0
$\begingroup$

Currently I have a list of Books which I need to compare to a different list. For example in my local list I have:

 - Harry Potter and the Philosopher's Stone | J.K. Rowling
 - Harry Potter and the Deathly Hallows | J.K. Rowling
 - Harry Potter and the Goblet of Fire | J.K. Rowling
 - Harry Potter and the Prisoner of Azkaban | J.K. Rowling

When I compare I end up having a simple operation:

Book Title + Author (local list) vs [Book Title + Author] (remote list)

I'm currently using Levenshtein distance, based on that I can find out if Book is in remote system or not.

In this case if Book1 vs Book2 result score exceeds X, I assume is the same Book.

levenshtein(Book1 Title + Author, Book2 Title + Author)

I read about Word Embeddings and before I start coding want to check if converting my text information to embeddings may improve my results.

$\endgroup$
  • $\begingroup$ I do not think you need an embedding; just simple string similarity in title and author. Your database should be able to do this. $\endgroup$ – Emre Aug 17 '17 at 18:22
1
$\begingroup$

You do not need to use Word Embedding to solve this problem.

You are performing set membership check. In this problem, each book title and author combination is an element. You can create a hash for each element and perform constant time lookup. This assumes you have consistent normalization for each data source.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.