TCRKDS: Towards Integration of Semantic Intelligence for Course Recommendation in Support of a Knowledge Driven Strategy

A personalized learning recommendation system has always been an area of focus and a lot of work had been done in the past related to this field, suggesting a proper course needs to deal with various problems such as cold start, data sparsity and not a diverse recommendation. In this paper, we have assimilated data from Kaggle Coursera course dataset, Udemy and open university to curate auxiliary knowledge and thereby formalize a knowledge centric approach along with deep learning classification using LSTM, Entity Population with WebChild, Aristo Tuple and Google Knowledge Graph API is applied with query words processed by structured topic modelling (STM) to compute semantic similarity and a highly relevant course is recommended to the user. The proposed architecture has a Precision of 96.81%, Recall 98.47%, Average 97.64%, F-Measure of 97.63% and has outperformed all the baseline models.