Somnambulism Classification Model Using Decision Tree Algorithm

Abstract

Somnambulism, commonly known as sleepwalking, is a sleep disorder classified under parasomnias and poses potential dangers to both the individual affected and those nearby. This condition often goes unnoticed by the person experiencing it, making early detection and intervention challenging. This study aims to develop a classification model for somnambulism using the C4.5 decision tree algorithm, focusing on identifying key risk factors and supporting early diagnosis and treatment strategies. The research adopts the CRISP-DM (Cross Industry Standard Process for Data Mining) methodology, which comprises six stages: business understanding, data understanding, data preparation, modeling, evaluation, and deployment. A dataset of 400 records was used, containing attributes such as age, sleep quality, stress level, and BMI category. Analysis results revealed that the "Age" attribute serves as the root node due to its highest information gain value, indicating its significant role in classification. The constructed model achieved an accuracy rate of 71.25% and a classification error rate of 28.75%. While the overall performance of the model is fairly satisfactory, it shows limitations in accurately identifying minority classes like Insomnia and Sleep Apnea. In conclusion, the model offers potential as a decision-support tool for analyzing sleep disorders, although further enhancement is necessary to improve its accuracy and generalizability, particularly for more diverse and imbalanced datasets