V
V
Vitaly Vitaly2018-02-02 13:54:22
Parsing
Vitaly Vitaly, 2018-02-02 13:54:22

How to sync products?

Good afternoon.
I think it is as easy as possible to implement the search for identical products when parsing.
Let's say there is a product:
1 Source: AMD RYZEN 3 1300X, OEM
2 Source: AMD Ryzen 3 1300X, SocketAM4 OEM The
product is the same, but the names are different. What algorithm can be used to more accurately determine that these lines (product) are similar (same)?
Tried Jaro-Winkler, did not help.

Answer the question

In order to leave comments, you need to log in

1 answer(s)
L
Leonid Nikolaev, 2018-02-02
@nikonor

we translate all words into upper, assign a degree of 2ki to each word, depending on the weight (number of words in the sample), sum up all the words.
when searching, we do and to check if there are any words in the title.
then somehow we count the weight of the hits and that's it.
I did it somehow. worked

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question