• Teja Reddy Gatla

The main purpose of this paper is to review the use of machine learning in credit risk assessment, focusing on the data pre-processing aspects which are often the most time-consuming activities in the development of machine learning models. The report begins with an overview of machine learning methods, assessing their advantages and disadvantages compared to traditional credit scoring, such as the Statistical Discrimination Method. An investigation of how the choice of performance measure used in credit scoring can influence the model that is selected for classification. Model settings and the issues regarding classification of the response variable are discussed, followed by an example of a test/train data split [1]. The main body of this document presents a demonstration of data pre-processing techniques and an evaluation of their effects on classification accuracy. Effects on model assessments and comparisons are discussed throughout. An example of variable selection method is investigated, demonstrating the accessibility to high-level statistical tools due to inter-platform data compatibility. Finally, we discuss the future use of predictive models that are automatically implemented using model classification strategies[1]. Variable coding is an important step of pre-processing which this report does not cover due to keeping to the constraints of the main objective of developing classification models. Machine learning is a technique of data analysis that makes automated model building by employing different algorithms. It is a subfield of artificial intelligence that involves some systems learning from data, since no humans are necessary for providing help on the most important decisions [1,2]. Over the last few years, banks and other financial organizations have been pushing harder to use machine learning to improve their credit risk models. One of the most attractive things about using machine learning is that it can help make better credit decisions in a shorter amount of time, which can be critical in an industry where being the first to make a decision on a potential customer can be the difference between millions of pounds [2]. Additionally, predictive models can help automatically implement the best strategies to consumers, again saving time and resources for the organization. This has led to great interest in developing and testing new machine learning models to be used as benchmarks for more traditional credit risk models.

Machine learning, Credit, Finance, Loans, Credit risk, Automation, Model building, Classification, Statistics, banks, Deby, Workflow SOAR platforms, Predictive models

"MACHINE LEARNING IN CREDIT RISK ASSESSMENT: ANALYZING HOW MACHINE LEARNING MODELS ARE TRANSFORMING THE ASSESSMENT OF CREDIT RISK FOR LOANS AND CREDIT CARDS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 6, page no.k746-k750, June-2023, Available :http://www.jetir.org/papers/JETIR2306A99.pdf

"MACHINE LEARNING IN CREDIT RISK ASSESSMENT: ANALYZING HOW MACHINE LEARNING MODELS ARE TRANSFORMING THE ASSESSMENT OF CREDIT RISK FOR LOANS AND CREDIT CARDS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 6, page no. ppk746-k750, June-2023, Available at : http://www.jetir.org/papers/JETIR2306A99.pdf