Bike Buyer Predictor

September 20, 2018

Report Abuse
97 % accuracy prediction of buying a bike based on only 7 variables. Method: two class decision jungle.
A 23 variable data set is reduced after exploration to only 7, thus strongly improving the classification accuracy. The numerical variables are transformed und normalised accordingly. Finally only the scored label is returned. The accuracy on the productive set of 500 entries evaluated to 97%.