
- command-line commands in shell, Shell Commands in IPython
- controlling exceptions, Controlling Exceptions: %xmode-Controlling Exceptions: %xmode
- debugging, Debugging: When Reading Tracebacks Is Not Enough-Partial list of debugging commands
- documentation, Help and Documentation in IPython-Beyond tab completion: Wildcard matching, Introduction to NumPy
- errors handling, Errors and Debugging-Partial list of debugging commands
- exploring modules with tab completion, Exploring Modules with Tab Completion-Tab completion when importing
- help and documentation, Help and Documentation in IPython-Beyond tab completion: Wildcard matching
- input/output history, Input and Output History-Related Magic Commands
- keyboard shortcuts in shell, Keyboard Shortcuts in the IPython Shell
- launching Jupyter notebook, Launching the Jupyter Notebook
- launching shell, Launching the IPython Shell
- magic commands, IPython Magic Commands-Help on Magic Functions: ?, %magic, and %lsmagic
- notebook (see Jupyter notebook)
- plotting from shell, Plotting from an IPython shell
- profiling and timing code, Profiling and Timing Code-Profiling Memory Use: %memit and %mprun
- shell commands, IPython and Shell Commands-Passing Values to and from the Shell
- shell-related magic commands, Shell-Related Magic Commands
- web resources, Web Resources
- wildcard matching, Beyond tab completion: Wildcard matching
Iris dataset
- as table, Data as table
- classification, Supervised learning example: Iris classification
- clustering, Unsupervised learning: Iris clustering
- dimensionality, Unsupervised learning example: Iris dimensionality
- pair plots, Pair plots
- scatter plots, Scatter Plots with plt.scatter
- visualization of, Target array
isnull() method, Detecting null values
Isomap
- dimensionality reduction, Dimensionality reduction: Inferring structure of unlabeled data, Unsupervised learning: Dimensionality reduction
- face data, Example: Isomap on Faces-Example: Isomap on Faces
ix attribute (Pandas), Indexers: loc, iloc, and ix

J

jet colormap, Choosing the colormap
joins, Concatenation with joins
- (see also merging)
- categories of, Categories of Joins-Many-to-many joins
- datasets, Combining Datasets: Merge and Join-Example: US States Data
- many-to-one, Many-to-one joins
- one-to-one, One-to-one joins
- set arithmetic for, Specifying Set Arithmetic for Joins
joint distributions, Histograms, KDE, and densities, Joint distributions
Jupyter notebook
- launching, Launching the Jupyter Notebook
- plotting from, Plotting from an IPython notebook

K

k-means clustering, Clustering: Inferring labels on unlabeled data, In Depth: k-Means Clustering-Example 2: k-means for color compression
- basics, Introducing k-Means-Introducing k-Means
- color compression example, Example 2: k-means for color compression-Example 2: k-means for color compression
- expectation-maximization algorithm, k-Means Algorithm: Expectation–Maximization-Example 2: k-means for color compression
- GMM as means of addressing weaknesses of, Motivating GMM: Weaknesses of k-Means-Motivating GMM: Weaknesses of k-Means
- simple digits data application, Example 1: k-Means on digits-Example 1: k-Means on digits
kernel (defined), Kernel Density Estimation in Practice
kernel bandwidth
- defined, Kernel Density Estimation in Practice
- selection via cross-validation, Selecting the bandwidth via cross-validation
kernel density estimation (KDE), In-Depth: Kernel Density Estimation-Using our custom estimator
- bandwidth selection via cross-validation, Selecting the bandwidth via cross-validation
- Bayesian generative classification with, Example: Not-So-Naive Bayes-Using our custom estimator
- custom estimator, Example: Not-So-Naive Bayes-Using our custom estimator
- histograms and, Motivating KDE: Histograms-Motivating KDE: Histograms
- in practice, Kernel Density Estimation in Practice-Using our custom estimator
- Matplotlib, Kernel density estimation
- Seaborn, Histograms, KDE, and densities
- visualization of geographic distributions, Example: KDE on a Sphere-Example: KDE on a Sphere
kernel SVM, Beyond linear boundaries: Kernel SVM-Beyond linear boundaries: Kernel SVM
kernel transformation, Beyond linear boundaries: Kernel SVM
kernel trick, Beyond linear boundaries: Kernel SVM
keyboard shortcuts, IPython shell, Keyboard Shortcuts in the IPython Shell
- command history, Command History Shortcuts
- navigation, Navigation Shortcuts
- text entry, Text Entry Shortcuts
Knuth, Donald, Profiling and Timing Code

L

labels/labeling
- classification task, Classification: Predicting discrete labels-Classification: Predicting discrete labels
- clustering, Clustering: Inferring labels on unlabeled data-Clustering: Inferring labels on unlabeled data
- dimensionality reduction and, Dimensionality reduction: Inferring structure of unlabeled data-Dimensionality reduction: Inferring structure of unlabeled data
- regression task, Regression: Predicting continuous labels-Regression: Predicting continuous labels
- simple line plots, Labeling Plots-Labeling Plots
Lambert conformal conic projection, Conic projections
lasso regularization (L1 regularization), Lasso regularization ()
learning curves, computing, Learning curves in Scikit-Learn
left join, Specifying Set Arithmetic for Joins
left_index keyword, The left_index and right_index keywords-The left_index and right_index keywords
legends, plot
- choosing elements for, Choosing Elements for the Legend
- customizing, Customizing Plot Legends-Multiple Legends
- multiple legends on same axes, Multiple Legends
- point size, Legend for Size of Points
levels, naming, MultiIndex level names
line plots
- axes limits for, Adjusting the Plot: Axes Limits-Adjusting the Plot: Axes Limits
- labeling, Labeling Plots-Labeling Plots
- line colors and styles, Adjusting the Plot: Line Colors and Styles-Adjusting the Plot: Line Colors and Styles
- Matplotlib, Simple Line Plots-Labeling Plots
line-by-line profiling, Line-by-Line Profiling with %lprun
linear regression (in machine learning), In Depth: Linear Regression
- basis function regression, Basis Function Regression-Gaussian basis functions
- regularization, Regularization-Lasso regularization ()
- Seattle bicycle traffic prediction example, Example: Predicting Bicycle Traffic
- simple, Simple Linear Regression-Simple Linear Regression
lists, Python, A Python List Is More Than Just a List-NumPy Standard Data Types
loc attribute (Pandas), Indexers: loc, iloc, and ix
locally linear embedding (LLE), Nonlinear Manifolds: Locally Linear Embedding-Nonlinear Manifolds: Locally Linear Embedding
logarithms, Exponents and logarithms

M

machine learning, Machine Learning
- basics, Machine Learning-Summary
- categories of, Categories of Machine Learning
- classification task, Classification: Predicting discrete labels-Classification: Predicting discrete labels
- clustering, Clustering: Inferring labels on unlabeled data-Clustering: Inferring labels on unlabeled data
- decision trees and random forests, In-Depth: Decision Trees and Random Forests
- defined, What Is Machine Learning?
- dimensionality reduction, Dimensionality reduction: Inferring structure of unlabeled data-Dimensionality reduction: Inferring structure of unlabeled data
- educational resources, Further Machine Learning Resources
- face detection pipeline, Application: A Face Detection Pipeline-Caveats and Improvements
- feature engineering, Feature Engineering-Feature Pipelines
- GMM (see Gaussian mixture models)
- hyperparameters and model validation, Hyperparameters and Model Validation-Summary
- KDE (see kernel density estimation)
- linear regression (see linear regression)
- manifold learning (see manifold learning)
- naive Bayes classification, In Depth: Naive Bayes Classification-When to Use Naive Bayes
- PCA (see principal component analysis)
- qualitative examples, Qualitative Examples of Machine Learning Applications-Dimensionality reduction: Inferring structure of unlabeled data
- regression task, Regression: Predicting continuous labels-Regression: Predicting continuous labels
- Scikit-Learn basics, Introducing Scikit-Learn
- supervised, Categories of Machine Learning
- SVMs (see support vector machines)
- unsupervised, Categories of Machine Learning
magic commands
- code block pasting, Pasting Code Blocks: %paste and %cpaste
- code execution timing, Timing Code Execution: %timeit
- help commands, Help on Magic Functions: ?, %magic, and %lsmagic
- IPython input/output history, Related Magic Commands
- running external code, Running External Code: %run
- shell-related, Shell-Related Magic Commands
manifold learning, In-Depth: Manifold Learning-Example: Visualizing Structure in Digits
- "HELLO" function, Manifold Learning: “HELLO”
- advantages/disadvantages, Some Thoughts on Manifold Methods
- applying Isomap on faces data, Example: Isomap on Faces-Example: Isomap on Faces
- defined, In-Depth: Manifold Learning
- k-means clustering (see k-means clustering)
- multidimensional scaling, MDS as Manifold Learning-MDS as Manifold Learning
- PCA vs., Some Thoughts on Manifold Methods
- visualizing structure in digits, Example: Visualizing Structure in Digits-Example: Visualizing Structure in Digits
many-to-one joins, Many-to-one joins
map projections, Map Projections-Other projections
- conic, Conic projections
- cylindrical, Cylindrical projections
- perspective, Perspective projections
- pseudo-cylindrical, Pseudo-cylindrical projections
maps, geographic (see geographic data)
margins, maximizing, Support Vector Machines: Maximizing the Margin-Tuning the SVM: Softening margins
masking, Additional indexing conventions
- (see also Boolean masks)
- Boolean arrays, Boolean Arrays as Masks-Boolean Arrays as Masks
- Boolean masks, Comparisons, Masks, and Boolean Logic-Boolean Arrays as Masks
MATLAB-style interface, MATLAB-style interface
Matplotlib, Visualization with Matplotlib, Matplotlib Resources
- axes limits for line plots, Adjusting the Plot: Axes Limits-Adjusting the Plot: Axes Limits
- changing defaults via rcParams, Changing the Defaults: rcParams
- colorbar customization, Customizing Colorbars-Example: Handwritten Digits
- configurations and stylesheets, Customizing Matplotlib: Configurations and Stylesheets-Seaborn style
- density and contour plots, Density and Contour Plots-Visualizing a Three-Dimensional Function
- error visualization, Visualizing Errors-Continuous Errors
- general tips, General Matplotlib Tips-Saving Figures to File
- geographic data with Basemap toolkit, Geographic Data with Basemap
- gotchas, Labeling Plots
- histograms, binnings, and density, Histograms, Binnings, and Density-Kernel density estimation
- importing, Importing matplotlib
- interfaces, Two Interfaces for the Price of One
- labeling simple line plots, Labeling Plots-Labeling Plots
- line colors and styles, Adjusting the Plot: Line Colors and Styles-Adjusting the Plot: Line Colors and Styles
- MATLAB-style interfaces, MATLAB-style interface
- multiple subplots, Multiple Subplots-plt.GridSpec: More Complicated Arrangements
- object hierarchy of plots, Customizing Ticks
- object-oriented interfaces, Object-oriented interface
- plot customization, Plot Customization by Hand-Plot Customization by Hand
- plot display contexts, show() or No show()? How to Display Your Plots-Plotting from an IPython notebook
- plot legend customization, Customizing Plot Legends-Multiple Legends
- plotting from a script, Plotting from a script
- plotting from IPython notebook, Plotting from an IPython notebook
- plotting from IPython shell, Plotting from an IPython shell
- resources and documentation for, Matplotlib Resources
- saving figures to file, Saving Figures to File
- Seaborn vs., Visualization with Seaborn-Seaborn Versus Matplotlib
- setting styles, Setting Styles
- simple line plots, Simple Line Plots-Labeling Plots
- stylesheets, Stylesheets-Seaborn style
- text and annotation, Text and Annotation-Arrows and Annotation
- three-dimensional function visualization, Visualizing a Three-Dimensional Function-Visualizing a Three-Dimensional Function
- three-dimensional plotting, Three-Dimensional Plotting in Matplotlib-Example: Visualizing a Möbius strip
- tick customization, Customizing Ticks-Summary of Formatters and Locators
max() function, Minimum and Maximum
maximum margin estimator, Support Vector Machines: Maximizing the Margin
- (see also support vector machines (SVMs))
memory use, profiling, Profiling Memory Use: %memit and %mprun
merge key
- on keyword, The on keyword
- specification of, Specification of the Merge Key-The left_index and right_index keywords
merging, Combining Datasets: Merge and Join-Example: US States Data
- (see also joins)
- key specification, Specification of the Merge Key-The left_index and right_index keywords
- relational algebra and, Relational Algebra
- US state population data example, Example: US States Data-Example: US States Data
min() function, Minimum and Maximum
Miniconda, Installation Considerations
missing data, Missing Data in Pandas-NaN and None in Pandas
- feature engineering and, Imputation of Missing Data
- handling, Handling Missing Data-Trade-Offs in Missing Data Conventions
- NaN and None, NaN and None in Pandas
- operating on null values in Pandas, Operating on Null Values-Filling null values
Möbius strip, Example: Visualizing a Möbius strip-Example: Visualizing a Möbius strip
model (defined), Classification: Predicting discrete labels
model parameters (defined), Classification: Predicting discrete labels
model selection
- bias–variance trade-off, The bias–variance trade-off-The bias–variance trade-off
- validation curves in Scikit-Learn, Validation curves in Scikit-Learn-Validation curves in Scikit-Learn
model validation, Hyperparameters and Model Validation-Summary
- bias–variance trade-off, The bias–variance trade-off-The bias–variance trade-off
- cross-validation, Model validation via cross-validation-Validation curves in Scikit-Learn
- grid search example, Validation in Practice: Grid Search
- holdout sets, Model validation the right way: Holdout sets
- learning curves, Learning Curves-Learning curves in Scikit-Learn
- naive approach to, Model validation the wrong way
- validation curves, Validation curves in Scikit-Learn-Validation curves in Scikit-Learn
modules, IPython, Exploring Modules with Tab Completion-Tab completion when importing
Mollweide projection, Pseudo-cylindrical projections
multi-indexing (see hierarchical indexing)
multidimensional scaling (MDS), MDS as Manifold Learning-MDS as Manifold Learning
- basics, Multidimensional Scaling (MDS)-Multidimensional Scaling (MDS)
- locally linear embedding and, Nonlinear Manifolds: Locally Linear Embedding-Nonlinear Manifolds: Locally Linear Embedding
- nonlinear embeddings, Nonlinear Embeddings: Where MDS Fails
MultiIndex type, The better way: Pandas MultiIndex-MultiIndex as extra dimension
- creation methods, Methods of MultiIndex Creation-MultiIndex for columns
- data aggregations on, Data Aggregations on Multi-Indices
- explicit constructors for, Explicit MultiIndex constructors
- extra dimension of data with, MultiIndex as extra dimension
- for columns, MultiIndex for columns
- index setting/resetting, Index setting and resetting
- indexing and slicing, Indexing and Slicing a MultiIndex-Multiply indexed DataFrames
- keys option, Adding MultiIndex keys
- level names, MultiIndex level names
- multiply indexed DataFrames, Multiply indexed DataFrames
- multiply indexed Series, Multiply indexed Series
- rearranging, Rearranging Multi-Indices-Index setting and resetting
- sorted/unsorted indices with, Sorted and unsorted indices
- stacking/unstacking indices, Stacking and unstacking indices
multinomial naive Bayes classification, Multinomial Naive Bayes-Example: Classifying text

N

naive Bayes classification, In Depth: Naive Bayes Classification-When to Use Naive Bayes
- advantages/disadvantages, When to Use Naive Bayes
- Bayesian classification and, Bayesian Classification
- Gaussian, Gaussian Naive Bayes-Gaussian Naive Bayes
- multinomial, Multinomial Naive Bayes-Example: Classifying text
- text classification example, Example: Classifying text-Example: Classifying text
NaN value, From a list of dicts, Index alignment in Series, NaN: Missing numerical data
navigation shortcuts, Navigation Shortcuts
neural networks, Caveats and Improvements
noise filter, PCA as, PCA as Noise Filtering-PCA as Noise Filtering
None object, None: Pythonic missing data, NaN and None in Pandas
nonlinear embeddings, MDS and, Nonlinear Embeddings: Where MDS Fails
notnull() method, Detecting null values
np.argsort() function, Fast Sorting in NumPy: np.sort and np.argsort
np.concatenate() function, Concatenation of arrays, Duplicate indices
np.sort() function, Fast Sorting in NumPy: np.sort and np.argsort
null values, Operating on Null Values-Filling null values
- detecting, Detecting null values
- dropping, Dropping null values
- filling, Filling null values
NumPy, Introduction to NumPy
- aggregations, Aggregations: Min, Max, and Everything in Between-Example: What Is the Average Height of US Presidents?
- array attributes, NumPy Array Attributes
- array basics, The Basics of NumPy Arrays
- array indexing: accessing single elements, Array Indexing: Accessing Single Elements
- array slicing: accessing subarrays, Array Slicing: Accessing Subarrays
- Boolean masks, Comparisons, Masks, and Boolean Logic-Boolean Arrays as Masks
- broadcasting, Computation on Arrays: Broadcasting-Plotting a two-dimensional function
- comparison operators as ufuncs, Comparison Operators as ufuncs-Comparison Operators as ufuncs
- computation on arrays, Computation on NumPy Arrays: Universal Functions-Ufuncs: Learning More
- data types in Python, Understanding Data Types in Python
- datetime64 dtype, Typed arrays of times: NumPy’s datetime64
- documentation, Introduction to NumPy
- fancy indexing, Fancy Indexing-Example: Binning Data
- keywords and/or vs. operators &/|, Boolean Arrays as Masks
- sorting arrays, Sorting Arrays-Example: k-Nearest Neighbors
- standard data types, NumPy Standard Data Types
- structured arrays, Structured Data: NumPy’s Structured Arrays-RecordArrays: Structured Arrays with a Twist
- universal functions, Computation on NumPy Arrays: Universal Functions-Ufuncs: Learning More

O

object-oriented interface, Object-oriented interface
offsets, time series, Frequencies and Offsets
on keyword, The on keyword
one-hot encoding, Categorical Features
one-to-one joins, One-to-one joins
optical character recognition
- digit classification, Classification on digits-Classification on digits
- GMMs, Example: GMM for Generating New Data-Example: GMM for Generating New Data
- k-means clustering, Example 1: k-Means on digits-Example 1: k-Means on digits
- loading/visualizing digits data, Loading and visualizing the digits data
- Matplotlib, Example: Handwritten Digits-Example: Handwritten Digits
- PCA as noise filtering, PCA as Noise Filtering-PCA as Noise Filtering
- PCA for visualization, PCA for visualization: Handwritten digits
- random forests for classifying digits, Example: Random Forest for Classifying Digits-Example: Random Forest for Classifying Digits
- Scikit-Learn application, Application: Exploring Handwritten Digits-Classification on digits
- visualizing structure in digits, Example: Visualizing Structure in Digits-Example: Visualizing Structure in Digits
or keyword, Boolean Arrays as Masks
ordered set, Index object as, Index as ordered set
orthographic projection, Perspective projections
Out objects, IPython, IPython’s In and Out Objects
outer join, Specifying Set Arithmetic for Joins
outer products, Outer products
outliers, PCA and, Principal Component Analysis Summary
output, suppressing, Suppressing Output
overfitting, Learning Curves, Decision trees and overfitting

P

pair plots, Pair plots
Pandas, Data Manipulation with Pandas
- aggregation and grouping, Aggregation and Grouping-Grouping example
- and compound expressions, Motivating query() and eval(): Compound Expressions
- appending datasets, The append() method
- built-in documentation, Installing and Using Pandas
- combining datasets, Combining Datasets: Concat and Append-Example: US States Data
- concatenation of datasets, Combining Datasets: Concat and Append-Concatenation with joins
- data indexing and selection, Data Indexing and Selection
- data selection in DataFrame, Data Selection in DataFrame-Further Resources
- data selection in Series, Data Selection in Series-Indexers: loc, iloc, and ix
- DataFrame object, The Pandas DataFrame Object-From a NumPy structured array
- eval() and query(), High-Performance Pandas: eval() and query()-High-Performance Pandas: eval() and query()
- handling missing data, Handling Missing Data-Trade-Offs in Missing Data Conventions
- hierarchical indexing, Hierarchical Indexing-Data Aggregations on Multi-Indices
- Index object, The Pandas Index Object-Index as ordered set
- installation, Installing and Using Pandas
- merging/joining datasets, Combining Datasets: Merge and Join-Example: US States Data
- NaN and None in, NaN and None in Pandas
- null values, Operating on Null Values-Filling null values
- objects, Introducing Pandas Objects-Index as ordered set
- operating on data in, Operating on Data in Pandas-Filling null values
  - (see also universal functions)
- pandas.eval(), pandas.eval() for Efficient Operations-Other operations
- Panel data, Data Aggregations on Multi-Indices
- pivot tables, Pivot Tables-Further data exploration
- Series object, The Pandas Series Object-Constructing Series objects
- time series, Working with Time Series-Performance: When to Use These Functions
- vectorized string operations, Vectorized String Operations-Going further with recipes
pandas.eval() function, pandas.eval() for Efficient Operations-Other operations
Panel data, Data Aggregations on Multi-Indices
partial slicing, Multiply indexed Series
partitioning (partial sorts), Partial Sorts: Partitioning
pasting code blocks, magic commands for, Pasting Code Blocks: %paste and %cpaste
pd.concat() function
- catching repeats as error, Catching the repeats as an error
- concatenation with, Simple Concatenation with pd.concat-Concatenation with joins
- concatenation with joins, Concatenation with joins
- duplicate indices, Duplicate indices
- ignoring the index, Ignoring the index
- MultiIndex keys, Adding MultiIndex keys
pd.date_range() function, Regular sequences: pd.date_range()
pd.eval() function, pandas.eval() for Efficient Operations-Other operations
pd.merge() function, Combining Datasets: Merge and Join-Example: US States Data
- categories of joins, Categories of Joins-Many-to-many joins
- keywords, Specification of the Merge Key-The left_index and right_index keywords
- left_index/right_index keywords, The left_index and right_index keywords-The left_index and right_index keywords
- merge key specification, Specification of the Merge Key-The left_index and right_index keywords
- relational algebra and, Relational Algebra
- specifying set arithmetic for joins, Specifying Set Arithmetic for Joins
pdb (Python debugger), Debugging: When Reading Tracebacks Is Not Enough
Perez, Fernando, IPython: Beyond Normal Python, Visualization with Matplotlib
Period type, Pandas Time Series Data Structures
perspective projections, Perspective projections
pipelines, Validation curves in Scikit-Learn, Feature Pipelines
pivot tables, Pivot Tables-Further data exploration
- groupby() operation vs., Pivot Tables by Hand
- multi-level, Multilevel pivot tables
- syntax, Pivot Table Syntax-Additional pivot table options
- Titanic passengers example, Motivating Pivot Tables
- US birthrate data example, Example: Birthrate Data-Further data exploration
Planets dataset
- aggregation and grouping, Planets Data
- bar plots, Bar plots
plot legends
- choosing elements for, Choosing Elements for the Legend
- customizing, Customizing Plot Legends-Multiple Legends
- multiple legends on same axes, Multiple Legends
- points size, Legend for Size of Points
Plotly, Other Python Graphics Libraries
plotting
- axes limits for simple line plots, Adjusting the Plot: Axes Limits-Adjusting the Plot: Axes Limits
- bar plots, Bar plots
- changing defaults via rcParams, Changing the Defaults: rcParams
- colorbars, Customizing Colorbars-Example: Handwritten Digits
- data on maps, Plotting Data on Maps-Example: Exploring Marathon Finishing Times
- density and contour plots, Density and Contour Plots-Visualizing a Three-Dimensional Function
- display contexts, show() or No show()? How to Display Your Plots-Plotting from an IPython notebook
- factor plots, Factor plots
- from an IPython shell, Plotting from an IPython shell
- from script, Plotting from a script
- histograms, binnings, and density, Histograms, Binnings, and Density-Kernel density estimation
- IPython notebook, Plotting from an IPython notebook
- joint distributions, Joint distributions
- labeling simple line plots, Labeling Plots-Labeling Plots
- line colors and styles, Adjusting the Plot: Line Colors and Styles-Adjusting the Plot: Line Colors and Styles
- manual customization, Plot Customization by Hand-Plot Customization by Hand
- Matplotlib, Visualization with Matplotlib
- multiple subplots, Multiple Subplots-plt.GridSpec: More Complicated Arrangements
- of errors, Visualizing Errors-Continuous Errors
- pair plots, Pair plots
- plot legends, Customizing Plot Legends-Multiple Legends
- Seaborn, Visualization with Seaborn-Seaborn Versus Matplotlib
- simple line plots, Simple Line Plots-Labeling Plots
- simple scatter plots, Simple Scatter Plots-plot Versus scatter: A Note on Efficiency
- stylesheets for, Stylesheets-Seaborn style
- text and annotation for, Text and Annotation-Arrows and Annotation
- three-dimensional, Three-Dimensional Plotting in Matplotlib-Example: Visualizing a Möbius strip
- three-dimensional function, Visualizing a Three-Dimensional Function-Visualizing a Three-Dimensional Function
- ticks, Customizing Ticks-Summary of Formatters and Locators
- two-dimensional function, Plotting a two-dimensional function
- various Python graphics libraries, Other Python Graphics Libraries
plt.axes() function, plt.axes: Subplots by Hand-plt.axes: Subplots by Hand
plt.contour() function, Visualizing a Three-Dimensional Function-Visualizing a Three-Dimensional Function
plt.GridSpec() function, plt.GridSpec: More Complicated Arrangements-plt.GridSpec: More Complicated Arrangements
plt.imshow() function, Visualizing a Three-Dimensional Function-Visualizing a Three-Dimensional Function
plt.legend() command, Customizing Plot Legends-Multiple Legends
plt.plot() function
- color arguments, Adjusting the Plot: Line Colors and Styles
- plt.scatter vs., plot Versus scatter: A Note on Efficiency
- scatter plots with, Scatter Plots with plt.plot-Scatter Plots with plt.plot
plt.scatter() function
- plt.plot vs., plot Versus scatter: A Note on Efficiency
- simple scatter plots with, Scatter Plots with plt.scatter-Scatter Plots with plt.scatter
plt.subplot() function, plt.subplot: Simple Grids of Subplots
plt.subplots() function, plt.subplots: The Whole Grid in One Go
polynomial basis functions, Polynomial basis functions
polynomial regression model, Validation curves in Scikit-Learn
pop() method, DataFrame as a dictionary
population data, US, merge and join operations with, Example: US States Data-Example: US States Data
principal axes, Introducing Principal Component Analysis-Introducing Principal Component Analysis
principal component analysis (PCA), In Depth: Principal Component Analysis-General Machine Learning
- basics, Introducing Principal Component Analysis-PCA as Noise Filtering
- choosing number of components, Choosing the number of components
- eigenfaces example, Example: Eigenfaces-Example: Eigenfaces
- facial recognition example, Example: Eigenfaces-Example: Eigenfaces
- for dimensionality reduction, PCA as dimensionality reduction
- handwritten digit example, PCA for visualization: Handwritten digits-Choosing the number of components, PCA as Noise Filtering-PCA as Noise Filtering
- manifold learning vs., Some Thoughts on Manifold Methods
- meaning of components, What do the components mean?-What do the components mean?
- noise filtering, PCA as Noise Filtering-PCA as Noise Filtering
- strengths/weaknesses, Principal Component Analysis Summary
- visualization with, PCA for visualization: Handwritten digits
profiling
- full scripts, Profiling Full Scripts: %prun
- line-by-line, Line-by-Line Profiling with %lprun
- memory use, Profiling Memory Use: %memit and %mprun
projections (see map projections)
pseudo-cylindrical projections, Pseudo-cylindrical projections
Python
- installation considerations, Installation Considerations
- Python 2.x vs. Python 3, Python 2 Versus Python 3
- reasons for using, Why Python?

Q

query() method
- DataFrame.query() method, DataFrame.query() Method
- when to use, Performance: When to Use These Functions
question mark (?), accessing IPython documentation with, Accessing Documentation with ?
quicksort algorithm, Fast Sorting in NumPy: np.sort and np.argsort

R

radial basis function, Beyond linear boundaries: Kernel SVM
rainfall statistics, Example: Counting Rainy Days
random forests
- advantages/disadvantages, Summary of Random Forests
- classifying digits with, Example: Random Forest for Classifying Digits-Example: Random Forest for Classifying Digits
- defined, Ensembles of Estimators: Random Forests
- ensembles of estimators, Ensembles of Estimators: Random Forests-Ensembles of Estimators: Random Forests
- motivating with decision trees, Motivating Random Forests: Decision Trees-Decision trees and overfitting
- regression, Random Forest Regression
RandomizedPCA, Example: Eigenfaces
rcParams dictionary, changing defaults via, Changing the Defaults: rcParams
RdBu colormap, Choosing the colormap
record arrays, RecordArrays: Structured Arrays with a Twist
reduce() method, Aggregates
regression, Random Forest Regression-Summary of Random Forests
- (see also specific forms, e.g.: linear regression)
regression task
- defined, Categories of Machine Learning
- machine learning, Regression: Predicting continuous labels-Regression: Predicting continuous labels
regular expressions, Methods using regular expressions
regularization, Regularization-Lasso regularization ()
- lasso regularization, Lasso regularization ()
- ridge regression, Ridge regression ( regularization)
relational algebra, Relational Algebra
resample() method, Resampling and converting frequencies-Resampling and converting frequencies
reset_index() method, Index setting and resetting
reshaping, Reshaping of Arrays
ridge regression (L2 regularization), Ridge regression ( regularization)
right join, Specifying Set Arithmetic for Joins
right_index keyword, The left_index and right_index keywords-The left_index and right_index keywords
rolling statistics, Rolling windows
runtime configuration (rc), Changing the Defaults: rcParams

S

scatter plots (see simple scatter plots)
Scikit-Learn package, Machine Learning, Data Representation in Scikit-Learn-Target array
- API (see Estimator API)
- basics, Introducing Scikit-Learn-Summary
- data as table, Data as table
- data representation in, Data Representation in Scikit-Learn-Target array
- Estimator API, Scikit-Learn’s Estimator API-Unsupervised learning: Iris clustering
- features matrix, Features matrix
- handwritten digit application, Application: Exploring Handwritten Digits-Classification on digits
- support vector classifier, Fitting a support vector machine-Fitting a support vector machine
- target array, Target array-Target array
scipy.special submodule, Specialized ufuncs
script
- plotting from, Plotting from a script
- profiling, Profiling Full Scripts: %prun
Seaborn
- bar plots, Bar plots
- datasets and plot types, Exploring Seaborn Plots-Example: Exploring Marathon Finishing Times
- faceted histograms, Faceted histograms
- factor plots, Factor plots
- histograms, KDE, and densities, Histograms, KDE, and densities-Histograms, KDE, and densities
- joint distributions, Joint distributions
- marathon finishing times example, Example: Exploring Marathon Finishing Times-Example: Exploring Marathon Finishing Times
- Matplotlib vs., Visualization with Seaborn-Seaborn Versus Matplotlib
- pair plots, Pair plots
- stylesheet, Seaborn style
- visualization with, Visualization with Seaborn-Seaborn Versus Matplotlib
Seattle, bicycle traffic prediction in
- linear regression, Example: Predicting Bicycle Traffic-Example: Predicting Bicycle Traffic
- time series, Example: Visualizing Seattle Bicycle Counts-High-Performance Pandas: eval() and query()
Seattle, rainfall statistics in, Example: Counting Rainy Days
semi-supervised learning, Categories of Machine Learning
Series object (Pandas), The Pandas Series Object-Constructing Series objects
- as dictionary, Series as specialized dictionary, Series as dictionary
- constructing, Constructing Series objects
- data indexing/selection in, Data Selection in Series-Indexers: loc, iloc, and ix
- DataFrame as dictionary of, DataFrame as a dictionary-DataFrame as a dictionary
- DataFrame object constructed from, From a single Series object
- DataFrame object constructed from dictionary of, From a dictionary of Series objects
- generalized NumPy array, Series as generalized NumPy array
- hierarchical indexing in, A Multiply Indexed Series-Data Aggregations on Multi-Indices
- index alignment in, Index alignment in Series
- indexer attributes, Indexers: loc, iloc, and ix
- multiply indexed, Multiply indexed Series
- one-dimensional array, Series as one-dimensional array
- operations between DataFrame and, Ufuncs: Operations Between DataFrame and Series
shell, IPython
- basics, Quick Introduction to the Shell
- command-line commands, Shell Commands in IPython
- commands, IPython and Shell Commands-Passing Values to and from the Shell
- keyboard shortcuts in, Keyboard Shortcuts in the IPython Shell
- launching, Launching the IPython Shell
- magic commands, Shell-Related Magic Commands
- passing values to and from, Passing Values to and from the Shell
shift() function, Time-shifts-Time-shifts
shortcuts
- accessing previous output, Underscore Shortcuts and Previous Outputs
- command history, Command History Shortcuts
- IPython shell, Keyboard Shortcuts in the IPython Shell-Books
- navigation, Navigation Shortcuts
- text entry, Text Entry Shortcuts
simple histograms, Histograms, Binnings, and Density-Histograms, Binnings, and Density
simple line plots
- axes limits for, Adjusting the Plot: Axes Limits-Adjusting the Plot: Axes Limits
- labeling, Labeling Plots-Labeling Plots
- line colors and styles, Adjusting the Plot: Line Colors and Styles-Adjusting the Plot: Line Colors and Styles
- Matplotlib, Simple Line Plots-Labeling Plots
- simple (Matplotlib), Simple Line Plots-Labeling Plots
simple linear regression, Simple Linear Regression-Simple Linear Regression
simple scatter plots
- California city populations, Customizing Plot Legends-Multiple Legends
- Matplotlib, Simple Scatter Plots-plot Versus scatter: A Note on Efficiency
- plt.plot, Scatter Plots with plt.plot-Scatter Plots with plt.plot
- plt.plot vs. plt.scatter, plot Versus scatter: A Note on Efficiency
- plt.scatter, Scatter Plots with plt.scatter-Scatter Plots with plt.scatter
slice() operation, Vectorized item access and slicing
slicing
- MultiIndex with sorted/unsorted indices, Sorted and unsorted indices
- NumPy arrays, Array Slicing: Accessing Subarrays-Creating copies of arrays
- NumPy arrays: accessing subarrays, Array Slicing: Accessing Subarrays
- NumPy arrays: multidimensional subarrays, Multidimensional subarrays
- NumPy arrays: one-dimensional subarrays, Array Slicing: Accessing Subarrays
- NumPy vs. Python, Subarrays as no-copy views
- Pandas conventions, Additional indexing conventions
sorting arrays, Sorting Arrays-Example: k-Nearest Neighbors
- along rows or columns, Sorting along rows or columns
- basics, Sorting Arrays
- fast sorting with np.sort and np.argsort, Fast Sorting in NumPy: np.sort and np.argsort
- k-nearest neighbors example, Example: k-Nearest Neighbors-Example: k-Nearest Neighbors
- partitioning, Partial Sorts: Partitioning
source code, accessing, Accessing Source Code with ??
splitting arrays, Splitting of arrays
string operations (see vectorized string operations)
structured arrays, Structured Data: NumPy’s Structured Arrays-RecordArrays: Structured Arrays with a Twist
- advanced compound types, More Advanced Compound Types
- creating, Creating Structured Arrays
- record arrays, RecordArrays: Structured Arrays with a Twist
stylesheets
- Bayesian Methods for Hackers, Bayesian Methods for Hackers style
- default style, Default style
- FiveThirtyEight style, FiveThirtyEight style
- ggplot, ggplot
- Matplotlib, Stylesheets-Seaborn style
- Seaborn, Seaborn style
subarrays
- as no-copy views, Subarrays as no-copy views
- creating copies, Creating copies of arrays
- slicing multidimensional, Multidimensional subarrays
- slicing one-dimensional, Array Slicing: Accessing Subarrays
subplots
- manual customization, plt.axes: Subplots by Hand-plt.axes: Subplots by Hand
- multiple, Multiple Subplots-plt.GridSpec: More Complicated Arrangements
- plt.axes() for, plt.axes: Subplots by Hand-plt.axes: Subplots by Hand
- plt.GridSpec() for, plt.GridSpec: More Complicated Arrangements-plt.GridSpec: More Complicated Arrangements
- plt.subplot() for, plt.subplot: Simple Grids of Subplots
- plt.subplots() for, plt.subplots: The Whole Grid in One Go
subsets, faceted histograms, Faceted histograms
suffixes keyword, Overlapping Column Names: The suffixes Keyword
supervised learning, Categories of Machine Learning
- classification task, Classification: Predicting discrete labels-Classification: Predicting discrete labels
- regression task, Regression: Predicting continuous labels-Regression: Predicting continuous labels
support vector (defined), Fitting a support vector machine
support vector classifier, Fitting a support vector machine-Fitting a support vector machine
support vector machines (SVMs), In-Depth: Support Vector Machines
- advantages/disadvantages, Support Vector Machine Summary
- face recognition example, Example: Face Recognition-Example: Face Recognition
- fitting, Fitting a support vector machine-Fitting a support vector machine
- kernels and, Beyond linear boundaries: Kernel SVM-Beyond linear boundaries: Kernel SVM
- maximizing the margin, Support Vector Machines: Maximizing the Margin-Tuning the SVM: Softening margins
- motivating, Motivating Support Vector Machines-Support Vector Machine Summary
- simple face detector, HOG in Action: A Simple Face Detector
- softening margins, Tuning the SVM: Softening margins-Tuning the SVM: Softening margins
surface plots, three-dimensional, Wireframes and Surface Plots-Example: Visualizing a Möbius strip

T

t-distributed stochastic neighbor embedding (t-SNE), Some Thoughts on Manifold Methods, Example 1: k-Means on digits
tab completion
- exploring IPython modules with, Exploring Modules with Tab Completion-Tab completion when importing
- of object contents, Tab completion of object contents
- when importing, Tab completion when importing
table, data as, Data as table
target array, Target array-Target array
term frequency-inverse document frequency (TF-IDF), Text Features
text, Text Features
- (see also annotation of plots)
- transforms and position of, Transforms and Text Position-Transforms and Text Position
text entry shortcuts, Text Entry Shortcuts
three-dimensional plotting
- contour plots, Three-Dimensional Contour Plots
- Möbius strip visualization, Example: Visualizing a Möbius strip-Example: Visualizing a Möbius strip
- points and lines, Three-Dimensional Points and Lines
- surface plots, Wireframes and Surface Plots-Example: Visualizing a Möbius strip
- surface triangulations, Surface Triangulations-Example: Visualizing a Möbius strip
- wireframes, Wireframes and Surface Plots
- with Matplotlib, Three-Dimensional Plotting in Matplotlib-Example: Visualizing a Möbius strip
ticks (tick marks)
- customizing, Customizing Ticks-Summary of Formatters and Locators
- fancy formats, Fancy Tick Formats-Fancy Tick Formats
- formatter/locator options, Summary of Formatters and Locators
- major and minor, Major and Minor Ticks
- reducing/increasing number of, Reducing or Increasing the Number of Ticks
Tikhonov regularization, Ridge regression ( regularization)
time series
- bar plots, Bar plots
- dates and times in Pandas, Dates and times in Pandas: Best of both worlds
- datetime64, Typed arrays of times: NumPy’s datetime64
- frequency codes, Frequencies and Offsets
- indexing data by timestamps, Pandas Time Series: Indexing by Time
- native Python dates and times, Native Python dates and times: datetime and dateutil
- offsets, Frequencies and Offsets
- Pandas, Working with Time Series-High-Performance Pandas: eval() and query()
- Pandas data structures for, Pandas Time Series Data Structures-Regular sequences: pd.date_range()
- pd.date_range(), Regular sequences: pd.date_range()
- Python vs. Pandas, Dates and Times in Python-Dates and times in Pandas: Best of both worlds
- resampling and converting frequencies, Resampling and converting frequencies-Resampling and converting frequencies
- rolling statistics, Rolling windows
- Seattle bicycle counts example, Example: Visualizing Seattle Bicycle Counts-High-Performance Pandas: eval() and query()
- time-shifts, Time-shifts-Time-shifts
- typed arrays, Typed arrays of times: NumPy’s datetime64
Timedelta type, Pandas Time Series Data Structures
Timestamp type, Pandas Time Series Data Structures
timestamps, indexing data by, Pandas Time Series: Indexing by Time
timing, of code, Timing Code Execution: %timeit, Timing Code Snippets: %timeit and %time-Timing Code Snippets: %timeit and %time
transform() method, Transformation
transforms
- modifying, Transforms and Text Position-Transforms and Text Position
- text position and, Transforms and Text Position-Transforms and Text Position
triangulated surface plots, Surface Triangulations-Example: Visualizing a Möbius strip
trigonometric functions, Trigonometric functions
tshift() function, Time-shifts-Time-shifts
two-fold cross-validation, Model validation via cross-validation

U

ufuncs (see universal functions)
unary ufuncs, Exploring NumPy’s UFuncs
underfitting, The bias–variance trade-off, Learning Curves
underscore (_) shortcut, Underscore Shortcuts and Previous Outputs
universal functions (ufuncs), Computation on NumPy Arrays: Universal Functions-Ufuncs: Learning More
- absolute value, Absolute value
- advanced features, Advanced Ufunc Features
- aggregates, Aggregates
- array arithmetic, Array arithmetic
- basics, Introducing UFuncs
- comparison operators as, Comparison Operators as ufuncs-Comparison Operators as ufuncs
- exponentials, Exponents and logarithms
- index alignment, UFuncs: Index Alignment-Index alignment in DataFrame
- index preservation, Ufuncs: Index Preservation
- logarithms, Exponents and logarithms
- operating on data in Pandas, Operating on Data in Pandas-Filling null values
- operations between DataFrame and Series, Ufuncs: Operations Between DataFrame and Series
- outer products, Outer products
- slowness of Python loops, The Slowness of Loops
- specialized ufuncs, Specialized ufuncs
- specifying output, Specifying output
- trigonometric functions, Trigonometric functions
unstack() method, MultiIndex as extra dimension
unsupervised learning
- clustering, Clustering: Inferring labels on unlabeled data-Clustering: Inferring labels on unlabeled data, Unsupervised learning: Iris clustering
- defined, Categories of Machine Learning
- dimensionality reduction, Example: Handwritten Digits, Dimensionality reduction: Inferring structure of unlabeled data-Dimensionality reduction: Inferring structure of unlabeled data, Unsupervised learning example: Iris dimensionality, Unsupervised learning: Dimensionality reduction
- PCA (see principal component analysis)

V

validation (see model validation)
validation curves, Validation curves in Scikit-Learn-Validation curves in Scikit-Learn
variables
- dynamic typing, Understanding Data Types in Python
- passing to and from shell, Passing Values to and from the Shell
variance, in bias–variance trade-off, The bias–variance trade-off-The bias–variance trade-off
vectorized operations, Computation on Arrays: Broadcasting
vectorized string operations, Vectorized String Operations-Going further with recipes
- basics, Introducing Pandas String Operations
- indicator variables, Indicator variables
- methods similar to Python string methods, Methods similar to Python string methods
- methods using regular expressions, Methods using regular expressions
- recipe database example, Example: Recipe Database-Going further with recipes
- tables of, Tables of Pandas String Methods-Indicator variables
- vectorized item access and slicing, Vectorized item access and slicing
Vega/Vega-Lite, Other Python Graphics Libraries
violin plot, Example: Exploring Marathon Finishing Times
viridis colormap, Choosing the colormap
Vispy, Other Python Graphics Libraries
visualization software (see Matplotlib) (see Seaborn)

W

Wickham, Hadley, GroupBy: Split, Apply, Combine
wildcard matching, Beyond tab completion: Wildcard matching
wireframe plot, Wireframes and Surface Plots
word counts, Text Features-Text Features

Previous Chapter

5. Machine Learning

About the Author