Word classification based on combined measures of distributional and semantic similarity