Random logistic machine (RLM): Transforming statistical models into machine learning approach

Yu Shan Li*, Chao Yu Guo

*此作品的通信作者

研究成果: Article同行評審

摘要

Data science is booming with big data, and machine learning provides better predictive analyses. However, conventional statistical models can effortlessly interpret the effect estimates, and the prediction models are generally in a closed form. Therefore, this research integrates the logistic regression model’s core with the Random Forest structure to create a blended novel machine learning method, the Random Logistic Machine (RLM). In this way, the new approach preserves the statistical and machine learning advantages. Computer simulation studies examined the predictive ability of RLM, random forest, and Logistic Regression under various scenarios. The results showed that the RLM delivers a comparable performance to Random Forests and Logistic Regression. An application to the Breast Cancer Wisconsin (Diagnostic) Data Set also demonstrates the superior performance of the new approach.

指紋

深入研究「Random logistic machine (RLM): Transforming statistical models into machine learning approach」主題。共同形成了獨特的指紋。

引用此