R
R
Ruslan2017-11-13 11:56:59
Database
Ruslan, 2017-11-13 11:56:59

How to search for similar products when parsing several online stores?

You need to parse goods from 30 online stores.
But the same products need to be combined, that is, there is a template product and a product from their store.
If the database already has a product template, then link the product to it, if not, then create a new product template.
That is, the Philips 223V5LSB Monitor (or MAUNFELD Ancona 90 stainless steel Chimney Hood, etc.) in the templates must be one, and all goods from stores are tied to it.
There are a lot of goods, the parser takes the name, picture, characteristic, description, articles.
In fact, it will turn out like in the Yandex market there is a product template and offers from different stores.
Maybe there are ready-made algorithms, libraries for comparison, identifying the percentage of similarity, even more interested in the percentage of similarity.
Or tell me which way to dig for the best comparison of goods.

Answer the question

In order to leave comments, you need to log in

1 answer(s)
I
Ilya, 2017-11-13
@mitrm

In fact, the problem here is only in the search for similarity:
Actually, that's the whole algorithm. But again, such typos must be taken into account, or look in the direction of inaccurate string comparisons.

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question