Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Using evaluating environment for Itemprediction with further informations. #462

Open
aoberegg opened this issue Jan 5, 2016 · 4 comments

Comments

@aoberegg
Copy link

aoberegg commented Jan 5, 2016

Hey guys!

I'm currently working on a recommender that includes the impact of weather conditions on item choosing. Therefore I'd like to use the Evaluation of the ItemRecommender engine of MyMediaLite to test my recommender. For this reason I'd need a testing environment that calls a function like:

IList<Tuple<int, float>> Recommend(int user_id, IList current_weather, int n = -1,
ICollection ignore_items = null,
ICollection candidate_items = null);

this. So for giving a recommendation I'd need the current weather situation. Is that somehow possible with MyMediaLite?

Best regards,

Alexander Oberegger

@zenogantner
Copy link
Owner

Hi Alexander,

Modelling context (like weather) is currently not directly supported by
MyMediaLite.

It could be added, though.
If you have something nice, let me know.

One question, why is the current weather represented as a list?

By the way, LibFM does model context in a very nice way -- check it out:
http://www.libfm.org/

Cheers,
Z.

On Tue, Jan 5, 2016 at 5:40 PM, aoberegg [email protected] wrote:

Hey guys!

I'm currently working on a recommender that includes the impact of weather
conditions on item choosing Therefore I'd like to use the Evaluation of the
ItemRecommender engine of MyMediaLite to test my recommender For this
reason I'd need a testing environment that calls a function like:

IList> Recommend(int user_id, IList current_weather, int n = -1,
ICollection ignore_items = null,
ICollection candidate_items = null);

this So for giving a recommendation I'd need the current weather situation
Is that somehow possible with MyMediaLite?

Best regards,

Alexander Oberegger


Reply to this email directly or view it on GitHub
#462.

MyMediaLite Recommender System Library: http://mymedialite.net

@aoberegg
Copy link
Author

aoberegg commented Jan 5, 2016

Ok thanks for your reply.

Weather is in that case represented as a list of current temperature, precip intensity, windspeed,...

@zenogantner
Copy link
Owner

OK.

Then a nice, flexible API would be a list of fixed lenght containing
float-valued context parameters.

Would that make sense?

Z.

On Tue, Jan 5, 2016 at 6:07 PM, aoberegg [email protected] wrote:

Ok thanks for your reply.

Weather is in that case represented as a list of current temperature,
precip intensity, windspeed,...


Reply to this email directly or view it on GitHub
#462 (comment)
.

MyMediaLite Recommender System Library: http://mymedialite.net

@aoberegg
Copy link
Author

aoberegg commented Jan 5, 2016

That's exactly what I would need yes. In my case it would be:
[humidity, precip_intensity, temperature, cloud cover, moonphase, pressure, visibility, wind_speed]

On the other hand it would also be possible to hand the time and geolocation so that I can look up the weather data in a database. With that approach it's not fixed to weather data and maybe it would offer more flexibility for other contexts.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants