Linear Models, Variable Selection, Artificial Intelligence
Researchers have developed a novel Artificial Intelligence approach for variable selection in linear regression models. This method utilizes an Artificial Neural Network (ANN) trained to assess variable significance based on Ordinary Least Squares (OLS) estimates. A simulation study demonstrated the ANN's accuracy and compared its performance against traditional techniques like Forward/Backward selection, AIC, BIC, and LASSO. The approach was further illustrated using a World Health Organization dataset on Life Expectancy, with code and a pre-trained ANN made available on GitHub. AI
IMPACT Introduces an AI-driven alternative for statistical model selection, potentially improving accuracy and efficiency over traditional methods.