Combining Technologies of AI and Fuzzy Logic System for Huntington Disease Prediction

• Published in: 2024 4th International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE) pp. 697 - 702
• Main Authors: Kumar, Vikram, Verma, Apurv, Raman, Ramakrishnan, Rastogi, Shailesh, Saini, Dimple, Meenakshi, R.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.05.2024
• Subjects: Artificial Neural Network (ANN); Artificial neural networks; Data models; Functional Capacity; Fuzzy logic; Huntington Disease; Hybrid Neurofuzzy Model; Mamdani Type-1 Fuzzy Mathematical Expert System; Mathematical models; Mean Squared Errors; Predictive models; Reaction State; Regression Values; Support vector machines; Training
• DOI: 10.1109/ICACITE60783.2024.10616600

We provide a novel approach to predicting the deterioration in the response state of patients suffering from neurological movement disorders, which include hand tremors and involuntary motions, similar to the symptoms seen in patients with Huntington's disease (HD). We provide a hybrid neurofuzzy model approach that combines a fuzzy logic system (FLS) to determine the response stage and an ANN, which is an artificial neural network, to forecast the function level (FCL). A dataset consisting of 20 participants-a mix of healthy and HD individuals-was carefully analyzed. Subjects were given smartphones or tablets were instructed to recognize circular items on the screen. Complete explanations are provided for the neural network's initial data gathering and labeling procedures, training algorithm selection, fuzzy logic controller modeling, hybrid model construction, and hybrid model deployment. The neural network using feed-forward backpropagation (FFBP) produced remarkable regression R values of 0.98 and mean squared errors (MSEs) of \mathbf{0 . 0 8}. In addition, the FLS provides a thorough evaluation of a person's response state with respect to FCL.