Machine learning model

#1

I have a doubt, yesterday I read the comments of a winner in Kaggle, he said that he has nearly trained about 200 models. I doubt my understanding of the term Machine learning model. Correct me if I am wrong, In a machine learning algorithm we tweak a hyperparameter and we predict the output it is considered as a model, so by this way tweaking 200 times trying to tweak different parameters we get 200 models right?

#2

Good questions!.

  1. It can also be addition or deletion of parameters.
  2. Using different ML algorithms to make model.
  3. It can also be setting up the right Hyperparameters and parameters!
  4. Doing the PCA or taking the relevant and most impacting columns.
    and many more are required to get a best model and this may take more number of iterations and at last a model is a file of weights numbers and the architecture will be stored.