Author(s) :

Shivani Abrol, Monika Gupta.

Volume/Issue :

Abstract :

Using data analytics and machine learning to combat fraud is a strategy that many businesses have already considered. Fraud may be detected, investigated, and prevented with the aid of big data analytics and machine learning. The purpose of this research is to systematically review the 219 Scopus-indexed publications in context of data analytics in detecting financial crime during 1999 to 2022. The findings indicate that a significant portion of the literature focuses on the utilization of big data analytics, specifically machine learning and deep learning techniques, for the purpose of detecting credit fraud or financial statement fraud. Previous studies have primarily concentrated on the utilization of hybrid technology in the realm of financial fraud detection, thereby indicating its potential as a promising avenue for future research. This study highlights the prominent research gap existing for a predictive model that can issue a warning as soon as a vulnerability for fraudulent behavior is noted. Moreover, findings highlight the accentuated need for data-driven financial investment model and stock market anomalies in context of data analytics and text mining, along with key future research agenda.

No. of Downloads :

14

KEYWORDS:

Financial Fraud Detection, Big Data, Big Data Analytics, Machine Learning, Deep Learning, Bibliometric Analysis

INTRODUCTION & OBJECTIVES:

Unstructured data is the norm in the modern day, and it typically features high levels of volume, velocity, variety, and variability. The rapid pace at which new data is being generated makes it challenging to analyze using conventional methods. Now that so much data is readily available, businesses in virtually every sector are concentrating on finding ways to leverage it to their advantage. Therefore, it is vital that we analyze this pile of data in order to find answers to several business issues. “Big data analytics is the use of advanced analytic techniques against very large, diverse data sets that include structured, semi-structured and unstructured data, from different sources, and in different sizes from terabytes to zettabytes” (IBM). Pattern recognition, machine learning, predictive analytics, text analytics, deep learning, data modelling, data mining, statistics, and computational linguistics are some of the advanced analytics techniques that companies can use on their own or in conjunction with their existing enterprise data to uncover previously undiscovered insights. The Association of Certified Fraud Examiners (ACFE) report says that “the average company loses more than $1.5 million to fraud” (Association of Certified Fraud Examiners, 2022). Beyond the obvious monetary losses, fraud can also negatively impact customer satisfaction, company image, operational problems, and more. We can no longer rely on tried-and-true techniques for detecting fraud. The methods used to manage fraud risk need to develop in parallel with the increasing sophistication of criminal attacks. Data analytics and machine learning aid in the identification and detection of fraud by reducing costs and increasing income, saving time and effort by automating fraud analysis, stopping fraud in its tracks, increasing the percentage of successful fraud detections by minimizing false alarms, and also helping to locate and correct flaws in systems or business processes. Data mining is used to find previously unknown connections and patterns in large datasets. The methods of data analysis focus mostly on the extraction of numerical and statistical features of the data. Using these methods, you can gain a deeper understanding of the processes underlying your data and make more informed decisions as a result. Capturing, storing, analyzing, and correctly visualizing massive amounts of diverse data is made possible by big data analytics. It is helpful to construct a predictive model that can issue a warning as soon as a vulnerability for fraudulent behaviour is identified.

DOI:

10.58426/cgi.v5.i1.2023.82-111