RCTS Math Seminar

DATE2015-08-24 10:10-11:00


SPEAKER關聖威 教授(Department of computer science and software engineering at Xi’an Jiaotong-Liverpool University

TITLE Input Space Partitioning for Machine Learning

ABSTRACT This talk introduces an input attribute grouping method to improve the performance of learning. During training for a specific problem, the input attributes are partitioned into groups according to the degree of inter-attribute promotion or correlation that quantifies the supportive or negative interactions between attributes. After partitioning, multiple sub-networks are trained by taking each group of attributes as their respective inputs. The final classification result is obtained by integrating the results from each sub-network. Experimental results on several UCI datasets demonstrate the effectiveness of the proposed method.