Cultural recommendations: the experience of the Moscow hackathon

At the end of last year, I took part in the Digital Transformation Leaders hackathon with the support of the Moscow Government. We solved the problem from the Department of Culture - a recommendation system for its services, that is, books in libraries, as well as circles and events in cultural centers. A special piquancy is that for one of these services it was necessary to recommend others. Our decision came in only second place, but making it was informative.





How our system works - and an example of a piece of its issue
How our system works - and an example of a piece of its issue

About hackathon

, 10 , , - . . - , . , . , - . , , . API , - , Excel, , , . , .





: ML , , . Artificial Intelligentsia. , : , - , . , . , , , . , , , , - .





, . , , , . , , . : , , , - , - , , . - , . , , , . : , , . " " , " , 1" - .





, . -, : , . -, : , " " " ", " 2" - . , , -, . ( ), .





content-based , : , , . , fasttext- . : fasttext - , word2vec, , . , . , . , , , . , , , , . , . , .





: / , , … , , , . . , , , , . , , .





- , , . , - . : , , - , , , , . . - , , .





, : ( NLP), . , collaborative filtering , . , . , , .





!








All Articles