Analysis and Classification of Word Co-Occurrence Networks From Alzheimer’s Patients and Controls

In this paper we construct word co-occurrence networks from transcript data of controls and patients with potential Alzheimer’s disease using the ADReSS challenge dataset spontaneous speech. We examine measures structure these for significant differences, finding that have a lower heterogeneity centralization, but higher edge density. then use measures, network embedding method some frequency distribution to classify transcripts into control or Alzheimer’s, estimate cognitive test score participant based on transcript. find it is possible distinguish between AD alone, achieving 66.7% accuracy set, predict scores root mean squared error 5.675. Using more successful than method. However, if are shuffled relatively few different, indicating drives many properties. This observation borne out by classification experiments, where perform similarly measures.