search

UBC-MDS / DSCI_522_Group_18

Group project repo for DSCI 522
MIT License
0 stars 5 forks source link

Peer review comments #37

Open huan-ds opened 3 years ago

huan-ds commented 3 years ago

Hi,

First, the analysis is clear and easy to understand.

Then, I have a couple of suggestions regarding the report.

  1. For "Table 1. Cross validate scores of model performance on train data.", you interpreted the process of choosing a model. However, it is not necessary for the final report. It is ok only providing the best model.

  2. For the model performance table, I feel that you gave a lot of information but not enough interpretation. If you can add more interpretation on the metric (f1, recall, etc.), it would be more helpful to the reader, especially for people without a data background. You can also make a tweak on the performance table to make it simpler.

  3. For Figures 1 and 2, you use the encode 0 and 1 as x labels. It would be more clear to relabel to >50k and <50k.

Last, for the team communication, I see most PR is merged by the person who created. It is better to review by other team members.

Thank you!

AishwaryaGopal12 commented 3 years ago

@huan-ds Thank you so much for the valuable feedback. @fei-chang @yhchen20 I added basic definitions for precision and recall based on the confusion matrix in the report.