Ml 10 develop a baseline model

[promatty]

No description provided.

[justin-phxm]

• Your Analysis looks good. I would throw it into ChatGPT to clean it up though.
  There's a few things we can do to improve it.

1. Clean the data
2. use different algorithms
3. consider using a meta-model
4. train on different subsets of the training data (cross validation)
5. ask chatGPT lmao

[justin-phxm]

Even after reevaulation of performance, there should be no negative minimum distance:

Great summary on the various regression algorithms. My prediction is that Random Forest is going to perform the best.

Can you also include an interpretation of the data at the end... What does this data mean in the real world? How can we make data driven decisions from this? For example, maybe an explanation of how this model can predict the distance of a given set of data with +-10m accuracy.

You can honestly remove this one section as it is not very relevant. Instead, add a section at the top about how we want a regression algorithm instead of a classification model.

You can also turn this into a function as it is duplicated multiple times.

I believe it is standard practice to include all the imports at the beginning of the file. Please move this to the top.

Look into Hyperparameter tuning... We won't implement this in the current ticket but we will definitely use it in the future to evaluate our models. It's not nearly as important in LinearRegression models but will be huge when it comes to decisionTrees/Forests.

[justin-phxm]

I was hoping that removing the unrealistic negative values would decrease MAE but it's fine. This can happen sometimes