10 different ways to improve a Machine Learning Model

After building a successful machine learning model and deploying to production. Maybe after few months you gathered a new set of training data and you want to improve the metrics of your model. In this post, I have mentioned 10 different ways on how you can try different ones to see which does a better job for you.

  1. Collect more training data
  2. Try collect more diverse training data that closely represents the real world data
  3. Try running optimisation algorithms longer — gradient descent.