Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{206210,
        author = {Thamina anzum A and Thirumahal R and Gobika R and Vinithaa and Naveen Ragav K and Naveen P},
        title = {Intelligent Depression Classification Based on Hybrid Models Using Actigraphy},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {2},
        pages = {392-404},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=206210},
        abstract = {Depression is a serious mental health problem that affects an individual’s emotion and daily activities which requires early and accurate diagnosis. Traditional methods depend on clinical interviews and questionnaires, which may not capture continuous behavioural changes. By using wearable devices, actigraphy-based data provides continuous values of human behaviour that is used for automated depression assessment. In this work, we develop an “Intelligent depression classification system” using actigraphy data based on two hybrid machine learning models: CatBoost-Artificial Neural Network (CatBoost-NN) and LightGBM-Artificial Neural Network (LightGBM-NN). The system classifies individuals into bipolar I, bipolar II and unipolar categories. Tree based models (CatBoost and LightGBM) are used to learn from clinical patient information, while Neural Network is used for identifying deep non-linear relationships from activity-based and behavioural features. The dataset contains minute-level data collected from depressed and healthy subjects, collected using wearable actigraphy sensors. The hybrid models combine the strengths of both approaches to improve classification. Explainable AI techniques using SHAP are used to interpret model predictions and identify key features that result in depression detection. The proposed hybrid approach aims to achieve higher accuracy when compared to traditional singlemodel methods, providing the effectiveness of combining wearable sensor data with hybrid and explainable machine learning techniques for automated depression classification.},
        keywords = {Depression classification; Actigraphy data; Wearable sensors; Hybrid machine learning; Explainable AI; Mental health diagnosis},
        month = {July},
        }