Which option to chose to set parameters to LightGBMClassifier()?

[Nitinsiwach]

There are three ways to set pameters to a LightGBMClassifier()

1. In the class constructor itself
2. Using the params argument in the fit method
3. Using the setParams method available to LightGBMClassifier()

I can see why we have to endow the classifier with a setParams method but, can you please explain why we also have params argument available in the fit method.

Any use-case I can think of turns out to be a use-case that we need setParams for. So, why params? When should I choose it over setParams?

AB#1884830

[imatiach-msft]

@Nitinsiwach I believe all of these options are on the standard spark ml learners as well. You would normally set in the constructor, but these other methods are available so it can integrate with the spark ml ecosystem. Hope that helps resolve the question.

[ppruthi]

@Nitinsiwach -- if you are satisfied with the response, can we close the issue ?