Optimized algorithm for Credit Scoring

Abstract

The rapid expansion of credit scoring technologies is increased today. Credit scoring will be considered as the significant element in the financial industries. It plays an important role in modern affairs such as credit customer selection, risk measurement, post-loan and after-loan supervision, comprehensive performance evaluation etc. Credit scoring has been recognized as a binary classification technique distinguishing applicants into two classes: good credit and bad credit, based on characteristics such as gender, age, occupation, and salary. These determine the applicability of loans for applicants. There are two main stream classification techniques, statistical techniques and machine learning techniques. Linear discriminant analysis and logistic regression are the two most commonly used statistical techniques in credit scoring. Machine learning techniques include K-nearest neighbor, support vector machine, decision tree and neural network. Use different best algorithms for classify the credit scoring data sets. Here uses four algorithms for the classification of credit scoring data sets and then the accuracy of different algorithms on the data sets will be obtained.

Development of a Hybrid Credit Scoring Model for the Banking System

“…Some of the various classification models we can mention include the k-nearest neighbors ( [3]; [4]; [5]; [6]; [7]), support vector machine ( [8]; [9]; [10]; [11]), logit regression ( [12]; [13]; [14]; [15]; [16]; [17]), various neural networks such as feed-forward neural networks, back propagation and multilayer perceptron ( [18]; [19]; [20]; [21]; [22]; [23]; [24]; [25]; [26]), decision tree ( [27]; [28]; [29]), Boosting ( [30]; [31]; [32]; [33]; [34]) and the dynamic ensemble classification -soft probability technique [35]. Additionally, optimization-based methods and metaheuristic algorithms are presented separately ( [36]; [37]; [38]; [39]; [40]). The metaheuristic algorithms such as genetic algorithms (GA), and particle swarm optimization (PSO), are the algorithms which, inspired by nature, physics, and humans, are capable of performing optimization operations in the search space with high accuracy.…”

FAWPA: A FAW Attack Protection Algorithm Based on the Behavior of Blockchain Miners

“…Therefore, the clustering effect of existing algorithms is not ideal, and the running time of existing algorithms is too long. They cannot meet the precision rate and real-time requirements of credit rating classification [23]. The Aquila Optimizer (AO) algorithm [24] has a good search ability for global optimal solutions.…”