Taba Binary, Multinomial, and Ordinal Regression Models: New Machine Learning Methods for Classification.

Area of Science:

• Machine Learning
• Statistical Modeling

Background:

• Machine learning classification methods are broadly applied across disciplines.
• Taba regression is a novel classification technique designed for binary, multinomial, and ordinal data analysis.

Purpose of the Study:

• To introduce and evaluate the performance of the Taba regression classification method.
• To compare Taba regression against established classification models using real-world data.

Main Methods:

• Analysis of liver cirrhosis data from a Mayo Clinic study.
• Performance comparison of Taba regression with artificial neural network (ANN), random forest (RF), logistic regression (LR), and probit analysis (PA).
• Evaluation metrics included true positive rate, F-score, accuracy, and area under the receiver operating characteristic curve (AUC).

Main Results:

• Taba regression demonstrated competitive performance against ANN, RF, LR, and PA.
• The model achieved favorable results in terms of accuracy, recall, F-score, and AUC.
• The Taba regression model proved effective for analyzing the liver cirrhosis dataset.

Conclusions:

• Taba regression is a viable and reliable alternative classification method in machine learning.
• Researchers and practitioners can utilize Taba regression for analyzing various outcome types.
• The study confirms Taba regression's efficacy and reliability for classification tasks.