Could you release the code for transforming the raw data to the forms in data folder

[haoyanbin918]

Could you please release the code or give advice for transforming the raw review file (with each line be the document text) to the used form in data folder. Many thanks.

[hsqmlzno1]

You can use NLTK to transform raw data into the hierarchical form.

[haoyanbin918]

I am sorry that I do not have knowledge of this hierarchical form. Could you please give some detailed info, like the extract name of this form, or some relative website that can help understand it. Because I have some new raw data to be done. It is much appreciated that you can send the transform codes. Thanks.

[hsqmlzno1]

Please tell me your email, I will send the code to you.

[haoyanbin918]

my email: [email protected] many thanks

[MansteinLiliang]

Hello，I am just wondering about the setting of the data. As described in this paper "We conduct the experiments on the Amazon reviews dataset (Blitzer, Dredze, and Pereira 2007)". However, this is referred to Multi-Domain Sentiment Dataset(https://www.cs.jhu.edu/~mdredze/datasets/sentiment/). I think your data is collected from Amazon product data(http://jmcauley.ucsd.edu/data/amazon/). Is that right?

[hsqmlzno1]

I just do some random sampling for the unprocessing version of the Amazon reviews dataset. You can also try the original small-scale data, it works well too.

[MansteinLiliang]

So, you mean the data is really from http://jmcauley.ucsd.edu/data/amazon/? I think that's not related the data (Blitzer, Dredze, and Pereira 2007). Aha, I also wanna your data transform codes like haoyanbincityu. my email: [email protected] many thanks

[MansteinLiliang]

And by the way, Dear zheng. I surveyed some work on sentiment transfer and notice that the paper "Transfer Learning for Sequences via Learning to Collocate" was just accepted by ICLR2019. They also compare with your model HATN and report poor results. I think it was unfair because the setting is quite different, they got the target labeled data to do finetune not unsupervise transfer. And in the open review, they just said they use your code... What's your opinion about their acceptance?

[hsqmlzno1]

1. The data is from http://jmcauley.ucsd.edu/data/amazon/.
2. The code will be send to you soon.
3. i think the author of the paper has reported uncorrected results in the rebuttal. I have also verified the hatn model in the small-scale dataset, the results still remain to be superior, which can outperform the AMN model by a large margin.

[hsqmlzno1]

Thanks for your notification. I have also responded to the rebuttal of this paper (https://openreview.net/forum?id=ByldlhAqYQ).

[MansteinLiliang]

Thanks for your generosity! I think it was chaos in the domain adaptation area of NLP. We are comparing different models for adaptation and find that even some mSDA based methods may work better. Can you also send me the code of AMN? I also want to compare with it for research. many thanks again~

[hsqmlzno1]

Sorry for the 1st question，the data is not from ucsd dataset.

[hsqmlzno1]

Sorry , I have no time to clean the AMN model at present. I think you can easily revise the hatn model to the amn model according to the AMN paper.

[MansteinLiliang]

OK, I will try to reimplemented it.

[eyalbd2]

Hi, Can you please send me your code which transforms raw data into hierarchical form? my email is : [email protected] many thanks

[hsqmlzno1]

Hi, Can you please send me your code which transforms raw data into hierarchical form? my email is : [email protected] many thanks

Sent, If the data and code are useful for your research, please be kindly to give us stars and cite our paper.

[eyalbd2]

Thanks, Will do..

בתאריך יום ג׳, 3 בספט׳ 2019, 17:55, מאת LI ZHENG ‏<[email protected]

:

Hi, Can you please send me your code which transforms raw data into hierarchical form? my email is : [email protected] many thanks

Sent, If the data and code are useful for your research, please be kindly to give us stars and cite our paper.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/hsqmlzno1/HATN/issues/5?email_source=notifications&email_token=AJDPXTSSG6XCKX6QI3UQ6DLQHZ3H3A5CNFSM4GAOLK72YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD5YPFPA#issuecomment-527495868, or mute the thread https://github.com/notifications/unsubscribe-auth/AJDPXTQIDESB3TU6ZGAOGQ3QHZ3H3ANCNFSM4GAOLK7Q .

[fzy0728]

Dear hsqmlzo1, Can you please send me your code which transforms raw data into hierarchical form? my email is : [email protected] thanks a lot

[hsqmlzno1]

Dear hsqmlzo1, Can you please send me your code which transforms raw data into hierarchical form? my email is : [email protected] thanks a lot

The code (preprocess_HAN.py, not clean version) has already been uploaded in this project page.