Comparison of Tree-based ML methods
The project aims to compare the performance of Tree-based methods for anomaly detection problems i.e., the labels are heavily skewed. For this, we use data from credit card fraud detection taken from Data-Flair's website. They also perform their own detection, whereas we take a more comparative approach. The methods chosen are Decision tree, Random forests and Gradient Boosted trees.
The code is implemented in R as an R-Notebook titled comparison.Rmd, along with the knitted HTML version. For viewing the process, use the HTML file and for reproduction of results, use the notebook. The data can be downloaded directly from Data-Flair's website