Data Analysis on ERP System of Easy-Pay Payment Services

Abstract

Machine Learning is one of the leading technologies in this generation with unique techniques and problem-solving algorithms which are very helpful to predict the models accurately with higher positivity rates as compare to the less or false positives prediction of any given projects. In this chapter, we can assure that all of the proposed approaches are almost valid and able to detect fraud transactions. Through this chapter, our aim is to show and determine the possibility of balancing the ratio between fraudulent and non-fraudulent transactions. The methods and parameters which have been used to study the project has anticipated after a thorough research and every requirement have been fulfilled to obtain the desired output. Payment fraud detection is the solid example of classification. The datasets have been modelled after analyzing and pre-processing the existing datasets later to differentiate between fraudulent and non-fraudulent transactions. Such predictions can only be determined with the help of Machine Learning as we used along with the relevant algorithms and techniques which was required for the respective chapter. Numerous interpretations have been done, but acquiring 100% accuracy is still cannot be accomplished due to class imbalanced data related to the transactions. To monitor the usual transactions of people is a vast task and so is to track the fare and fraudulent transactions among billion of people. The main goal of this research is to prevent users from performing unusual transactions through unverified sources and to obtain the positive results in terms of data accuracy and data visualization.