A Proposed Framework for Automated Online Question and Answering System for Interview using Latent Semantic Analysis

Engku Fadzli Hasan, Syed Abdullah and Azrul Amri, Jamal and Norsyahira, Jano (2017) A Proposed Framework for Automated Online Question and Answering System for Interview using Latent Semantic Analysis. In: The 5th International Conference on Artificial Intelligence, Computer Science, & Information Technology, 31 July 2017- 1 August 2017, Batu Ferringhi, Pulau Pinang.

[img] Text
FH03-FIK-17-09889.pdf
Restricted to Registered users only

Download (317kB)

Abstract

Interview is an important process to select the best applicant during employee selection process for a company, student enrollment in university and others. The interview process is conducted to find out more about respondent knowledge, personality and background. Interview can be conducted by face to face interview session or online interview session. Online interview session can be structured or unstructured Question and Answer (Q&A). Online interview tends to use structured Q&A where candidates are not given a chance to reveal their knowledge. Because of limitations in the structured online interview, unstructured online interviews started being used by many organizations. However, using unstructured online interviews, it is difficult to determine accuracy of comparing the answer. In this research, we proposed a framework for automated online interview system using Latent Semantic Analysis (LSA) to evaluate the answer. This research will used Natural Language Processing techniques to pre-process the answer and LSA to evaluate the answer. Further in this study, the result of this process will be evaluated and compared with the expert evaluation to grade the answer.

Item Type: Conference or Workshop Item (Paper)
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Faculty of Informatics & Computing
Depositing User: Muhammad Akmal Azhar
Date Deposited: 18 Nov 2020 02:32
Last Modified: 18 Nov 2020 02:32
URI: http://eprints.unisza.edu.my/id/eprint/1559

Actions (login required)

View Item View Item