Towards the fully automatic merging of lexical resources: A step forward