A Descriptive Answer Evaluation System Using Cosine Similarity Technique

Use of computer in spite of human in knowledge discovery and engineering is a trend from last decade. In situation of COVID-19, it is need of hour to use such computer assisted assessment tools which will reduce the work of teachers in education domain. In Literature various techniques were developed to assess the answers of objective types questions by using machine learning but very less work carried out on assessment of answer of descriptive type questions. This paper provides a solution for the assessment of answer of descriptive type questions where teacher need not to use any paper or pen but the computer imitates like as teacher and evaluate answer of students. The primary objective is to represent results of subjective answers in the form of pictorial representations using cosine similarity technique. Cosine similarity is one of similarity measures which find similarity between two objects irrespective of their size. The results of web-based application are sufficient enough to withdraw the traditional teacher centric assessment approach and develop a computer-aid solution which is beneficial for educational institutions.