Text this: Multi-Data Aspects of Protein Similarity with a Learning Technique to Identify Drug-Disease Associations