Skip to content
/ Big-data-Credit-card-fraud-detection Public

In this project, three prospective approaches are demonstrated for pre-processing large data sets in practical time-frames, that can attempt to address the class imbalance by improving the running time of the relevant SMOTE+ENN oversampling techniques, with the aim of improving or enabling classifier performance. The focus of our study was to im…

1 star 1 fork Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

SMJajoo/Big-data-Credit-card-fraud-detection

BranchesTags

Repository files navigation

BIg-data-Credit-card-fraud-detection

Dataset:

Link for the dataset: https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud

The data-set used here consists of 284,807 European credit card transaction instances of which 492 (0.172%) are fraudu- lent (denoted by the class label ’1’). 2 of the 30 predictors are the output of PCA transformation applied in part to anonymize the data set. The remaining 2 are the ’Time’ and ’Amount’ features.

Three aproaches are used to solve this problem.

  1. Local - Local Map Reduce

image

Local SMOTE+ENN

  1. Global - Smote_Enn_Global

image

GlobalSMOTEEN

  1. Hybrid - Smote_Enn_Cluster_Global

image

HybridSMOTEENN

About

In this project, three prospective approaches are demonstrated for pre-processing large data sets in practical time-frames, that can attempt to address the class imbalance by improving the running time of the relevant SMOTE+ENN oversampling techniques, with the aim of improving or enabling classifier performance. The focus of our study was to im…

Topics

cluster bigdata mapreduce smote smoteenn

Resources

Readme
Activity

Stars

1 star

Watchers

1 watching

Forks

1 fork
Report repository

Releases

No releases published

Packages

No packages published

Languages