Abstract

The early detection of moderate cognitive impairment (MCI) allows for timely management, which in turn leads to improved patient outcomes. Creating a categorization system through the use of an ensemble of deep learning classifiers is the approach that this research takes in order to fulfil this goal. Using dimensionality reduction and feature selection as our primary concern, we perform preliminary processing on a diverse dataset that contains information on demographics, cognitive test scores, and brain imaging. Within the framework of an ensemble method, many deep learning architectures are utilised, with each design concentrating on a particular aspect of cognitive impairment prediction. A training and fine-tuning process is performed on each individual model on its own unique training set before the ensemble is constructed and evaluated based on a comprehensive set of performance criteria. The research presented here contributes to the development of a robust ensemble model for early MCI classification by integrating multiple different deep learning algorithms. This results in an improvement in diagnostic accuracy. The process ensures that the models may be utilised in therapeutic settings and that they are comprehensible to other individuals. When contrasted with models that are used on their own, the ensemble demonstrates superior performance, exhibiting greater memory, precision, and accuracy.

Authors

A.S. Shanthi
Dr.N.G.P. Institute of Technology

Keywords

Cognitive Impairment, Deep Learning, Ensemble Classification, Early Detection, Medical Imaging