Predicting Institute Graduation Rate using Evolutionary Computing and Machine Learning

Abstract

There are diverse parameters available for measuring performance of an academic institute. Graduation rate of an institute is an important indicator of institute’s success. It is essential to understand which factors lead to better graduation rates. Hence, a prediction system which helps institutes well in advance to avoid poor graduation rate is required. In this study, a novel adaptive dimensionality reduction model is proposed using evolutionary computing and machine learning to better predict institute graduation rate. This work has explored the feature optimization capacity of evolutionary algorithm with weight assignment approach to each dimension. A high dimensional dataset is considered for analyzing attributes that affect institute graduation rates. Proposed model uses adaptive approach of incrementing weights of contributing features which lead to minimum error. Experimental results show that proposed model yields optimum dimensions, low execution time and minimum error. Predictive analysis presented could lead to useful future directions for education domain stakeholders.