Mobile IT-diagnostic system for Alzheimer's disease recognition

Abstract

The article presents the structure of an IoT-based mobile system for the IT-diagnosis of Alzheimer's Disease (AD), including mobile applications, server management, algorithms development, and data protection. The authors conducted analytical and predictive work on the mobile diagnosis of Alzheimer's disease based on decoded textbased speech data from patients using machine learning algorithms and a neural network. The data used in the article was taken from the Address 2020 Challenge kit, which contains speech data from patients with Alzheimer's disease and healthy people. The system combines the technologies of IoT, machine learning and blockchain, supporting early detection of AD and its continuous monitoring. The mobile application module performs voice data recording patients, preprocessing, and predictive analysis by contacting the server via requests. The data management and preprocessing module structures the input data by converting the raw voice data into JSON format for server-side analysis via the Flask backend, with additional integration of the EMQX broker for real-time data distribution. In the algorithms and analysis module (on the server), text features are highlighted and classiication algorithms are used to obtain probabilistic predictions for patients with Alzheimer's disease. Users receive diagnostic feedback in real time through the results visualization module, which can transmit diagnostic results to medical professionals. The system's security and privacy module uses IPFS-based decentralized storage and blockchain-driven access control to ensure data integrity and support authorized data retrieval.