Natural Language Processing based Recommendation System for Courses *

The goal of Course Recommendation System based on Natural Language Processing, is to suggest courses to users based on their various searches. Users now have an easier time locating the appropriate courses based on their searches thanks to this system. To determine whether a course is available, the system employs the TF-IDF algorithm and cosine similarity. The suggested method made use of natural language processing. The dataset must first be correctly prepared. The first module involves the dataset's preparation. Stopwords are removed, and the number of repeated rows has been reduced sufficiently. The dataset is thus ready. The first two fields of the dataset are combined to form a new field. The CourseTitle field is made by combining the Title and Description fields from the prepared dataset. Utilizing TF -IDF, the new dataset has been vectorized. The cosine similarity of the search to the TF-IDF vector matrix is considered when a user searches for a particular word. The cosine similarity determines the score. Now, the scores are given in decreasing order. Basedon the score values, the users preferred course is recommended. A methodology for recommendation systems is proposed that is more efficient and can overcome the difficulties posed by the cold start problem andthe data sparsity problem.