How to create knowledge graphs on unstructured data?
Take a look at:
@bratanic_tomaz thanks for sharing the articles . but in this article all these done using old approach like Spacy. do you have any leads to do these things with langchain or llma_index or any other approach
I have an article that uses GPT-4 as well.
We've also added today the option to create a knowledge graph using LangChain:
Thanks for sharing the information @bratanic_tomaz .
Few more questions I have.
How to retrieve the information which is not presented in knowledge graph nodes or relations ?
Like we have complete article but while creating article we might miss few information and user ask question on that is there any way to receive that information?
@bratanic_tomaz
Please I have a question regarding this article. I see that relationships (verbs) are represented in the graph as nodes, not edges.
I want to do the same thing with "relations/verbs" because I want them to inherit higher abstract categories from the ontology of verbs.
That is, the verb “to provide” is from the category of “to supply,” which is from the category of motion verbs that can be represented as “to do.”
Which option is better in terms of implementation, in your opinion: to make the hierarchical categories associated with the “verb” using the attributes, or to make the verb (relationship) a node so that it is easy to inherit from the ontological categories?