University ranking prediction system by analyzing influential global performance indicators

In this research, we present a technique of developing university ranking prediction system by analyzing global university performance indicators. Here, we consider standardized dataset of Times higher education world university rankings. Firstly, we perform country wise university ranking data analysis to observe the variation of performance indicators to find out the top influential features. To build the proposed prediction model, we split the ranking dataset into training and test data. Then, based on score of previous years we generate predicted score for each influential feature using our proposed outlier detection and rank score calculation algorithm. Later on, all the universities are ranked globally based on the predicted total score. Then, we evaluate the prediction system accuracy based on ROC curve, recall, number of matched rank against rank deviation. Finally it is justified that our proposed university ranking prediction system is well suited to assess the upcoming global university ranking.