This repository contains a complete analysis of a financial dataset and the benchmark of different outliers detection algorithm !
This project is about Machine Learning and Cybersecurity.
The dataset is not in the repository. It is the IEEE CIS Fraud Detection contest 2019 : the dataset entails user information as well as transaction information, linked together with a TransactionID.
The objective of this notebook is twofold: • Deploy data analysis for fraud detection in a highly incomplete dataset • Perform anomaly detection using IsolationForest and Local Outlier Factor (LOF) algorithm and evaluate their complementarity for unsupervised anomaly detection
The analysis is made with python on the Secu&IA - Lab session 1 – anomaly detection.ipynb notebook. We are going through the whole analysis process, with the dataset exploration, the cleaning, the feature engineering, statistical analysis, visualization and finally the machine learning part.
The results are not great, and we can deduce multiple reason. The main one would be that we don't have enough data for this kind of algorithms. The unsupervised algorithm are not stable to perform well between each other. But we still find a correlation between fraud and outliers, so there is something to get from this implementation.