An Intelligent System for Evaluation of Descriptive Answers

Examination is an efficient and conventional way to test the knowledge gained by the students. The evaluation of examinations is a tiresome and arduous task. Automatic answer-script evaluation makes this task convenient for teachers as it reduces the efforts and time taken. The existing systems provide teachers the facility to conduct automatic objective and short answers exams but evaluating descriptive answers automatically is a challenge. For assessment of descriptive answers, the machine needs to consider the important factors which a human assessor takes into account for manual evaluation. To solve these problems, this research work proposed an intelligent assessment platform which considers the question type, necessary keywords, structural, conceptual and language aspects to evaluate an answer. Proposed model used concept graphs, fuzzy string matching, grammar checking and other similarity metrics for the assessment. The system also generates personalized feedback and analysis reports for students and teachers to promote focused learning.