Statistical Models for Short Text Clustering