Representation of User Interest Model Based on Attribute Coordinate Analysis
It was discussed how to establish a machine-learning user interest model based on attribute coordinate analysis, and enable the information system to cluster by personalized data and realize user interest-oriented initiative recommendation service. The inference relationship was established by the characteristic that there was a mutual information correlation between data attributes, and thus an attribute simplex model was constructed. The center of gravity of attribute simplex subdivision was a stationary point which could show the essence of user interest. Digital lithography of attribute simplex could be performed by attribute linear coordinate system, and clustering similarity could be obtained by evaluating the Euclidean geometric distance between the coordinates of the stationary point. The results showed that the method - to subdivide and infer the center of gravity of the attribute in the attribute simplex at first to obtain the center of gravity of attribute and then cluster data for the center of gravity by utilizing attribute coordinate could mine and utilize the covert semantic information between data attributes more sufficiently than the similarity calculation method of directly using data attribute weights.
machine learning user interest model personal clustering attributive Coordinate analysis
ZHOU Ru Qi
Department of Computer Science Guangdong University of Education Guangzhou, Guangdong Province, China
国际会议
武汉
英文
66-69
2010-12-04（万方平台首次上网日期，不代表论文的发表时间）