PROJECT TITLE :

Classifying Text-Based Computer Interactions for Health Monitoring

ABSTRACT:

Detecting early trends indicating cognitive decline can allow older adults to raised manage their health, however current assessments present barriers precluding the employment of such continuous monitoring by shoppers. To explore the consequences of cognitive status on pc interaction patterns, the authors collected typed text samples from older adults with and without pre-delicate cognitive impairment (PreMCI) and constructed statistical models from keystroke and linguistic features for differentiating between the two teams. Using both feature sets, they obtained a 77.one percent correct classification rate with seventy.vi p.c sensitivity, eighty three.3 p.c specificity, and a zero.808 space beneath curve (AUC). These results are consistent with current assessments for MC--a more advanced disease--but using an unobtrusive methodology. This analysis contributes a combination of options for text and keystroke analysis and enhances understanding of how clinicians or older adults themselves might monitor for PreMCI through patterns in typed text. It's implications for embedded systems which will enable healthcare providers and consumers to proactively and continuously monitor changes in cognitive operate.