A robust hybrid methodology between applied linear regression model (alrm) and multilayer perceptron (mlp)

Background: The goal of this study is to illustrate an optimum variable selection method using established Multiple Linear Regression (MLR) models and validate the Multilayer Perceptron Neural Network (MLP) models. Initially, all selected variables will be passed through bootstrap methodology, they were screened for significant relationships. Objective: work analyze construct a model factor linked with total crime cases by combining Applied Model (ALRM) (MLP). Material Methods: Around 200 data was simulated build methodology. Advanced computational statistical modeling methodologies used evaluate descriptions several in retrospective study, including victim, gender, age, marital status, social class, adult household, children burglary’s sexual’s victim’s report, household location. case developed implemented R-Studio program syntax. Results: demonstrated that regression surpasses R-squared mean square error test most situations. Researchers observed when divided into two datasets training testing, hybrid approach performs significantly better at predicting experiment’s outcome. When it came time determine validity, well-established bootstrap-integrated MLR applied. Ten characteristics are taken consideration case: Gender (: -0.4369700; p< 0.25), age -0.0086757; status 0.2646097; class ( : 0.0602540; -0.0211293; p> -0.0025346; victim 1.3473593; 1.0382444; report -0.3176104; location -0.1355046; 0.25).There 0.07745823 MSE linear scenario. Conclusion: neural network’s Predicted Mean Square Error (PMSE) assess MLP’s performance (MSE-forecasts Network). PMSE how far our projections from actual data, lowest MLP indicates best achievement. R syntax also included research article.As result, study’s conclusion establishes superiority technique. Bangladesh Journal Medical Science Vol. 22 No. 01 January’23 Page 38-46