Detecting Cheats In Online Student Assessments Using Data Mining

We can find several online assessment applications, Windows oriented or Web based, licensed or gnu free software, proprietary or standardized. All of them executing basic questions and test interoperability stages: providing assessment items, training and/or evaluation, and the assignment of a grade. Tons of information resulting of this educational process is stored into databases, including starting times, local or remote IP addresses, finishing times and, the student's behavior: frequency of visits, attempts to be trained, and preliminary grades for specific subjects, demographics and perceptions about subject being evaluated. We propose the use of data mining to identify students (persons) that commit cheat in online assessments (cyber cheats) and identify patterns to detect and avoid this practice.