search

TUHH-PROJECT / ISM_ML_Project

0 stars 0 forks source link

Feature extraction third iteration #3

Closed asingh1098 closed 2 years ago

asingh1098 commented 2 years ago

As of now, we have extracted features from the dataset but as discussed in our last meeting, on 03/12/2021, we have collectively decided to go for another round so as to extract more (better) features from the dataset which would have a significant positive impact on the final prediction result.

asingh1098 commented 2 years ago

@GiulianoTaccogna I'd strongly suggest you put all your findings (like articles, research papers, tutorials, etc.) regarding this issue/task here by putting the links of the material in the comments section so that all the reading materials related to this issue will be right here which would make it easier for you to later refer and also would be beneficial for all of us to get some insight and knowledge about the methodology you'd use.

Lemme know what you think about it 😄

PS. make sure to work only on your branch and NOT ON the main branch. I've also created a milestone and put this issue along with all other issues in that milestone.

asingh1098 commented 2 years ago

@GiulianoTaccogna why did you close this issue? weren't you and Gursel supposed to do it?

GiulianoTaccogna commented 2 years ago

@shrey112710 I finished the task, sorry should have said that.

asingh1098 commented 2 years ago

@GiulianoTaccogna Oh wow! that's great. The convention which most people follow on GitHub is that one would work on a specific branch and then create a PR and link that PR to the issue which that PR addresses. After that, the PR is reviewed by another person (typically of the same team) and once the PR is approved by the reviewer then the PR is merged with master. Once the PR is merged with the master then the issue, linked to that PR, automatically closes.

click to expand Screenshot 2021-12-10 at 00 44 10

but the thing is that the PR(/pull/10) to address this issue needs to be first linked with this issue and then would need to be reviewed and when it is approved then it needs to be merged. In this way, there'll be no need for manually closing the issue, and also since with this approach the master will always have the latest changes therefore we won't have the overhead of hoping to different branches for training/testing specific models or just to run the code.

Hence I'd strongly advocate working this way. Please, Lemme, know your feedback 🙌