ActiveLearningProj

query_strategies modify the active learning strategies to serve a streaming pool

How to run this project experiments

simple_linreg_exp.py is the experiment for linear regression. Specify parameters as follows:

-n := num_rounds (default = 10)
-c := num_coeffs (default = 5)
-s := initial_sample_sz (default = 20)
-p := pool_sz (default = 1000)
-b := budget (default = 10)
-i := iter_per_algo (default = 10)
-v := verbose mode (default = false)
-m := flag to turn on measurement error (default=false)

Example run command:

python3 simple_linreg_exp.py -n 1000 -c 2 -b 1 -p 1000 -i 25

logreg_experiment.py recreates our results for the logistic regression experiement. Example run command:

python3 logreg_experiment.py

multivar-exp.py is the experiment for multiple linear regression. Specify parameters as follows:

-c := num_coeffs (default = 5)
-m := flag to turn on measurement error (default=false)
-v := verbose mode (default = false)

Example run command:

python3 multivar-exp.py -c 5

---

Developer's Corner:

Contributing Summary

• To format code, run black .
• To lint code, run flake8 .
• To install requirements, run pip install -r requirements.txt

Contributing

1. Fork this Repo
2. Clone the Repo onto your computer -- You may need to setup an SSH Key on your device.

• Run pip install -r requirements.txt to get all the packages you need.

3. Create a branch (git checkout -b new-feature)
4. Make Changes
5. Run necessary quality assurance tools

• Formatter, Linter, and Test your code.

6. Add your changes (git commit -am "Commit Message" or git add <whatever files/folders you want to add> followed by git commit -m "Commit Message")
7. Push your changes to the repo (git push origin new-feature)
8. Create a pull request

---

Code Quality Tools

black formatter automatically formats code

1. Run pip install black to get the package.
2. After making changes, run black ./.

flake8 lints code

Notifies you if any code is not currently up to Python standards.

1. Run pip install flake8 to get the package.
2. After making changes, run flake8.

---