Behavior Sequence Analysis Based on Software Operation Testing

• Main Authors: Wang, Yu-Hsiang, 王郁翔
• Other Authors: Konng, Chorng-Shiuh
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/pkb4af

碩士 === 國立臺中教育大學 === 資訊工程學系 === 102 === Conventional paper and pencil testing are mostly True/False and multiple choices that the achievement and final score can only be revealed at the end of the test. It is hard to investigate how students realize and familiarize the question, and there is also lack of the operational records for further investigating. For the past decade, with the rapid development of information technology, Interactive online testing is gradually replacing the conventional paper and pencil testing. It cannot only increase students’ learning motivation, and also allows more students to perform test at the same time. Besides, it also helps decrease the time and manpower while teachers correct the students’ answers; the auto-marking and the operational processes can be performed automatically. To analysis the log file can understand the behavior when student operation in test, that can help system designer to modify the system module which is difficultly to use. Our previous studies – ITIS records students’ operations while they are performing testing. Through investigating the operational logs, which help the understanding of students’ behavior. It can also help recognize the true answer operations but not by randomly guessing answers. In this thesis, three analysis models are applied and formalized to analyze students’ operational behavior, including (1) The answering model between questions, (2) The answering model of interactive testing answering skills, and (3) The answering model of step-based test items. The corresponding automatic analysis tools are developed to investigate their behavior patterns. This thesis also conducts experiments to collect the experimental data, which are automatically encoded from their operational patters for further analysis. The sequential behavior analysis method is also introduced to find the variance of diverse behavior patterns. Therefore, several long-term variances of different students’ sequential behaviors are investigated.