BHUVANESWARI J , THAMILVANAN G , GURU P , CHRISTY A N , LAKSHMI BALA M , SWATHI S , KEERTHANA K
DOI: https://doi.org/The work life balance (WLB) is one of the most important determinants of professional performance and personal wellbeing of women teachers in Tamil Nadu. As the needs of teaching, administration, and family requirements continue to rise, the forecasting of the levels of WLB may offer useful information to the policy-makers and the institutions, to develop an effective support system. The proposed research hypothesizes that it is possible to predict and model WLB among the women teachers with the help of real-time data based on the structured questionnaire that will be administered with the use of Google Forms. The numerical responses in the form of a dataset were pre-treated, applying the methods of working with missing values, normalizing features, and equal distribution of classes. They used two machine learning algorithms, namely, the k-Nearest Neighbors (K-NN) and the Logistic Regression (LR) to categorize teachers into three types of WLB: poor, moderate and good. Model analysis had been conducted through k-fold cross-validation and the measures of accuracy, precision, recall, F1-score and ROC-AUC. Comparative findings revealed the strengths of K-NN in the local data structure capture, whereas Logistic Regression could offer interpretable information on important predictors of WLB. The paper emphasizes that quantitative predictors of WLB outcome include workload, teaching hours, family commitments, and institutional support, which have a considerable effect on the outcomes. The findings do not only add to the usage of machine learning in educational and social studies, but also offer evidence-based suggestions to better the work-life balance of women educators in Tamil Nadu.
