Depression and Anxiety Monitoring using Intelligent Neural Detection

Identifying mental disorders such as depression and anxiety remains a major challenge, especially in the era of social media and digital communication. Conventional methods are lengthy, subjective, and inaccessible. To address these, the current study introduces a new Multilayer Steerable Dense Net-Spiking Transformers Neural Network with Meerkat Optimization Algorithm (MSDSTN2Nets+MOA) framework for Detection of Depression and Anxiety (DDA) in Social Media and Voice Data. The input voice and social media data, obtained from the DAIC-WOZ and RSDD datasets, are initially preprocessed using the Quasi-Cross Bilateral Filtering (QCBF) method. After preprocessing, the Inverse Z-transform and Wiener-Hopf Factorization (IZ-WHF) method is utilized for feature extraction. Depression and Anxiety Detection relies on Multilayer Steerable Dense Net-Spiking Transformers Neural Network (MSDSTN2Nets) and optimization via the Meerkat Optimization Algorithm (MOA). DAIC-WOZ and RSDD datasets are applied to assess the performance of the proposed MSDSTN2Nets+MOA model, which depicts an accuracy rate of 99.9% andan F1-score of 99.8%. The developed model is implemented using the Python programming language. The created model successfully identifies anxiety and depression from voice and social media data showing good accuracy, strong feature extraction, effective optimization, and solid performance on a variety of benchmark datasets