Application Of Decision Tree Algorithm In Classifying The Level Of Impulsivity In EEG Signals

• Published in: 2024 2nd International Conference on Technology Innovation and Its Applications (ICTIIA) pp. 1 - 6
• Main Authors: Yennimar, Manday, Dhanny Rukmana, Sembiring, Anita Christine, Lumbantoruan, Nurima, Ginting, Arico Sempana, Simanjuntak, Ruth Marsaulina, Sitanggang, Delima
• Format: Conference Proceeding
• Language: English
• Published: IEEE 12.09.2024
• Subjects: Classification; Decision Tree; Electroencephalography; Genetics; Impulsivity; Partitioning algorithms; Planning; Prediction algorithms; Predictive models; Recording; Regression tree analysis; Sleep; Technological innovation
• DOI: 10.1109/ICTIIA61827.2024.10761187

Impulsivity is the tendency to act spontaneously without proper planning. Everyone has different levels of impulsivity, influenced by factors such as genetics, environment, and psychological state. Measuring and detecting impulsivity is crucial due to its substantial impact on various aspects of an individual's life. The factors that influence impulsivity levels include the social environment, stress levels, mental health, and genetic factors. People with high levels of impulsivity tend to make quick decisions without considering risks, have difficulty restraining themselves, act on impulse, have difficulty maintaining focus for long periods, and often take unnecessary risks. The level of impulsivity reflects how often a person acts without thinking about the impact of their actions. EEG (Electroencephalography) is a medical technique for recording the brain's electrical activity. EEG is frequently employed to diagnose and monitor conditions affecting brain activity, including epilepsy, sleep disorders, and various neurological issues. The decision tree algorithm in machine learning constructs predictive models by sequentially making decisions based on data features. It partitions the data into subsets by choosing features that best differentiate the data at each stage, creating a tree-like framework where each branch signifies a decision and each leaf denotes the outcome or prediction. Decision trees are versatile, supporting both classification and regression tasks by effectively managing intricate and non-linear datasets. Key benefits include their strong interpretability, capacity to handle categorical and numerical data, and ease of comprehension for humans. The findings from the study, which classified subjects into four categories, non-impulsive, impulsive, potentially impulsive, and potentially impulsive demonstrated a high accuracy rate of 92.9%.