Inspiration

I was inspired to take on the Barclays problems statement as I was interested to explore how the different types of regression models would impact the model performance and to evaluate which model would be the most appropriate to use.

What it does

It provides an evaluation on the linear regression, ridge, decision tree and SVR model and its impact on the model performance based on the given dataset.

How we built it

The solution focuses on the 3 main steps when it comes to machine learning, data preparation, training and evaluation. Time was taken to analyze the dataset and better understand ways to refine the features to improve training of the model. Different models were also evaluated against each other since the results using linear regression was not ideal.

Challenges we ran into

Since the dataset seems to be non-linear, there is a need to understand and explore other types of models more suited for this purpose. Further understanding and research of the different models needed to be done, however, might not have fully grasp the full uses of the other different models and the advantages of each model over another.

Accomplishments that we're proud of

Learnt quite a few ways of looking at the data and formulating a way to run the different models in a single loop.

What we learned

Non-linear regression models like decision tree.

What's next for Barclays

To spend some time to try out this dataset using classification models instead.

Built With

Share this project:

Updates