Is it possible to patent a particular implementation and use case of a well-known statistical technique?

I have to be intentionally vague for obvious reasons.

The technique is standard and reasonably well known amongst Bayesian statisticians. However, I have adapted the technique to a particular use case and the details of the implementation are quite different from what existing papers talk about.

I've seen some papers or blogs say "[this technique] can be extended to [use case]" however none of them go into any more detail than that.

Is this kind of thing patentable?