feature_importances deems gender umimportant

Hi all I’m very new to all of this, but I just can’t get my head around this. It is clear that Sex has a big impact on the prediction accuracy, as simply stating that all females will survive results in about 0.76 accuracy. So I’m using sk-learn and have been experimenting with the GradientBoostClassifier. When I train on 80pct of the training set, and use the other 20pct for testing, I usually get a score of around 0.81. classifier = GradientBoostingClassifier(n_estimators=700) classifier.fit(X_train, y_train) print classifier.feature_importances_ What I don’t understand is that feature_importances are showing me something like this: [ 0.04224973 0.03318767 0.42367892 0.05388107 0.03522496 0.07889192 0.01812447 0.11699406 0.0070662 0.01845042 0.01065627 0.11439321 0.03316843 0.01403267] These are the features I used for above run (and no, I didn’t move any features…


Link to Full Article: feature_importances deems gender umimportant