Semantic Analysis for Document Validation
Abstract
Every year many competitive, institutional and non-institutional examinations are conducted. Most of the institutional exams demand descriptive type answers. However, only objective answers are being evaluated by the machines effectively. Evaluation of descriptive answers has always been a hurdle in the development of Electronic learning. The existing systems consider just the keywords or frequency of the terms but not the meaning of the answers while evaluating. The proposed system is expected to solve the above problem by using dependency parsing. Here we make use of the relationship between the terms in a text to extract its meaning. The system is expected to enhance the evaluation process by reducing the time and manpower thereby improving the quality.
Keywords: Automated evaluation, Descriptive answers, Dependency parsing, Semantic similarity