| Autor: |
Jabłoński, Ryszard, Turkowski, Mateusz, Szewczyk, Roman, Ślubowski, R., Lewenstein, K., Ślubowska, E. |
| Zdroj: |
Recent Advances in Mechatronics; 2007, p216-221, 6p |
| Abstrakt: |
The results of detection of alcohol addiction based on the analysis of human sleep are presented in this paper. Sleep was described by numerical parameters calculated from the standard processed records of polysomnography (PSG) signals. The database used in the experiments consisted of almost 200 examinations: 50% of healthy and alcoholic addicted patients, and 50% males and females, with normal age distribution. We have used two different methods: statistical estimator and neural networks to evaluate the diagnostic value of the sleep parameters. We have proposed the set of 13 basic parameters to detect alcohol addiction. The differences in diagnostic value of these features are noticeable, but not very significant (the differences of the diagnosis correctness lie between +2% and −4%), but each of them improves the total quality of learning process. Finally, we have obtained about 75% correctness of alcohol addition diagnoses. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
