Statistics Seminar: STA 290
Research Training Group at the Department of Statistics University of California, Davis
Friday January 11th, 2013 at 4.10pm, MSB 1143 (Seminar Room)
Refreshments 3:30pm, prior to seminar in MSB 4110 (Statistics Lounge)
Speaker: James Long, University of California, Berkeley
Title: "Classification of Sparse, Irregularly Sampled Time Series and Heterogeneous Feature Noise"
Abstract: In high dimensional classification problems, a common practice is to extract features from each observation and then train a classifier on the features and associated observation classes. Here, I consider classification of periodic variable stars, which are essentially sparse, irregularly sampled time series. The sampling of the time series introduces heterogeneous noise into derived features. Classifiers which do not account for the feature noise often have high misclassification rates. In this talk, I will discuss methods for addressing the feature noise, present results from empirical studies, and highlight the relationship between feature noise and regularization of classifiers.