3 ms·

Good that it discusses overfitting. I can't help but wonder about those recommender systems. With so little material on statistics I have to assume it's only a

Good that it discusses overfitting. I can't help but wonder about those recommender systems. With so little material on statistics I have to assume it's only about observational data, which is the best way to make the millionth+1 useless recommendation engine. And why is it that the "data science" books never discuss DoE?

x0x0 12y ago
Most data scientists, particularly those coming from the cs department, lack most probability and statistics fundamentals. I doubt many of them have even heard of an anova, or sampling distributions, F tests, chi2, etc. To be fair, ml tends to focus very heavily on prediction, not inference / interpretation of betas. In many tree models how to even understand coefs is an open question.
- ced 12y ago
  anova, or sampling distributions, F tests, chi2 Most of the big ML books are heavily Bayesians, and these subjects are less discussed (though IIRC Gelman's book has a "Bayesian ANOVA"). Even Elements of Statistical Learning, which is very frequentist in its approach, only references ANOVA in passing. Do you have any book to recommend about these fundamentals?
  - x0x0 12y ago
    basic level, very approachable, filled with case studies (and with R code to run them easily found), but stupidly expensive: _statistical sleuth_ by ramsey (but, you know, pdfs can be found on the internets) intermediate level, covers some blocking IIRC: _Statistics for Experimenters_ by Box et al advanced: I thought quite good, but classmates did not universally love. Unfortunately does not come with case studies or R code to run them; I have a bunch but (very unfortunately) printed instead of computerized and, in any case, probably copyrighted by my professors. _Experiments: Planning, Analysis, Operation_ by Wu and Hamada. The math is not complex but can be involved for various types of blocking designs.
    - ced 12y ago
      Thanks a lot, statistical sleuth looks very readable and interesting! How do you feel about the Bayesian approach to these questions? (cf. Gelman's Bayesian Data Analysis)
ced 12y ago
Do you have any reading recommendation about DoE?
Goladus 12y ago
What is DoE? I hear DoE and I think "Department of Education" or "Department of Energy" which is also what comes up in a web search...
- christopheraden 12y ago
  Most likely http://en.wikipedia.org/wiki/Design_of_experiments http://en.wikipedia.org/wiki/Design_of_experiments
memilanuk 12y ago
I've been wondering just about exactly the same thing - why DoE gets no love, while everyone is crazy for ML.

x0x0 12y ago

Most data scientists, particularly those coming from the cs department, lack most probability and statistics fundamentals. I doubt many of them have even heard of an anova, or sampling distributions, F tests, chi2, etc. To be fair, ml tends to focus very heavily on prediction, not inference / interpretation of betas. In many tree models how to even understand coefs is an open question.

ced 12y ago
anova, or sampling distributions, F tests, chi2 Most of the big ML books are heavily Bayesians, and these subjects are less discussed (though IIRC Gelman's book has a "Bayesian ANOVA"). Even Elements of Statistical Learning, which is very frequentist in its approach, only references ANOVA in passing. Do you have any book to recommend about these fundamentals?
- x0x0 12y ago
  basic level, very approachable, filled with case studies (and with R code to run them easily found), but stupidly expensive: _statistical sleuth_ by ramsey (but, you know, pdfs can be found on the internets) intermediate level, covers some blocking IIRC: _Statistics for Experimenters_ by Box et al advanced: I thought quite good, but classmates did not universally love. Unfortunately does not come with case studies or R code to run them; I have a bunch but (very unfortunately) printed instead of computerized and, in any case, probably copyrighted by my professors. _Experiments: Planning, Analysis, Operation_ by Wu and Hamada. The math is not complex but can be involved for various types of blocking designs.
  - ced 12y ago
    Thanks a lot, statistical sleuth looks very readable and interesting! How do you feel about the Bayesian approach to these questions? (cf. Gelman's Bayesian Data Analysis)

ced 12y ago

Do you have any reading recommendation about DoE?

Goladus 12y ago

What is DoE? I hear DoE and I think "Department of Education" or "Department of Energy" which is also what comes up in a web search...

christopheraden 12y ago
Most likely http://en.wikipedia.org/wiki/Design_of_experiments http://en.wikipedia.org/wiki/Design_of_experiments

memilanuk 12y ago

I've been wondering just about exactly the same thing - why DoE gets no love, while everyone is crazy for ML.