Statistical Models for Mining Chinese Text