Within statistics, I have heard that almost all analysis can be broken into two general classes:
- Prediction : E.g. Statistical Modelling, Machine Learning
- Inference
I have seen the term "inference" being used to describe the following:
- Inferring the parameters of an unknown probability distribution.
- Inferring the independent, actual effect of a particular phenomenon.
- Making inferences about a population are made based on certain statistics calculated from a sample of data drawn from that population.
My Question: Although this is more of a pedantic issue - but in the above three points, is there anything stopping you from replacing "inference" with "prediction"? For example:
- Predicting the parameters of an unknown probability distribution.
- Predicting the independent, actual effect of a particular phenomenon.
- Making predictions about a population are made based on certain statistics calculated from a sample of data drawn from that population.
Can someone please comment on this?
Thanks!