306829
Using Data Mining Technique to Study the Likelihood of Dental Students from a Disadvantaged Background Practicing in a Disadvantaged Community
Methodology: The dataset consisted of 213 dental students from 2006-2009 matriculation years and of these students, 188 graduated in 2010-2013. The geographical information system (GIS) was used to extract variables (e.g., median household income, household size) within the census tracts which are linked with the addresses of parents’ home and dentist practicing location. K-means cluster analysis was used to classify the students into three groups, disadvantaged, moderately disadvantaged, and non-disadvantaged. The means of median household income are for disadvantaged group- $33,000, moderately disadvantaged group- $62,000, and non-disadvantaged group- $112,000. The means of unemployment rates for the three groups are 13%, 8.6%, and 6%, respectively. Decision tree analysis was used as a benchmarking tool to verify the accuracy of classification. Partial least square (PLS) regression was implemented to demonstrate prediction validity of clustering variables.
Results and Conclusion: Over 90% of students from a disadvantaged background, 88% of students from a moderately disadvantaged background, and 52% of students from a non-disadvantaged background are practicing dentistry in a disadvantaged community. The vast majority of students who have graduated practice in a disadvantaged community, indicating that they significantly contribute to the College’s mission in serving underserved community and reducing health disparities.
Learning Areas:
Biostatistics, economicsPublic health or related education
Learning Objectives:
Demonstrate to the audience the various artificial intelligence models such as cluster analysis, decision tree algorithm, and partial least squares regression. Analyze the relationship between students from a disadvantaged background and the likelihood of practicing in an underserved community.
Keyword(s): Health Disparities/Inequities, Oral Health
Qualified on the content I am responsible for because: Biostatistics professor at Meharry Medical College
Any relevant financial relationships? No
I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.