Could you please provide an example of Learning-to-Rank (LTR) Model?

[linh47]

Thank you for your implementation.
I want to know how to run Personalized Re-ranking Model based on your recsys?
Could you please provide a Google Colab version example for me ?Thank you a lot!!!
It would be better if the code could have detailed comments.

[linh47]

我本想根据文档写拓展的，但是文档里空空如也。
想请求你完善一下排序学习的示例代码，拜托了。

[p768lwy3]

Thank you for your proposal and I am very sorry that there are too many missings in the documentation and repository since the framework is still in the first stage and I have not decided the interface yet, but I have created an example here:

https://github.com/p768lwy3/torecsys/blob/master/example/notebook/Example-1-Training-a-Bayesian-Persoanlized-Ranking-Matrix-Factorization-in-torecsys.ipynb

It would be great if you find any other issues or opinions since I haven't check it well actually.

[linh47]

Thank you for your reply!!!You are so kind!!!
Could you please also provide an example of this?

Personalized Re-ranking Model | Changhua Pei et al, 2019.
Personalized Re-ranking for Recommendation

[linh47]

I tried to install torecsys which version was 0.0.5 from pip ,and I found there was no model about this PRM( Personalized Re-ranking Model | Changhua Pei et al, 2019.)。
Maybe torecsys in PYPI needs to be updated? I also couldn't find the DSIN and other new models you implemented in this repository.
我从pip下载的torecsys里面没有您最新实现的一些模型，比如DSIN、PRM等等。也许是上面的包没有更新？感谢您的回复，谢谢！

[p768lwy3]

I tried to install torecsys which version was 0.0.5 from pip ,and I found there was no model about this PRM( Personalized Re-ranking Model | Changhua Pei et al, 2019.)。
Maybe torecsys in PYPI needs to be updated? I also couldn't find the DSIN and other new models you implemented in this repository.
我从pip下载的torecsys里面没有您最新实现的一些模型，比如DSIN、PRM等等。也许是上面的包没有更新？感谢您的回复，谢谢！

I am very sorry that the package on pypi is an old version for a long time since it was a test for me to try to upload a package to pypi :). I have planned to update it after more development (however I have no time to get the work done :(.

If you want to play the current version, you may need to clone the repository and import it locally. I am sorry about this but I have no plan to update the wheel at this stage.

[p768lwy3]

Thank you for your reply!!!You are so kind!!!
Could you please also provide an example of this?

Personalized Re-ranking Model | Changhua Pei et al, 2019.
Personalized Re-ranking for Recommendation

Regarding personal_reranking, it is still a long way from completion, because the design of it is quite special.

The main reason is that it is a reranking algorithm. In addition, it requires a pre-trained model to embed products in the input step, so it is very different from other algorithms that are temporarily implemented. Therefore, I still have no time to finish it. So far, I'm very sorry that only the part of the calculation is completed.

As for whether it is possible to complete it in the near future, I have not yet given a clear reply since this project is done in my spare time.