6 ms·
A lot of "applied statistics" is now performed under the headings of "machine learning" and "data mining". Both fields are thriving. Furthermore, Bayesian meth
by ced 16y ago
A lot of "applied statistics" is now performed under the headings of "machine learning" and "data mining". Both fields are thriving.
Furthermore, Bayesian methods have come back to the fore in the past ~15 years. They are quite likely the future of statistics, especially in academia.
- xtho 16y agoYou can't solve all problems with machine learning and data mining. How would you apply those methods, e.g., to psychological experiments or test planing when every single test costs a considerate amount of money?
- moultano 16y agoSituations like that are not why statistics is becoming the new in-demand skill. Expensive trials have existed for a long time. Petabytes of barely-structured data haven't.
- xtho 16y agoI didn't comment on the article but on the statement above about applied statistics being dominated by ml and about Bayesian statistics being the future.
- jey 16y agoBayesian experimental design? https://secure.wikimedia.org/wikipedia/en/wiki/Bayesian_experimental_design https://secure.wikimedia.org/wikipedia/en/wiki/Bayesian_expe...