Evaluating the Performance of Machine Learning Algorithm for Classification of Safer Sexual Negotiation among Married Women in Bangladesh

Abstract

Safer sexual practice is essential for improving women’s reproductive and sexual health outcomes. The goal of this study is to identify the contributing factors influencing safer sexual negotiations (SSN) through the application of machine learning algorithms. The algorithms include logistic regression (LR), random forest, Naïve Bayes, linear discriminant analysis, classification and regression trees, support vector machines (SVM), and K-nearest neighbors. This study utilized data from the 2017-18 Bangladesh Demographic and Health Survey, encompassing 19,457 married women within the ages of 15–49 years. The analysis reveals that the SVM algorithm achieved the highest classification accuracy (99.66%), along with high sensitivity (99.98%) and the lowest specificity. Conversely, the LR model produced the highest area under the curve statistics (0.6699), indicating good performance in distinguishing SSN among married women. The outcome illustrated that women’s autonomy, engagement with financial institutions, educational attainment, and their partner’s education play a significant role in SSN with their partners. The findings highlight the significance of empowering women, enhancing reproductive health awareness, and improving socio-economic conditions and education to encourage SSN. The government needs to consider all these risk factors to promote greater SSN for preventing sexually transmitted diseases among women in Bangladesh.