I read http://stackoverflow.com/questions/15869919/random-forest-predict-using-less-estimators and http://stackoverflow.com/questions/14192284/random-forests-probability-estimates-scikit-learn-specific
Does that mean if I want the probability for a single tree I can just access to RandomForestClassifier.estimators_ which represents a single tree, then e.g. call predict_proba() for each tree's probability? Given a test like for idx, tree in enumerate(model.estimators_): proba = tree.predict_proba(test_setx) print "idx %d:\n%s" % (idx, proba) It looks what I am searching for, but can't be very sure because I notice the probability output for each tree looks like either 0 or 1 idx 0: [1.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0, 1.0, 1.0, 1.0, 0.0, ... idx 1: [1.0, 1.0, 1.0, 1.0 ... 0.0, 0.0, 0.0, 0.0, 0.0] idx 2: [1.0, 1.0, 1.0, 1.0 ... 0.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0] idx 3: [1.0, 1.0, 1.0, 1.0, 1.0 ... 1.0, 1.0, 1.0, 0.0] ... Original output for predict_proba from a single tree is as below idx 0: [[ 1. 0.] [ 1. 0.] [ 0. 1.] ..., [ 1. 0.] [ 1. 0.] [ 0. 1.]] idx 1: [[ 1. 0.] [ 1. 0.] [ 1. 0.] ..., [ 0. 1.] [ 0. 1.] [ 0. 1.]] If this is the probability for each tree, how to calculate the probability for the ensemble? I appreciate any advice. Thanks ------------------------------------------------------------------------------ Introducing Performance Central, a new site from SourceForge and AppDynamics. Performance Central is your source for news, insights, analysis and resources for efficient Application Performance Management. Visit us today! http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
