Abstract
Recently, drug addict has become a serious social problem in the Kingdom of Thailand. Up until now the drug rehabilitation centers are making effort to research and improve a model of treatment, even though the Therapeutic community model and the Matrix model are evaluated as effective which are considered the culture background of the Kingdom of Thailand. But even if the patients are cured with these two rehabilitation models, getting effective evaluation when the rehabilitation ends, after they return to society, a recurrence problem still occurs frequently. The result of a survey of the curative affect after rehabilitation shows that the effectiveness of each rehabilitation model depends on background and personality of each patient, nature of their past habit and their drug abuse period. According to the aforementioned, the process to choose an effective rehabilitation model for each patient is extremely important subject of concern in the Kingdom of Thailand.
The approach of pattern recognition, which will be reported in this paper, is to predict the effectiveness of each rehabilitation model. The prediction is made based on the information from doctoral measuring and social activity measure analysis.
In this paper, the item selection model using the multi-dimensional pattern recognition MT-system was used. The items were selected in such a way that maximizes the difference in Mahalanobis distance between two groups of patients of each rehabilitation model: one of which consists of patients who recovered after 4 months and the other group consists of patients who could not.
We propose the method of judging which rehabilitation method is suitable for each drug addict in this paper.
