Context Aware Question Answer Analysis For Student Education Using Sentiment

Abstract

The currently available approaches for making recommendations based on content have two significant drawbacks. In the first place, the suggestion results are quite limited because of flaws in both the objects themselves and the algorithms that match user models. The second issue is that the recommendation system isn't aware of its context since the situation isn't given much thought. Increasing customer enjoyment via the provision of high-quality suggestions is an essential need. This study aims to improve recommendation performance by analyzing and expanding two state-of-the-art recommendation systems. The first method incorporates context information into the suggestion process; it's called the context-aware recommender. The second method considers domain semantics; it's called a recommender based on semantic analysis. The issue lies in combining them in a way that will completely take use of their potential, despite the fact that they are compatible with one another. An strategy based on Spark and MapReduce has been suggested for the context-aware recommendation system. In this work, we suggested a context aware similarity aware strategy utilizing Matlab to evaluate its accuracy, precision, recall, and F-score. Specifically, we looked at these metrics.