A Novel Framework for Descriptive Summary Evaluation
7 Pages Posted: 17 Feb 2018
Date Written: November 14, 2017
Abstract
Automatic evaluation in an academic institution plays a major role in the analysis of answer scripts using text analytics. In this particularly descriptive summary involves extraction of implications given by students in the exams. The prevailing evaluation methods are based on pattern matching between sentences and words in a summary. The challenge lies in interpreting the exact responses conveyed in the form of natural language. With this intuition the proposed method focuses on extracting the exact meaning by the question classification and extraction of grammatical patterns. Question classification is highly essential to infer the type of answers the student has summarized for the questions raised. The grammatical pattern analysis is the deciding factor in identifying the semantically accurate answers. This method adopts feature selection to determine appropriate features corresponding to different question types. Support Vector Machines (SVM) is used for the classifying the questions, which is more efficient than existing systems.
Keywords: Online evaluation, SVM, Grammatical patterns, question classifier
Suggested Citation: Suggested Citation