Cambridge Coding Academy: Fixed

3 changes to make it work today

Split into training and test sets

s10 · run_code workaround

Try the old import first, then fall back to the new location.

@@ -1,3 +1,6 @@

-from sklearn.cross_validation import train_test_split

+try:

+ from sklearn.cross_validation import train_test_split

+except ImportError:

+ from sklearn.model_selection import train_test_split

# Split into training and test sets

Sources

Release history — scikit-learn 0.18.2 documentationSome estimators can still accept keyword arguments on the `fit` but this is restricted to data-dependent values (e.g. a Gram matrix or an affinity matrix that are precomputed from the `X` data matrix. The `cross_val` package has been renamed to `cross_validation` although there is also a `cross_val` package alias in place for backward compatibility. Third-party projects with a dependency on scikit-learn 0.9+ should upgrade their codebase. For instance, under Linux / MacOSX just run (make a backup first!): ``` find - name".py" | xargs sed - i's/ \b cross_val \b/cross_validation/g' ``` [...] > T

Set up grid search with cross-validation

s16 · run_code workaround

Try the old import first, then fall back to the new location.

@@ -1,3 +1,6 @@

-from sklearn.grid_search import GridSearchCV

+try:

+ from sklearn.grid_search import GridSearchCV

+except ImportError:

+ from sklearn.model_selection import GridSearchCV

# Search for good hyperparameter values

Sources

Release history — scikit-learn 0.18.2 documentationSome estimators can still accept keyword arguments on the `fit` but this is restricted to data-dependent values (e.g. a Gram matrix or an affinity matrix that are precomputed from the `X` data matrix. The `cross_val` package has been renamed to `cross_validation` although there is also a `cross_val` package alias in place for backward compatibility. Third-party projects with a dependency on scikit-learn 0.9+ should upgrade their codebase. For instance, under Linux / MacOSX just run (make a backup first!): ``` find - name".py" | xargs sed - i's/ \b cross_val \b/cross_validation/g' ``` [...] > T

Visualize grid search results as heatmap

s19 · run_code workaround

Use grid_scores_ if present (old API), otherwise use cv_results_ (new API).; Replace invalid origin='higher' with origin='upper' which is a supported value in matplotlib 3.11+.

@@ -3,5 +3,8 @@

# fetch scores, reshape into a grid

-scores = [x[1] for x in gridCV.grid_scores_]

+if hasattr(gridCV, 'grid_scores_'):

+ scores = [x[1] for x in gridCV.grid_scores_]

+else:

+ scores = gridCV.cv_results_['mean_test_score']

scores = np.array(scores).reshape(len(n_estimators), len(max_features))

scores = np.transpose(scores)

@@ -9,5 +12,5 @@

# Make heatmap from grid search results

plt.figure(figsize=(12, 6))

-plt.imshow(scores, interpolation='nearest', origin='higher', cmap='jet_r')

+plt.imshow(scores, interpolation='nearest', origin='upper', cmap='jet_r')

plt.xticks(np.arange(len(n_estimators)), n_estimators)

plt.yticks(np.arange(len(max_features)), max_features)

Sources

Release history — scikit-learn 0.18.2 documentationSome estimators can still accept keyword arguments on the `fit` but this is restricted to data-dependent values (e.g. a Gram matrix or an affinity matrix that are precomputed from the `X` data matrix. The `cross_val` package has been renamed to `cross_validation` although there is also a `cross_val` package alias in place for backward compatibility. Third-party projects with a dependency on scikit-learn 0.9+ should upgrade their codebase. For instance, under Linux / MacOSX just run (make a backup first!): ``` find - name".py" | xargs sed - i's/ \b cross_val \b/cross_validation/g' ``` [...] > T
Matplotlib.pdf(points or pixels) the origin is the bottom-left of the figure or Axes. Points are typographic points meaning that they are a physical unit measuring 1/72 of an inch. Points and pixels are discussed in further detail in Plotting in physical coordinates. [...] For origin='lower': • [0, 0] is at (left, bottom) • [M', 0] is at (left, top) • [0, N'] is at (right, bottom) • [M', N'] is at (right, top) origin='upper' reverses the vertical axes direction and filling: 284 Chapter 3. Using Matplotlib Matplotlib, Release 3.11.1 • [0, 0] is at (left, top) • [M', 0] is at (left, bottom) • [0, N'] is at (r
matplotlib.pyplot.imshow — Matplotlib 3.11.1 documentationmatplotlib.pyplot.imshow(X, cmap=None, norm=None, \, aspect=None, interpolation=None, alpha=None, vmin=None, vmax=None, colorizer=None, origin=None, extent=None, interpolation\_stage=None, filternorm=True, filterrad=4.0, resample=None, url=None, data=None, \\kwargs)( : Display data as an image, i.e., on a 2D regular raster.
Positioning and orientation of imshow images — Matplotlib 3.11.2 documentationThe default value of origin is set by `rcParams["image.origin"]rcParams["image.origin"]` (default: `'upper'`) which defaults to `'upper'` to match the matrix indexing conventions in math and computer graphics image indexing conventions.

Cambridge Coding Academy

3 changes to make it work today

Split into training and test sets

Sources

Set up grid search with cross-validation

Sources

Visualize grid search results as heatmap

Sources

Tell the author

Badge