tumbleweed
tumbleweed

Reputation: 4640

How to plot performance percentages using percentage on y axis and each metric on x axis?

Let's say I have the following metrics which I obtained from an estimator:

aproach 1:

Accuracy: 0.492307692308
score: 0.492307692308
precision: 0.368678121457
recall: 0.492307692308
hamming loss: 0.0536130536131
Jaccard similarity: 0.946386946387
F-Beta Score: 0.902376921174

aproach 2:

Accuracy: 0.07692308
score: 0.307692308
precision: 0.8678121457
recall: 0.492307692308
hamming loss: 0.0536130536131
Jaccard similarity: 0.946386946387
F-Beta Score: 0.902376921174

aproach 3:

Accuracy: 0.432307692308
score: 0.412307692308
precision: 0.68678121457
recall: 0.2307692308
hamming loss: 0.0536130536131
Jaccard similarity: 0.946386946387
F-Beta Score: 0.902376921174

This metrics where obtained like this:

from sklearn.metrics.metrics import precision_score, recall_score, confusion_matrix, classification_report, accuracy_score, roc_auc_score, auc
print '\nAccuracy:', accuracy_score(y_test, prediction)
print '\nscore:', classifier.score(testing_matrix, y_test)
print '\nprecision:', precision_score(y_test, prediction)
print '\nrecall:', recall_score(y_test, prediction)
print 'Hamming loss:',hamming_loss(y_test,prediction)
print 'Jaccard similarity:',jaccard_similarity_score(y_test,prediction)
print 'F-Beta Score:',fbeta_score(y_test, prediction, average='macro', beta=0.5)

How can I plot this different aproaches performance with matplotlib?. Let's say on the y axis the percentage and on the x the aproach?.

Upvotes: 0

Views: 759

Answers (1)

mbatchkarov
mbatchkarov

Reputation: 16049

@cel'answer is the correct one if you want to know what to plot. If your question is more about how to plot your numbers, seaborn has something called factor plot. Have a look at the tutorial here.

You can easily produce a graph like this (pretend the x axis has labels, and they are accuracy, f1, precision, recall): enter image description here

Upvotes: 1

Related Questions