Abstract
The Credit risk prediction research domain has been evolving with different predictive models and these models have been developed using various tools. But the problem is that many of the tools are used in the wrong situation orwith the wrong data conditions. A systematic review of 62 journals articles published during 2010 to 2020 has been carried out in this paper. This review paper focuses on performance shown by elevenpromising and popular tools based on 13 key criterions used in credit risk prediction. It includes the following machine learning tools: SVM(Support vector machines), MDA(Multiple discriminant analysis),RS(Rough sets), LR(Logistic regression), ANN(Artificial neural network), CBR( case based reasoning), DT(Decision tree), GA(Genetic algorithm), KNN(K-Nearest Neighbor), XGBoost algorithm and DGHNL(Deep Genetic Hierarchical Network of Learners) .Various parameters used so far to identify criterions include result transparency accuracy, fully deterministic output, , data size capability, data dispersion, variable types applicable etc. A framework with the help of tables and diagrams has been proposed for the selection of tools that best fit different situations. It is determined that no single tool is predominantly better than the other tools. Their performance varies as scenario/situation changes. This review paper contributes towards a detailed and complete understanding of various tools developed till date for credit risk prediction and their limitations.

Key words: CRPM, BPM, Hybrid tools, Machine learning tools.