This paper presents a framework for data mining and automatic diagnosis in the context of Alzheimer's disease and frontotemporal dementia. Using multi-objective genetic algorithms, we select the most discriminant features based on neuropsychological and neuroimaging data. Subsequently we develop multiclass diagnostic models based on Machine Learning techniques.

In this paper we use Machine Learning techniques for the development of diagnostic models of frontotemporal dementia and Alzheimer's disease based on data from neuropsychological assessments. We combine feature engineering techniques with classical Machine Learning models demonstrating how Artificial Intelligence approaches can be successfully used for the automatic diagnosis of these types of diseases.

In this research, genetic algorithms were applied as a feature engineering technique in conjunction with Machine Learning algorithms for the differential diagnosis of patients with Alzheimer's disease, frontotemporal dementia and different subtypes of Primary Progressive Aphasia.