Page 540 - Python Data Science Handbook
P. 540
as table, 343 Seaborn, 314
classification, 351 visualization of geographic distributions,
clustering, 353 498-501
dimensionality, 352 kernel SVM, 411-414
pair plots, 317 kernel transformation, 413
scatter plots, 236 kernel trick, 413
visualization of, 345 keyboard shortcuts, IPython shell, 8
isnull() method, 124 command history, 9
Isomap navigation, 8
dimensionality reduction, 341, 355 text entry, 9
face data, 456-460 Knuth, Donald, 25
ix attribute (Pandas), 110
L
J labels/labeling
jet colormap, 257 classification task, 333-335
joins, 145 clustering, 338-339
(see also merging) dimensionality reduction and, 340-342
categories of, 147-149 regression task, 335-338
datasets, 146-158 simple line plots, 230-232
many-to-one, 148 Lambert conformal conic projection, 303
one-to-one, 147 lasso regularization (L1 regularization), 399
set arithmetic for, 152 learning curves, computing, 372
joint distributions, 316, 320 left join, 153
Jupyter notebook left_index keyword, 151-152
launching, 2 legends, plot
plotting from, 220 choosing elements for, 251
customizing, 249-255
K multiple legends on same axes, 254
k-means clustering, 339, 462-476 point size, 252
basics, 463-465 levels, naming, 133
color compression example, 473-476 line plots
expectation-maximization algorithm, axes limits for, 228-230
465-476 labeling, 230-232
GMM as means of addressing weaknesses line colors and styles, 226-228
of, 477-480 Matplotlib, 224-232
simple digits data application, 470-473 line-by-line profiling, 28
kernel (defined), 496 linear regression (in machine learning), 390
kernel bandwidth basis function regression, 392-396
defined, 496 regularization, 396-400
selection via cross-validation, 497 Seattle bicycle traffic prediction example,
kernel density estimation (KDE), 491-506 400
bandwidth selection via cross-validation, simple, 390-392
497 lists, Python, 37-41
Bayesian generative classification with, loc attribute (Pandas), 110
501-506 locally linear embedding (LLE), 453-455
custom estimator, 501-506 logarithms, 55
histograms and, 491-496
in practice, 496-506 M
Matplotlib, 248 machine learning, 331
522 | Index

