GSoC 2018- bug solving and idea understanding

[Abhijeet Panda]

Hi everyone,

This is Abhijeet Panda. I have been looking through the code and the documentation from the last 3 days and I am interested in participating in GSoC 2018 under statsmodels. For my first PR, I looked through the issues labeled as 'good as first PR' and I want to solve this issue:

https://github.com/statsmodels/statsmodels/issues/3584

Can someone help me out on how can I proceed forward? Or If someone suggests a better alternative to the issue, I would be grateful.

I wish to contribute to automatic forecasting as my GSoC project. I also need some help as to which paper or which particular algorithm to implement. I have gone through R J hyndman's paper on automatic forecasting through exponential smoothing methods.

Thank You,

Abhijeet Panda

[Chad Fulton]

Hi Abhijeet,

I'm glad to hear you're interested in a GSOC with Statsmodels, and the automatic forecasting project in particular. I will try to post more details soon about my thoughts for how this project might go.

I'm not too familiar with that particular part of the code, but  you might get the best response if you add a comment directly on the Github issue saying that you're interested.

Glad to hear you've read Hyndman's paper - that is indeed the best place to start. Now you will want to think about you feel is important and feasible to complete in a summer. My guess is that we will already have most of the required models (e.g. sarimax and exponential smoothing) and so the project would be about constructing the automatic forecasting infrastructure.

Best,

Chad

[Abhijeet Panda]

Hi Chad,

Thank you for the response.

I'll be waiting for your thoughts on how to proceed with this project.

  

Regarding the issue, that was a simple issue with numpy which got resolved. Could you please suggest me a bug which would be related to the automatic forecasting part of the code?

I have started doing some research on how we could proceed with the project. Since the sarimax and exponential smoothing models are already available, I'll be playing around with that part of the code to better understand how it works. Till now I was using the forecast package of R which included automatic forecasting.

Could you suggest some other places where I could look into so that I could have a clear understanding about how we could do the implementation?

Thank You,

Abhijeet