发明名称 |
SYSTEM AND METHODS FOR TRAJECTORY PATTERN RECOGNITION |
摘要 |
A multiple imputation (MI) based fuzzy clustering with visualization-aided MI validation that improves the accuracy and the stability of identified patterns, generally the structure of HD data with missing values. |
申请公布号 |
US2016358040(A1) |
申请公布日期 |
2016.12.08 |
申请号 |
US201515116570 |
申请日期 |
2015.02.06 |
申请人 |
UNIVERSITY OF MASSACHUSETTS |
发明人 |
Fang Hua |
分类号 |
G06K9/62;G06N7/02;G06T7/20 |
主分类号 |
G06K9/62 |
代理机构 |
|
代理人 |
|
主权项 |
1. A computer-implemented method for visualizing pattern recognition of longitudinal data without changing the classification of the longitudinal data in high dimensional space for analyzing a heterogeneity effect, comprising the steps of:
receiving one or more datasets associated with a population, or sample thereof, of subjects, wherein the dataset includes a plurality of components and one or more attributes associated with each of the plurality of components; imputing missing attribute values into the one or more datasets based on a multiple imputation technique; identifying one or more trajectory patterns or one or more data clusters from the imputed datasets; performing a multiple imputation, based validation on one or more trajectory patterns or one or more data clusters; comparing data associated with one or more individual subjects to the one or more trajectory patterns or one or more data clusters; and illustrating, on a display device, the one or more trajectory patterns or one or more data clusters. |
地址 |
Boston MA US |