AutoEval: A NLP Approach for Automatic Test Evaluation System

In this modern era, there is a need for automation in answer assessment systems as the world moves towards automation. Since online response evaluation is only available for option-based questions, the evaluator's job is made more difficult when evaluating theory answers. The teacher manually checks the answer and assigns the required grade. The current system necessitates more manpower and time to evaluate the response. In this paper, we demonstrate the working and operation of the ‘AutoEval - Automatic Test Evaluation System’, a computerized or automatic exam paper evaluation system. A manual paper evaluation is time-consuming and exhausting, but computerized exam paper evaluation aids in standardizing the paper correction process in an effective and reliable manner. The evaluation of response scripts is an essential part of evaluating students' results. In most cases, a response evaluation is performed manually, which can lead to bias. The assessment is based on several factors including the evaluator's mood swings and the inter-relationship between the student and evaluator. Furthermore, assessment and evaluation are a time-consuming and repetitive process. Natural language processing (NLP) methods are used to evaluate exam papers in a variety of ways. NLP is used to search for grammatical errors, as well as to perform syntactic analysis, semantic similarity and database storage.