Using Word2Vec or better yet, what it was based on, https://www.kaggle.com/c/word2vec-nlp-tutorial/forums/t/1234... might make this possible and would be a good approach toward achieving future goals.

Take a look at text sentiment analyzers and processors that employ advanced machine learning algorithms like this too, http://52.10.12.34/TuataraSum