beghbali / factsderivation Goto Github PK

Semantical facts derivation using grammatical parser and first order predicate logic resolution

see the docs folder for more discussion and graphs

example:

given predicates in plain language(scholarly articles):

TGF Beta-1 causes increase in PH level
increase in PH level contributes to development of cancer

it can be derived that:

TFG Beta-1 contributes to development of cancer

various parts of the system:

Abstracts: The abstracts that we are extracting information from (currently 100).

Chart Parser: It replaces compound nouns with their abbreviation and if an abbreviation is not specified, it creates an auxiliary abbreviation. It then does the parsing of the sentence to create a tree of the sentence structure.

Mental Representation: Given the sentence structure from Chart Parser, it derives the important parts of the sentence out and creates databases of relations.

Fact Derivation: Given the Ontology database it outputs 1) gammatical facts 2) derived facts

Lexicon: The dictionaries and thesauri in our case, WordNet, that provides word definitions, synonyms, antonyms, etc…

Ontologies Database: A database of ontology, relation of noun(s) by verb(s). Expands as new sentence information comes in.

NOTE: this project is very old and will not compile with latest java compilers. I hope to get arount to bridging it to the modern world once I have free cycles.