Comparison of decision trees , random forest and xgbbost in solving regressoin problem based on Kaggle 'House Prices' competition. Link to my leaderboard: https://www.kaggle.com/sylwiamielnicka

The code was prepared for my undergraduate disertation, it's purpose wasn't best result on Kaggle, but comparing performance of decisin trees, random forest and xgboost (feature engineering is just a sample). Tree-based models were compared in three stages:

1. raw data (messy, with NA)
2. data with filled-in NA's (MICE implementation)
3. sample feature engineering + fine-tuning models' parameters

Code is divided into three parts: tree_rf_xgb_models.R - contains main code. funkcje.R - contains custom functions which are used in tree_rf_xgb_models.R.