I am doing a binary classification using Random Forest. My Training dataset has 5 attributes and Feature importance for the data set is as below: A - 28%, B, 27%, C - 17%, D - 17%, E - 11%
My Testing dataset set has 5 rows and values for columns A,B,D and E are constant for each row. The only value I am changing is for Column C. I am increasing the value of Column C by 5% with each row. Although my model is predicting all 5 rows in the correct binary class the probabilities for the prediction seem to be all over the place. I thought the values would constantly increase from 87%
rf.predict_proba(X_test) array([[ 0.12502395, 0.87497605], [ 0.25752719, 0.74247281], [ 0.36002373, 0.63997627], [ 0.20251886, 0.79748114], [ 0.37752784, 0.62247216], [ 0.4625238 , 0.5374762 ], [ 0.36502739, 0.63497261]])
Am I doing something wrong?? Thanks in advance!