MS specialization project for Cornell-Technion Dual degree program in Information Systems by Noel Konagai.
We use a Python 3.7 virtual environment in Conda. To set this up:
conda create -n <your_env> python=3.7 anacondaTo activate:
source activate <your_env>Install the requirements from the repo:
pip install -r requirements.txtTo deactivate:
deactivateanalyze.py lets you pass in custom arguments to test certain parts of chkpt3_analysis.py.
| flag | desc | required? | type | choices | used in |
|---|---|---|---|---|---|
| -f | which function to run | no | str | vary_char, vary_time | N/A |
| -t | what time difference to use (mins) | no | multiple int | N/A | vary_time |
| -c | what minimum character length to use | no | multiple int | N/A | vary_char |
| -r | regression to run on all groups together or separate | no | str | all, separate | N/A |
| -d | what data type to use | no | str | vary_char, vary_time | N/A |
So, for example if you want to dissect the dataframe with specific character cutoffs:
python analyze.py -f vary_char -c 60 70 80These create individual CSV files either under data/chkpt3/conversation/time or ~/char. Once you have created the CSV files, you can run a regression to see how well the model predicts response rates. To do so:
python analyze.py -r separate -d vary_char
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent