Assessment Of Descriptive Answers in Moodle-based E-learning using Winnowing Algorithm

Abstract

Assessment in education allows for obtaining, organizing, and presenting information about how much and how well the student is learning. An automatic evaluation tool is proposed that allows the assessor to evaluate descriptive answer at any time and receive instant feedback of the students. Due to the lack of descriptive answer grading in Moodle-based E-learning system, there is a need to build a model and also add this feature as a plug-in for the E-learning system .Up until today, most assessors still choose to examine descriptive document manually for each student document. This method takes time where assessor needs to be focused and thorough while examining a number of descriptive documents. This reason often affects essay examinations to be less objective and not optimal. To improve objectivity, time efficiency and fair correction in descriptive answers assessment process, a system is needed that can automatically assess student documents, or in other words a descriptive answer evaluating system. An evaluation of descriptive answer system, works by analysing student answer document with model answer document. The higher the semantic similarity, the higher the score obtained.The purpose of this paper is to check the similarity between the teacher’s answer and the student’s answer using Winnowing algorithm. Winnowing algorithm is one of the document fingerprinting algorithms that can be used to detect document similarity by using hashing technique. The fingerprint document itself is a method used to detect document similarities with other documents. The Winnowing algorithm has fulfilled one of the requirements of the plagiarism algorithm, which is whitespace insensitivity, disposing of irrelevant characters such as punctuation. The similarity value is calculated using Jaccard Coefficient. Later this assessment is used for grading the student’s performance.