SAMS: Data and Text Mining for Early Detection of Alzheimer's Disease
Christopher Bull
SCC, Lancaster University
The SAMS project (Software Architecture for Mental health Self-management) is investigating whether monitoring data from everyday computer-use activity can be used to effectively detect subtle signs of cognitive impairment that may indicate the early stages of Alzheimer's disease.
In this talk I will discuss the SAMS project, the collection of data and text form participants, and our approach to mining the text to infer cognitive health. During the SAMS project, bespoke software is used to collect data and text from participants (installed on the participants' home PCs). The collection software passively and unobtrusively collects many forms of data and text from the participants' PCs (inc. typed email and document text), which is securely logged, and later transferred to our server for analysis. The analysis consists of various data and text mining techniques to attempt to map trends and patterns in the data with clinical indicators of Alzheimer's Disease, e.g. working memory, motor control.
Tools usage within the SAMS project will also be discussed, including the development of the bespoke collection and analysis software, as well as existing tools that are re-used (Part of Speech Tagger, Semantic Tagger).