Building a cross-border e-commerce talent training platform based on logistic regression model

Abstract

Student success is crucial to the process of building a Cross-border e-commerce (CBEC) talent development platform. Analysis of the important aspects impacting performance and performance prediction are carried out with the goal of enhancing students' academic outcomes. To better forecast student outcomes, a logistic regression model is used for factor analysis, and a penalty function is implemented. Parameters are reconciled using K-fold cross-validation, and then estimated using the coordinate descent technique. Model performance validation findings indicated that the Area Under the curve (AUC) for the minimax concave penalty (MCP) and smoothlyclippedabsolutedeviation(SCAD) penalized logistic regression models were 0.772 and 0.771, respectively. Both the MCP and SCAD penalized logistic regression models have overall accuracies of 0.738 and 0.739, respectively. Researchers found that for MCP, the correlation coefficient was 0.99949, and for SCAD, it was 0.99958, between the projected value and the anticipated value of students' performance. Due to their superior prediction accuracy, the MCP and SCAD penalized logistic regression models may be used as analytical tools in the development of the CBEC talent training platform.