The company has products which can be used for hiring assessments. The task is to predict the probability percentage that a client will purchase a product from the features provided in the dataset.
The dataset contains the following files :
1. train.csv : 7007 x 23
2. test.csv : 2093 x 22
3. sample_submission.csv : 5 x 2
The main challenge in this problem was to clean the dataset. The dataset had multiple missing values in various columns. Also. some of the features were categorical in nature. So it had to be converted into integral values. Once the dataset was cleaned, training it and making prediction was quite easy.
I have used XGBoost algorithm to train the data. It performed excellent with an accuracy of 95.38027
The link for the competition is as follows : https://www.hackerearth.com/challenges/competitive/hackerearth-machine-learning-challenge-reduce-marketing-spend/problems/