Abstract:Regression modeling is one of the most widely used statistical processes to estimate the relationships between dependent and independent variables, which have been frequently applied in a wide range of applications successfully. This method includes many techniques for modeling and analyzing several variables to cover real-world problems. The performance basis in conventional regression modeling is based on the assumption that maximum accuracy in inaccessible data is obtained from models with the least amount of error in modeling available data. In this type of regression modeling, in order to maximize the generalization ability of simulations, which are the main factor influencing the quality of decisions made in real-world problems, the principle of maximization of the accuracy of available historical data is used. However, in this type of modeling process, the model's reliability and results have not been considered. On the other, the generalization capability of a model is simultaneously dependent on the accuracy of the model and the reliability level of the accuracy. In this paper, a new methodology is proposed for multiple linear regression (MLR) modeling in which in contrast to traditionally developed models, the models' reliability is maximized instead of its accuracy. To comprehensively evaluate the proposed model's performance, 30 benchmark data sets are considered from the UCI. Empirical results indicate that, from a general perspective, in 19 cases, i.e., 63.333% of cases, the proposed model has better generalization ability than traditional ones. It is clearly illustrated the importance of the reliability of results and their accuracy that is considered in none of the conventional MLR modeling procedures. Therefore, the proposed MLR model can be regarded as an appropriate alternative in modeling fields, especially when more generalization is desired.

Towards an MCDM-Based Evaluation Framework for Regression Algorithms

Evaluation of Classification Algorithms Using MCDM and Rank Correlation.

User Preferences Based Software Defect Detection Algorithms Selection Using MCDM.

Evaluating Software Reliability: Integration of MCDM and Data Mining

Recommendation of Regression Models for Real Estate Price Prediction using Multi-Criteria Decision Making

Mcdm Approach to Evaluating Bank Loan Default Models

Famcdm: A Fusion Approach of Mcdm Methods to Rank Multiclass Classification Algorithms

Evaluation of Feature Selection Methods for Text Classification with Small Datasets Using Multiple Criteria Decision-Making Methods

Multiple Attribute Decision Making Based on Cross-Evaluation with Uncertain Decision Parameters

Evaluation of clustering algorithms for financial risk analysis using MCDM methods

Algorithms for Multi-Criteria Decision-Making and Efficiency Analysis Problems

A MCDM-Based Evaluation Approach for Imbalanced Classification Methods in Financial Risk Prediction

Performance Evaluation and Comparison of a New Regression Algorithm

An Integrated machine learning and DEA-predefined performance outcome prediction framework with high-dimensional imbalanced data

Multicriteria Decision Making Approach for Cluster Validation

Evaluating biometrics by using a hybrid MCDM model

Construction of an AI-Driven Risk Management Framework for Financial Service Firms Using the MRDM Approach

An empirical study of classification algorithm evaluation for financial risk prediction

Regression Diagnostics meets Forecast Evaluation: Conditional Calibration, Reliability Diagrams, and Coefficient of Determination

Multi-project Regression Based Approach for Software Defect Number Prediction

Etemadi multiple linear regression