1. Analyzing the Accuracy of Answer Sheet Data in Paper-based Test Using Decision Tree
- Author
-
Aris Puji Widodo, Edy Suharto, and Suryono Suryono
- Subjects
education ,Computer science ,Decision tree ,Paper based ,data mining ,computer.software_genre ,lcsh:QA75.5-76.95 ,lcsh:HD72-88 ,Test (assessment) ,lcsh:Economic growth, development, planning ,Comprehension ,paper-based test ,Order (business) ,decision tree ,Data mining ,lcsh:Electronic computers. Computer science ,Raw data ,computer ,Single layer ,Test data - Abstract
In education quality assurance, the accuracy of test data is crucial. However, there is still a problem regarding to the possibility of incorrect data filled by test taker during paper-based test. On the contrary, this problem does not appear in computer-based test. In this study, a method was proposed in order to analyze the accuracy of answer sheet filling out in paper-based test using data mining technique. A single layer of data comprehension was added within the method instead of raw data. The results of the study were a web-based program for data pre-processing and decision tree models. There were 374 instances which were analyzed. The accuracy of answer sheet filling out attained 95.19% while the accuracy of classification varied from 99.47% to 100% depend on evaluation method chosen. This study could motivate the administrators for test improvement since it preferred computer-based test to paper-based.
- Published
- 2019