Unsupervised Content-Based Book Recommendation System Using DeepWalk

Reading is a beneficial activity that anyone can do to increase their skills such as critical thinking skills and knowledge, unfortunately, low literacy rates are still prevalent in developing country such as Indonesia, to address this problem a book recommendation system is needed to increase people's interest in reading, this paper introduces the utilization of DeepWalk for creating an unsupervised content-based book recommendation system. It's starts by first doing pre-processing to the the dataset by cleaning the dataset and turning the dataset into graph, implement the DeepWalk's algorithm to the graph, and then conduct a study case to use the model to recommend books similar to the book “Anna Karenina” as the basis for the recommendation. It is found that based on the results, the model achieves its goal in recommending a sensible and robust collection of recommended books and thus makes DeepWalk a promising algorithm in making a book recommendation system.