Interval-Valued Examples Learning Based on Fuzzy C-Mean Clustering
Mingzhi Chen, Guolong Chen, Shui-Li Chen · 2006
In this paper, a new approach to generate decision tree from those examples with interval-valued attributes is presented and then rule matching is made. Considering that the interval values of the same attribute of all examples probably fall into certain distributing rule so as to form some center points, we cluster the interval-valued attributes of all examples by using the algorithm of FCCID (fuzzy c-mean clustering for interval-valued data). Consequently, the attributes represented by interval data are transformed into those represented by fuzzy degree of membership. On the basis of that, the fuzzy ID3 algorithm is adopted to generate a decision tree for rule matching