MyMediaLite

I still see it although it is not that significant. I used larger data training data: 48912 users,

This may be interesting for some of you: https://tirania.org/blog/archive/2019/Apr-22.html https://

Hello, I am not sure what you do not understand. Please have a look at the documentation http://

Not implemented for the command-line tool. You'd need to code it yourself, either standalone, or

Hi, just have a look at the classes in the ItemRecommendation namespace. There are plenty of pointers

may you share a link to this description slides? do you know the algorithm for item prediction from

http://tirania.org/blog/archive/2018/Apr-11.html This may be interesting for MML users. Anyone

Hello Evgenia, Which if the 5 options for item selection (see output of --help) did you choose for

Hello, I fear there is no simple, straightforward solution. We would need to look at the algorithm

Thank you for the help. I've already submitted my master's thesis, for which I modified

Initialization is random, but you set the random seed to get repeatability (eg for debugging). Sorry

I got it figured out. 1. I needed to use absolute paths (I thought I had done that) and 2. My input

I did use --recommender-options=correlation=Pearson On Thursday, June 29, 2017 at 12:20:18 PM UTC-4,

Thank you for your fast reply! much appreciated! On Jun 29, 2017, at 10:02 AM, Zeno Gantner <zeno.

Hi Julien: I wonder how do I call this function? Can you give me an example of the command line if I

I've already asked this question before on this forum (roughly the same intuition as with WRMF

Am 04.04.2017 12:08 nachm. schrieb "Djoels" <juv...@gmail.com>: I tried feeding the

Incidentally, I happened to stumble upon a recent (2017-01) implementation by R. Devooght: https://

I need to do some more testing to be really sure, but got it to compile without errors. Fix consisted

The commandline arguments and logs are all in the ZIP file that I added to this message. Op maandag

Hi Julien, I'd also suggest to try those toolkits (and tell us about your experiences). If you

Feel free to implement and contribute them via clean, minimalistic (1 per feature) pull requests on

Hi Marco, Please try the --repeated-items command line option. I hope that helps. Z. Am 13.03.2017 7:

On Sun, Feb 12, 2017 at 4:38 PM, Djoels <juv...@gmail.com> wrote: On Sun, Feb 12, 2017 at 3:51

Am 01.02.2017 3:03 nachm. schrieb "Wang Dawei" <weeo...@gmail.com>: Thank you so

Hi, Sorry for the late reply. If you want to report metrics, you should use either --test-file or --

Sorry for the late follow-up -- if you are still interested in this: What kind of RMSE do you see for

Hello list, After releasing MyMediaLite 3.12 (during the next days), I intend to use C# 6.0 language

Hi Julien, You just have too many users and items for MyMediaLite's kNN implementation. The kNN

Thanks. Being able to evaluate rating-predictors results with top-n metrics is helpful for my case. I