In response to the growing popularity of social web apps, much research has gone into analyzing and developing an AI-based responsive suggestion system. Machine learning and neural networks come in many forms that help online students choose the best texts for their studies. However, when training recommendation models to deal with massive amounts of data, traditional machine learning approaches require additional training models. As a result, they are deemed inappropriate for the personalized recommender generation of learning systems. In this paper, we examine LSTM-based strategies in order to make useful recommendations for future research.