An Automated Approach to Subjective Answer Evaluation Using ML and NLP

There are many approaches to subjective answer evaluation and in this project, one of these effective approaches is used and put into best use. In simple words machine evaluates and grades the answer that is provided. In this proposed system cosine similarity algorithm is used to get the amount of similarity between the answers and grade them accordingly. When the error percentage is observed after comparing the model with human valuation nearly there is 87% accuracy. Which means it has pretty low error percentage. There are many algorithms that is used for subjective answer evaluation but proposed algorithm has least time complexity and also proposed system has more accuracy than the existing systems. The main aim is to bring the subjective answer evaluator into use as much as possible and work on further improvement.