Computer-assisted valuation of descriptive answers using weka with randomforest classification

Abstract

In the present scenario, digitization is one of the most important goals of India. It has a serious impact on the education system, and exams are the basic requirement of education. To make exams digitized, a large number of resources are required. But a smaller step can be taken to digitize exams with minimum resources. In this research work, a classifier-based approach is proposed for the valuation of descriptive answers. An experiment is carried out in our academic organization to build the required dataset. Dataset consists of 530 training samples and 159 testing samples that are classified by using RandomForest classification. It is observed that 61.63% answers are exactly evaluated without any predefined threshold value and 96.22% with a manual threshold value of ±1 mark.