Self-organizing maps for cluster analysis of risk factors of chronic absenteeism

Chronic absenteeism is one of the important predictors for assessing student performance in school as well as social and psychological issues. Absenteeism is studied in two schools from Portugal for Mathematics class. That study provides student grades and absences for the subject and 33 variables including demographic information, family status, which includes parent's marital status and guardian information, education status, which includes higher education interests, past class failures. In addition, personal predictors such as romantic relationships, internet usage, average time spent going out in a week, and alcohol consumption are included. This research is the first of its kind that uses self-organizing maps to study school absenteeism. The study helps researchers to visually identify and interpret the factors that cause school absenteeism. One of the schools had high absenteeism. The results of this study indicate that chronically absent students tend to live in urban areas, have separated parents or live with a guardian, have internet access, and are involved in romantic relationships. Those students generally spend less than two hours per week studying, go out often, and consume alcohol during weekends and weekdays.