Applications of corpus-based semantic similarity and word segmentation to database schema matching