AI Case Study

Duke University is researching how to predict the onset of neurological disorders such as Parkinsons by analysing Microsoft-Bing search logs using machine-learned classifiers

Duke University in partnering with Microsoft to detect evidence of Parkinson’s disease (PD) by analysing longitudinal log data from search engines.

Industry

Healthcare

Healthcare Providers And Services

Project Overview

Data from search logs was used to check for

* Symptom or synonym related search
* Motor symptoms such as cursor movements, including speed, direction changes, tremors
* Presence of repeat queries and result clicks
* Other risk factors such as age, gender or prior trauma, etc.

Using this data researchers have found promising results in using search patterns to predict onset of PD.

Reported Results

During research phase, Classifier sensitivities for PD detection are 94.2/83.1/42.0/34.6% at false positive rates (FPRs) of 20/10/1/0.1%

Technology

Function

R And D

Product Development

Background

Nearly one in six people suffer from neurological disorders such as Parkinsons. Researchers at Duke University are developing a model to predict early onset of such conditions.

Benefits

Data

Deidentified logs of United States search activity from the Microsoft Bing web search engine, of 31,321,773 English speaking search engine users.