Is Maximum likelihood estimated parameters used only for dataset preparation, what do we do after getting maximum likelihood values?

Question

[kiran]

I'm new to statistics, i'm confused about what after finding maximum likelihood estimators.

what do we do with those values, is only for preparing most probable dataset, like if i'm using any statistical model i'm making sure that my data is most best probable to respond to that model (linear, logistics regression, or any other model) ?

After finding the ML estimation parameters, i'm so confused where are they used ?

Can anyone give me clarity about this please.

thanks for your time.

regards, kiran.

[liorglic]

Well, that's a bit of a broad question, but generally I'd say that ML can be helpful for optimizing model parameters. That's important if these parameters have some biological or other meaning. Another use of ML estimates is when comparing models - you can calculate the MLE of multiple models and see which one gets the best likelihood and thus best fits the data. Note that to perform such comparisons you need to apply likelihood ratio tests or the Akaike Information Criteria (AIC).