Installation requirements and usage instructions: README.md

This is an initial implementation of the Hashemi procedure, which leverages mathematical theories (ontologies) that are used to decipher complex data relationships. Through ontology reasoning, background information can be deduced from user-provided models (examples and counterexamples) to answer natural language questions (e.g., in financial applications) that cannot be answered by simple retrieval via a query language (i.e., SQL).

The script searches through an existing hierarchy of theories represented as a chain decomposition, and returns brackets from each chain, which enclose all matching theories. The search is done by checking consistency for all examples and inconsistency for all counterexamples using the Natural Language Toolkit API to access Prover9/Mace4.

• removing requirement of user to provide translation definitions
• further testing on bipartite incidence structure ontologies
• expanding the search to multiple hierarchies