database - How to relate word concordances and docuements using graph data base -

i saw tutorials referring importing data {neo4j, titan,...} graph dbs. have questions how apply our academic project.

in case have set of documents in pdf , microsoft word formats (which not important @ moment) want relate global topic. in case, easy visualize solution using {neo4j, titan,...}; however, there exists case think little more special. concepts present in documents must related in graph. in case have pinpointed such concepts, represented multiword terms (like instance "artificial neural netwok"), , relate them within document set of documents, possibility of making query context in document or documents in queried term used (concordances).

the last task (concordances) made using relational db, want centralize tasks unique database, think must graph db {neo4j, titan,...}.

we hope little orientation how think can adapt our problem {neo4j, titan,...} graph db. maybe documentation similar examples (if there exists), general view possible ways of data structure importing... know.

well hope not ambiguous. thank in advance.

the question still bit broad, i'll try give quick answer i've understood.

you can start in easy way "schema" (or structure of graph):

• create node multiword term
• create node document
• everytime there's concordance can create link between 2 respective ends (term, document).

i recommend normalize data in single format before proceeding processing/importing: generic format csv, can have @ graphml (which suported graphdbs) or graphson.

if want browse "schema" graphs, have @ neo4j gist collection: it's wide collection various topics can use inspire structure.