[Mar 31, 2017] Machine Learning in Population Research
Mar 31, 2017
from 12:00 PM to 01:00 PM
|Where||Room B005, 206 W. Franklin|
|Add event to calendar||
Chirayath M. Suchindran, Ph.D., Professor, Biostatistics, UNC-Chapel Hill
Suchindran’s recent publications include a paper that demonstrates the use of mixture models with linear predictors to identify incorrect gestational age in US birth certificates, a regression analysis of interval censored complex survey data, an examination of the redistribution techniques to identify the errors in causes of death data, a paper in Demography that demonstrates the use of event history data to obtain estimates of multistate life table parameters and their standard errors and a paper dealing with the appraisal of biomarker selection methods applicable to HIV/AIDS research.
Suchindran’s current, substantive research projects include genetic by context influence on trajectories of adolescent health risk behaviors, effects of cash transfer and community mobilization on HIV incidence and gender norms among South African young women, effect of neighborhood SES on coronary heart disease burden in communities and obesity development and CVD risk factor clustering in Filipino women and offspring, and promoting safe sex among HIV+ women.
Suchindran’s work will focus on developing methods for data analysis and conduct of collaborative research with CPC researchers that have been initiated currently. The methodological focus will be on the issues related to complex sampling designs and estimation of random effects models with specific complex survey designs with different modes of data collection, currently being proposed to collect the Add Health Survey wave V. Suchindran will also pursue his research in developing indices of longevity based on Kullback-Leiber divergence measures that involve moments of order three and above (for example, skewness and kurtosis). Suchindran will also join in the new initiatives in the Department of Biostatistics in developing state-of-the-art methodology for analyzing ‘big data’ (for example, machine learning techniques for predictive modeling, dimension reduction) with focus on demographic research. On the collaborative research side, Suchindran will continue his involvement in several intervention studies where the data collections are about to end. Suchindran will also provide statistical help, if necessary, by developing methods for data analysis for CPC researchers.