Open questions - Githubissues

BioinfoNet / Data-mining

Data mining to discover trends in Open Science in Kenya

5 stars 14 forks source link

Open questions #19

Open Shuyib opened 4 years ago

Shuyib commented 4 years ago

Hi,

I've just seen that the repo has been updated to reflect changes till now. I have several questions on the way forward.

Are we going to make an application for this? What if someone wants to do this in future? I'm thinking ipywidgets could help.
Are we going to do use Natural language Processing to look at the main topics being posted on the papers?

What do you think? @kipkurui

Shuyib commented 4 years ago

I'll do it anyway. When I get the time.

kipkurui commented 4 years ago

Hi @Shuyib that is a wonderful Idea. It would be fun to have a widget that would provide real-time visualisation. Is that what you had in mind. For natural language processing, which kind of question can we prioritise?

Are we analysing the abstract only? It can be tricky to get the full text, especially or closed access journals.

Shuyib commented 4 years ago

Yes, we'll need to wrap everything in a while loop at the querying and returning the results + timing in between. I think it would be better to make one version where the user can just query what they want. I can wrap that with ipywidgets.

For NLP, I'll use Non-negative matrix factorization which is an unsupervised learning method. We can adjust the K that is, the number of clusters to potentially find out the topics in the abstracts only. Unfortunately. Just a by the way, we could use this in the next session mybinder

How does that sound?