reporttitle:project management: manifold learning for financial data
reporter(institution):gang kou(southwestern universityof financeand economics)
time:02:00.pm,6th nov. 2015location:b-201, building of economics & management, jiulonghu campus
abstract:identification of characteristics and structure of high-dimensional data is an important task for financial analysis. through manifold learning method, we extract the intrinsic features and reduce the dimensions of the high-dimensional financial data, to display the internal relations and structure of the financial big data. so that people can achieve the effective objective data support for financial analysis. first, a kernel entropy manifold learning algorithm is introduced to measure the relationships between two financial data points and produce a low-dimensional representation of high-dimensional financial data. second, an information metric-based manifold learning (imml) method is proposed to identify the intrinsic manifold structure of a financial market.
