Text Segmentation with Feature Similarity for Exam Assessment using Machine Learning

Abstract

The need of green computing in order to reduce the excess use of paper to assess the theoretical answer is a serious demand. We therefore intend to provide a solution by building a model which helps in evaluating the theoretical answers online to reduce the human efforts. The paper involves the use of machine learning, NLP, keyword extraction and matching aggregation for checking the similarity between the user answer and the specimen answer. The user written answer is tokenized into bag of words and the meaning of words are extracted and matched with the specimen answer for semantic analysis. The machine learning algorithm analyses the answer and gives the percentage of similarity between the two answers with this system we can automatically evaluate the theoretical answers easily and efficiently, thus reducing the use of paper.