Prediction the sale price of the car using XGBoost
This repo contains script for prediction sale price of the car. It includes next steps:
Preprocessing: *renaming column; *processing missing values (full drop of object if any value not exists); *feature reduction; *convert text data in appropriate format; *removing outliers; *splitting our dataset in train and test data;
Running XGBoost: *running XGBoost in 2 loops, which change parameters "gamma" and "max_depth"; *printing score of model.
Best result for model - 71% on max_depth=2 and gamma=0. In comparing with linear regression, which gives us 32% accuracy, this result is much better).
All necessary information about packages consists in file "requirements.txt".