Original scientific paper
APPLICATION OF KNOWLEDGE DISCOVERY IN DATABASES TO AN ANALYSIS OF SURVEY DATA ON SUBSTANCE ABUSE
Goran Bubaš
; Faculty of Organization and Informatics, University of Zagreb, Varaždin, Croatia
Božidar Kliček
; Faculty of Organization and Informatics, University of Zagreb, Varaždin, Croatia
Željka Čolović Rodik
; Center for Prevention and Outpatient Treatment of Addiction, Varaždin, Croatia
Zvjezdana Fulir
; Center for Prevention and Outpatient Treatment of Addiction, Varaždin, Croatia
Abstract
Knowledge discovery in databases (KDD) was evaluated for its potential application to the analysis of survey data on substance abuse. Data for 238 psychosocial variables were collected from high school adolescents (N=4J98). One criterion variable (marijuana/hashish abuse) and 77 potential predictor variables were selected for linear regression and decision tree analyses. The results obtained by both methods are presented and compared. It is concluded that KDD techniques offer much opportunity for the discovery of new patterns in survey data, model development of substance abuse, and prevention activities.
Keywords
knowledge discovery in databases; data mining; substance abuse; survey
Hrčak ID:
78880
URI
Publication date:
14.12.1998.
Visits: 780 *