Add a Bayesian Regression Model to an existing Machine Learning Workflow

The summary of this task is to add a new model (Bayesian Regression) to my current code. Background: I have a Python machine learning model script which is split into two parts: - Training - Deployment Each is an end to end solution that takes a csv as an input and then is customizable as to which model and settings to use: Model type: - Regression - Classification - Knn Regression - Knn Classification Scaling options: - Minmax - Standard - Robust - MaxAbsolute Feature selection: - PCA - LASSo - Random forest - LightGBM When you run the training script I am able to choose between the above options and the code has inbuilt functions such as one hot encoding and label encoding, train test split etc. At the end of training it produces a folder with the model details which are call upon in the deployment. Task: What I would like added is a new model that produces posterior probability (a Bayesian type model) whereby the end results contains more detail around the predictions. Eg the model predicts a price of $3 but I want it to also contain information like a Gaussian distribution to be able to do things like “95% confident the price is greater than $2.50”. In this model i would also like the option of doing a time based sample rather than random sampling (e.g training on the first 66% of the data and testing on the last 34%). I would like it to work within the existing framework for both the training and deployment Attached is a sample code for training and deployment as well as the training data and unseen data for furture predictions. With the training data I have three types of columns (Information Columns, Target columns and feature columns which are defined at the start of the code). Any further questions please let me know!