Image Created by Author through DallE

Practical Machine Learning: Part 1

This is a series of articles to help other Data Scientists based on my learning

“How do I improve my model?”

This is the most common question I get asked by other data scientists. This question, to me, is the equivalent of “What should I do now?”

There is no one right answer to this question. Just like for the latter question, I would start answering the former with “It depends.”

The truth is, it really depends on multiple considerations, for example:

1. What do you actually mean by accuracy here?
2. Why is improving accuracy important? Are you sacrificing something else just to improve accuracy?
3. What does improving accuracy mean to your end user?
4. Can you afford to add more training data to improve accuracy?

In this post, I wanted to provide a logical Decision Tree to help you decide on what step to take to improve your model accuracy. I will limit the Part 1 to just the ML models.

Troubleshooting an ML Model

1 Is it a Regression Model or a Classification Model