Hybrid Logistic Super Newton Model for Predicting Small Sample Size Data

Abstract

Logistic regression is a model commonly used for predicting data with large sample sizes. However, in real-world scenarios, many cases involve small datasets that need to be addressed using logistic regression. The aim of this research is to develop a hybrid logistic regression model to address issues with small sample sizes by combining the Newton Raphson and Super Cubic methods. This hybrid model is applied to predict student dropout at Universitas Duta Bangsa Surakarta. The performance of the hybrid model is evaluated using two main metrics: the convergence of the parameter approximation to measure the precision of parameter estimation, and the ROC curve to assess prediction accuracy. Experimental results show that the Hybrid Logistic Super Newton model outperforms the logistic regression Newton Raphson model, requiring only three iterations to converge, thus improving computational efficiency. Moreover, this model achieves higher accuracy, with an AUC of 0.8833. These findings suggest that the developed model has the potential to be applied in various fields, such as healthcare, finance, and others, offering an effective solution for accurate, real-time predictive analytics. Further research could focus on optimizing the model’s computational efficiency and exploring its application in other domains with small dataset challenges, such as healthcare and finance.