Search for semantic similarity and relationships between different words

I am working on a project that involves calculating the relationship between different concepts.

Examples: "landing" - related to: "arrive, come" and "land", "phone" - related to: "electronic, electrical" and "talk, communicate", "movie" - related to: "to move, to change "and" picture; image ", etc.

I checked the structure of WordNet and WordNet :: Perl's similarity module, but in some cases they are unusable, especially in cases where I need to connect nouns with verbs (for example, the noun "dog" and the verb "to" bark). Moreover, WordNet organizes all the syntaxes in a hierarchical structure (eg: computer -> machine -> device -> devices) and this is not what I really need in my project (of course I could be wrong).

The question is which structure or database, which approach should be used to solve the problem? Where to begin?

It would be very helpful if you could give me some advice.


source to share

1 answer

Word2vec attachments might do a good job for your task. This gives a much softer concept of similarity than wordnet. You can get some pretrained attachments linked here .



All Articles