Machine Learning Techniques with Fairness for Prediction of Completion of Drug and Alcohol Rehabilitation

TL;DR

This study applies various machine learning models with fairness measures to predict completion of drug and alcohol rehab, addressing bias in demographic data and evaluating multiple fairness metrics.

Contribution

It introduces a comprehensive approach combining kernel methods, decision trees, and neural networks with fairness mitigation techniques for rehab outcome prediction.

Findings

Kernel methods outperform other models in accuracy.

Fairness measures effectively reduce demographic bias.

Multiple fairness metrics provide a nuanced bias assessment.

Abstract

The aim of this study is to look at predicting whether a person will complete a drug and alcohol rehabilitation program and the number of times a person attends. The study is based on demographic data obtained from Substance Abuse and Mental Health Services Administration (SAMHSA) from both admissions and discharge data from drug and alcohol rehabilitation centers in Oklahoma. Demographic data is highly categorical which led to binary encoding being used and various fairness measures being utilized to mitigate bias of nine demographic variables. Kernel methods such as linear, polynomial, sigmoid, and radial basis functions were compared using support vector machines at various parameter ranges to find the optimal values. These were then compared to methods such as decision trees, random forests, and neural networks. Synthetic Minority Oversampling Technique Nominal (SMOTEN) for categorical data was used to balance the data with imputation for missing data. The nine bias variables were then intersectionalized to mitigate bias and the dual and triple interactions were integrated to use the probabilities to look at worst case ratio fairness mitigation. Disparate Impact, Statistical Parity difference, Conditional Statistical Parity Ratio, Demographic Parity, Demographic Parity Ratio, Equalized Odds, Equalized Odds Ratio, Equal Opportunity, and Equalized Opportunity Ratio were all explored at both the binary and multiclass scenarios.