Comments (4)
negative numbers are documented to indicate "no particular size requested" however, it is convention (which I want to rely on) that -1 should indicate "rank all items"
Hmm. I sympathize. However, I am very hesitant to increase the guarantees implementations are required to maintain.
It is not made clear that implementations are allowed to return fewer than the requested n items (such as if the item recommender enforces a constraint and cannot or will not return more items)
Definitely need to clear that up.
from lenskit.
Quoting you on PR997 on the subject of -1 == all:
It does. Actually, any negative value is supposed to be all.
Currently there is no "correct" way to ask for an absolute ranking that I can think of. I consider this a problem.
from lenskit.
Currently there is no "correct" way to ask for an absolute ranking that I can think of. I consider this a problem.
I do not automatically disagree.
In conversation with Qian, we've discussed the relationship of retrieval, ranking, and recommendation.
What would you think about specifying that -1 means 'recommend as many as possible'? Leaves the door open for some hypothetical recommender with a strong belief that 10 items are all you need, but much more strongly indicates that any recommender without such a belief should generate as long a list as it can.
from lenskit.
I like this resolution. I agree that not all recommenders can produce a complete list (I just built one that will provide a truncated list due to constraints)
So the changes:
- Document that recommenders are not guaranteed to return n items
- Document that negative means "as many as possible" which often implies "all", but not always.
from lenskit.
Related Issues (20)
- Support query/runtime data in train-test evaluation
- Support emitting query data from crossfolder
- Support Bellogin's evaluation methods
- Bad import detection is broken HOT 1
- Add option for evaluation to continue after a failed job
- Add setting to restrict parallel evaluations
- Create general-purpose score/recommend/rank APIs
- which algorithm does use the item feature(e.g. some features in ML-100k's u.item files) in Lenskit HOT 3
- Support frequency-based recommendation
- Implement hit rate metric
- Rating summary is asking for Rating entities for implicit feedback data HOT 5
- Isolated train-test sets do not work correctly
- Implement new-style JDBC DAO HOT 2
- Write eval results to a database
- Adding Parameter to IntelliJ IDEA HOT 1
- Investigate switching to LA4J HOT 3
- Remove SparseVector
- Create general-purpose Lucene-based recommender HOT 1
- Support count attributes in popularity statistics
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from lenskit.