Semi-supervised Learning of Database Annotated Data Clustering Method

摘要

When using the traditional K-means clustering method for clustering, this number of clusters must be obtained in advance. Aiming at the above problems, improved K-means clustering algorithm based on semi-supervised learning is proposed for database annotation data clustering. First, the spanning minimum tree of the graph is established with a small amount of label data, and the cluster initial and number cluster required for the K-means clustering are alliterative split, and then, according to the defined K-means method flow. Experiments show the number of iterations is smaller than traditional, the stability is greatly improved, and the clustering error is reduced.