The purpose of this project is to identify data outliers and anomalies, compare data balancing methods and provide evaluation. • Data visualization before and after analysis • Applying different fraud identification methods with different hypotheses • Checking whether the data is a time series or not • Using different balancing methods • Providing a comparison table of different quality assessment criteria • Applying dimension reduction methods and comparing the results of each one separately The data set used in this report is of tabular type and is checked in order to find fraud, which is 374823 samples with 18 columns.