Use Jax linear regression to predict on probabilities of software defects

I really enjoy working on the Kaggle community competitions because they enable me to try out new techniques in a competitive environment. In this particular competition, which is season 3 episode 23, I was making predictions on the probability of a software defect.

Python’s machine learning library, sklearn, has functionality to make predictions on probabilities on some of its models with the predict_proba() method. Not every algorithm and not every library has predict_proba, however, so I had to improvise.I read on stackoverflow that to predict on the probability is a regression, so I decided to give it a try. Therefore, instead of using the logistic regression algorithm to solve this problem, I used a linear regression algorithm.

I wrote the program in Kaggle’s free online Jupyter Notebook, which is stored in my account for that platform.

After I created the Jupyter Notebook, I imported the libraries I would need to execute the program, being:-

1. Pandas to create dataframes and process data,
2. Os to go into the operating system to retrieve the files used in the program,
3. Sklearn to provide machine learning functionality,
4. Jax to create the linear regression model,
5. Matplotlib to visualise the data, and
6. Seaborn to statistically visualise the data.