Development of an automated tool for detecting errors in tenses

• Main Authors: Tengku Nor Rizan Tengku Mohd Maasum,, Siti Hamin Stapa,, Nazlia Omar,, Mohd Juzaiddin Ab Aziz,, Saadiyah Darus,
• Format: Article
• Language: English
• Published: Penerbit UKM 2012
• Online Access: http://journalarticle.ukm.my/4913/1/pp%2520427_441.pdf; http://journalarticle.ukm.my/4913/; http://pkukmweb.ukm.my/~ppbl/Gema/gemahome.html

The rapid growth of computer technologies creates a plethora of ways in which technology can be integrated into one of the alternatives to facilitate essay marking. Automated essay marking systems developed from the late 1960s have attempted to prove that computers can evaluate essays as competently as human expert. Several computer-based essay marking (CBEM) systems have been developed to mark students’ essays and they can be divided into semi-automated and automated systems. This paper illustrates the development of an Automated Tool for Detecting Errors in Tenses (ATDEiT™). The first phase analysed the errors found in 400 essays written by 112 English as second language (ESL) learners at tertiary level using Markin 3.1 software. The results showed that the most common errors were found in tenses. This finding led to the second phase of the research, which was the design of an automated marking tool. Consequently, the techniques and algorithm for error analysis marking tool for ESL learners were developed. An initial testing was conducted to evaluate the results of the marking tool using 50 essays. Findings showed that ATDEiT™ achieved a high level (93.5%) of recall and an average level (78.8%) of precision. This proves that ATDEiT™ has the potential to be used as an automated tool for detecting errors in tenses for ESL learners.