add new example regarding svm hyperparameter plotting #834
Conversation
| output_format='dataframe', | ||
| parameters_in_separate_columns=True, | ||
| ) | ||
| hyperparameters = ['sklearn.svm.classes.SVC(16)_C', 'sklearn.svm.classes.SVC(16)_gamma'] |
There was a problem hiding this comment.
How about printing the column names here to show that sklearn.svm.classes.SVC(16)_C & sklearn.svm.classes.SVC(16)_gamma or a check whether it is in the colnames list with the in operator?
| df = openml.evaluations.list_evaluations_setups( | ||
| 'predictive_accuracy', flow=[8353], task=[6], | ||
| output_format='dataframe', parameters_in_separate_columns=True, | ||
| ) # Choose an SVM flow, for example 8353, and a task. |
There was a problem hiding this comment.
Could we have the parameters on separate lines to maintain parity with a similar call earlier in the script?
| plt.colorbar(cntr, label="accuracy") | ||
| plt.xlim((min(C), max(C))); plt.ylim((min(gamma), max(gamma))) | ||
| plt.xlabel("C (log10)"); plt.ylabel("gamma (log10)") |
There was a problem hiding this comment.
It may be good to have a brief title to the plot.
Also, shouldn't there be a plt.show() for it to come up on the rendered website, and also as a thumbnail on the Examples page?
There was a problem hiding this comment.
I'll add a title.
The other two are not necessary for the gallery.
| import openml | ||
| import numpy as np | ||
| import matplotlib.pyplot as plt |
There was a problem hiding this comment.
We could maybe avoid duplicate imports by having a one-time dependency loading at the beginning.
Codecov Report
@@ Coverage Diff @@
## develop #834 +/- ##
==========================================
Coverage ? 89.31%
==========================================
Files ? 36
Lines ? 4718
Branches ? 0
==========================================
Hits ? 4214
Misses ? 504
Partials ? 0Continue to review full report at Codecov.
|
No description provided.