Optimized item selection

[takojunior]

• This PR is to add the public dataset used in CPAIOR'21 paper. The data is ready to be used to run Text-based Selection for applying optimized item selection framework proposed in the paper.
• The data includes GoodReads book reviews and MovieLens for movie recommendations. A README file is provided to introduce the data and show a usage example.

[takojunior]

Hi @skadio @bkleyn , I think I addressed all the current comments so far. Let me know if there would be further comments. If the PR looks good to go, I can merge the branch today or tomorrow.

[skadio]

@takojunior to re-iterate, we are NOT merging this branch, yes?

[takojunior]

@takojunior to re-iterate, we are NOT merging this branch, yes?

Got it. Sorry, I thought we would merge the branch with this ISP data.

I can withdraw the PR if it looks good for now and keep the branch.

[skadio]

@takojunior we need to clean this up though.. let me push a quick commit

[takojunior]

Hi @skadio @bkleyn , I made the following updates to the data repo per discussion:

• Update readme by adding citation, and a few advanced usage examples, i.e. using other selection methods by changing optimization_method and cost_metric configurations
• At the beginning mention the data is extracted from original public source and used in our research study, so ask users to refrain from distributing
• Uploaded the datasets to huggingface under skadio/optimized_item_selection, and updated the usage example by loading the data from there
• After the PR looks good, I think we can close it, and then we either keep this branch for further edits/reviews or delete and make changes on huggingface.