Exploring Faculty Performance in Engineering Institutions: Integrating National Board of Accreditation Attributes into Performance Score Prediction Models

Abstract

The performance score of academic faculty members often hinges on a blend of factors, encompassing experience, qualifications, and notably, performance evaluations. However, the specific benchmarks guiding salary determinations can diverge significantly across engineering institutions and departments. This study shifts focus towards predicting faculty’s performance score of engineering institutions solely based on performance metrics rather than a blend of factors. We examine data randomly gathered on faculty members' performance across metrics such as teaching effectiveness, research productivity, professional development, service to the institution and community, student mentoring, innovation and entrepreneurship, internationalization, and social impact. These metrics are utilized to formulate a performance score for each faculty member, subsequently utilized in predicting their performance score through a linear regression model.

Data Collection: This research centers on gathering Key Performance Indicators from the National Board of Accreditation in India, primarily aimed at assessing and elevating the quality standards of higher education institutions. The KPI framework encompasses various dimensions essential for faculty evaluation, including teaching effectiveness, research productivity, professional development, service to the institution and community, student mentoring, innovation and entrepreneurship, internationalization efforts, and social impact.

Feature Engineering: A composite score for each faculty member is computed based on their performance across diverse metrics. The composite score is derived using the formula:

Model Evaluation: The model's efficacy is assessed through metrics like Mean Absolute Error and Mean Squared Error, alongside employing cross-validation for more dependable estimates of its performance.

Model Deployment: The trained model is deployed to prognosticate performance score for new faculty members. A web application is to be developed, accepting a faculty member's performance metrics as input and generating a predicted score as output. Overall, this project furnishes a framework for performance-based prediction of faculty’s score, offering institutions a tool for crafting equitable and transparent salary structures for faculty members of Engineering Institutions.