some questions about the dataset
Closed this issue · 4 comments
in the sample.csv, the values of Egc and Egb seem to be reversed:
sample.csv:
paper:
https://www.polymergenome.org :
and by the way, how to get the full dataset? the website https://khazana.gatech.edu/ seems have ssl and nginx 400 mistake, or I should get some keywords to get the dataset
Hi @Don12138, thank you for bringing both issues to my attention. First, let me clear up any confusion. The polyGNN models reported in the companion paper are not the same as the models on Polymer Genome. So, your first question is really a Polymer Genome (PG) question, not a polyGNN question. That being said, I did route your question about PG to the people responsible for maintaining it. And indeed there was a bug! Thank you for finding it. It has been fixed.
Regarding Khazana, I also am having trouble accessing it. I will look into the issue and report back.
Hi @Don12138, thank you for bringing both issues to my attention. First, let me clear up any confusion. The polyGNN models reported in the companion paper are not the same as the models on Polymer Genome. So, your first question is really a Polymer Genome (PG) question, not a polyGNN question. That being said, I did route your question about PG to the people responsible for maintaining it. And indeed there was a bug! Thank you for finding it. It has been fixed.
Regarding Khazana, I also am having trouble accessing it. I will look into the issue and report back.
Thank for your response to my previous message !!! I truly appreciate your work in the paper and your assistance!!!
I can browse the website of Khazana by ignoring the ssl mistake, and how do I find the full dataset of this paper? I didn't find your paper in the Data Repository (Is it because this paper is newly published？) or maybe I made a mistake in my operation? The dataset would greatly benefit my ongoing research. I am excited about the opportunity to further analyze and build upon your work, and I am committed to ensuring appropriate use and proper attribution of the data.
@Don12138 , the data contained in the paper "Polymer Informatics with Multi-Task Learning" is a superset of the data used in my work. You should be able to download data corresponding to that paper on Khazana. Let me know if this suits your use case.
I am closing this for now. Feel free to reopen if you need.