Similarity Search

KNIME Analytics Platform
1

Hello all,

I'm having trouble using Similarity Search. It is, instead of crossing two different columns from different tables, crossing the same column with the same column, resulting in a column full of distance 0.

Attached is my workflow. My "non-watched movies per user" has its category bit vector for each movie, and so has my "User ID x Rating >= 4". What I need is to get the category bitvector distance between each movie the user hasn't watched yet with the movies with Ratings >= 4, so I can recommend him the similar movies it gives me.

 

Thank you in advance,

mhnjqm

 

2

(You forget to attach your workflow.)

3

Sorry! Size was over 5mb so it didn't upload. Here it is:

https://drive.google.com/file/d/0B16Lgcp22M9-RWl0SXVLNk5iVGs/view?usp=sharing

 

Thank you,

mhnjqm

4

Not having looked at your workflow I would guess that this is as intended.

What you need to do is to loop a single molecule at a time versus your comparison table.

I believe in some of the community workflows (example server) there is somewhere an example that does just that. I'll see if I can dig it out later.

5