The University of Basrah is exploring the use of artificial intelligence to estimate regression models:
A master's thesis at the College of Administration and Economics, Department of Statistics, explored the use of artificial intelligence to estimate some nonparametric regression models with applications.

The thesis, presented by student Doumou Mohammed Radhi, aimed to evaluate several machine learning models within the framework of artificial intelligence for estimating and analyzing earthquake and seismic data.

The study included a comparison of four nonparametric models: the modularized regression regression function (MARS) model, kernel regression, nearest neighbor algorithm, and related vector machine regression, using statistical performance evaluation indicators.

The results showed that the MARS model outperformed all indicators, achieving the highest coefficient of determination and the lowest error values.

The study concluded that the MARS model is best suited for handling complex data.