A neural net screening of psychiatric patients

Author

Yana, Kazuo ; Kawachi, Kohji ; Iida, Kazuhiro ; Okubo, Yoshio ; Tohru, Michio ; Okuyama, Fumio

Author_Institution

Dept. of Electron. Inf., Hosei Univ., Tokyo, Japan

fYear

1994

fDate

1994

Firstpage

1366

Abstract

Describes a method for screening psychiatric patients based on a questionnaire consisting of simple yes/no questions relating to physical, mental conditions and subjective symptoms which is provided at their first visit to the hospital. The analysis of the questionnaire is important to understand patients´ background. One hundred filled out questionnaires were utilized for constructing and evaluating a neural net classifier which classifies patients into three categories i.e. schizophrenic, emotional and neurotic disorders with average correct prediction rate of 74.7%. The rate was 18.0% higher than the result given by experienced medical doctors and the method will be a useful mean for automatic screening the psychiatric patients

Keywords

medical expert systems; automatic screening; average correct prediction rate; experienced medical doctors; first visit; hospital; mental conditions; neural net classifier; neural net screening; neurotic disorders; patient background; physical conditions; psychiatric patients; questionnaire; simple yes/no questions; subjective symptoms; three layer perceptron; Biomedical informatics; Dentistry; Hospitals; Laboratories; Medical diagnostic imaging; Mood; National electric code; Neural networks; Psychology; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Engineering in Medicine and Biology Society, 1994. Engineering Advances: New Opportunities for Biomedical Engineers. Proceedings of the 16th Annual International Conference of the IEEE

Conference_Location

Baltimore, MD

Print_ISBN

0-7803-2050-6

Type

conf

DOI

10.1109/IEMBS.1994.415475

Filename

415475