Build or Collect Analogy test data set

[sehsanm]

See :
Persian Word Embedding Evaluation Benchmarks
They have proposed analogy task. We can simply use their dataset.
Otherwise (if not available) or copyright issue we can build one ourselves. The data set must include both semantic and synthetic analogies. Also it must contain categories.

The file should be stored in data/analogy folder

[maryambiabani]

Hi,
These are some analogy test set:
word-analogy.zip

[sehsanm]

Hi please put the file in repo (/data/analogy) and create a pull request