Predict-Ad-Clicks

A Machine Learning challenge to predict the probability whether an ad will get clicked or not.

This repo contains solution of this challenge.

Relevant datasets can be found at the contest site.

File Description

• xg_boost_final.ipynb --> Final model which uses XGBoost to predict the probability of ad-clicks.

• other_algos.ipynb --> Other models like Naive bayes, Decision trees and Random Forest.

• submission_xg.csv.zip --> Output file after running the model on test.csv.

Results

• xg_boost_final gives AUC-ROC score of 0.806684 and overall Accuracy of 0.98011 using validation set as 25% of training set.