تقدير الانحدار الخطي المتعدد باستعمال الشبكات العصبية

Abstract

This thesis explores how to improve multiple linear regression estimates using neural networks through the R programming language. Linear regression, a widely recognized statistical method for scientific and accurate forecasting, has been innovatively analyzed here by applying neural networks to estimate relationships between independent variables and the dependent variable.

The thesis focuses on using neural networks as the sole method for data analysis, replacing traditional methods such as least squares with machine learning techniques that offer greater flexibility and improved accuracy in estimation. Through R programming, neural network models have been developed to interpret complex data and identify key factors influencing the dependent variable.

The results of this thesis demonstrate how the use of neural networks can significantly enhance prediction accuracy in multiple linear regression models, opening new avenues for exploring complex relationships in data. Through this study, we present a pathway for future research development, emphasizing the potential of using advanced quantitative analysis techniques to enhance our understanding of various phenomena.