Tohoku University

Hui Shen

Tetsuo Endoh

Yitao Ma

A clustering device includes: an evaluation score calculation section configured to calculate an evaluation score or evaluation scores for evaluating a classification result; a batch process section configured to classify multiple element data into clusters with an optimum number of clusters, based on the evaluation scores respectively obtained for different number of clusters by assigning each of the multiple element data to one of the clusters; an update process section configured to assign newly added element data to a cluster that is closest to the newly added element data among the clusters into which the multiple element data are classified by the batch process section; and a determination section configured to determine validity of a classification result after assigning the newly added element data to the cluster, based on the evaluation score obtained by assigning the newly added element data to the cluster by the update process section.

Clustering device and clustering method

Provided is a method for calculating an evaluation score of clustering quality, based on the number of clusters into which a plurality of data is clustered. The calculating the evaluation score includes: calculating a degree of internal compactness that is a sum of values, each being defined by normalizing a first index value by a first value that is based on a number of data within each cluster, the first index value indicating a degree of dispersion of data within each cluster; calculating a degree of external separation defined by normalizing a sum of a second index value for each cluster by a second value that is based on the number of clusters, the second index value indicating an index of a distance between the clusters; and calculating the evaluation score according to a predetermined formula having, as variables, the degree of internal compactness and the degree of external separation.

Growing community of inventors

Miyagi, Japan

Hui Shen