arXiv:1808.10173v3 [stat.AP]

Publicity
, , ,
1 
@misc{https://doi.org/10.48550/arxiv.1808.10173,
  doi = {10.48550/ARXIV.1808.10173},
  
  url = {https://arxiv.org/abs/1808.10173},
  
  author = {van Elst, Henk},
  
  keywords = {Applications (stat.AP), FOS: Computer and information sciences, FOS: Computer and information sciences},
  
  title = {An Introduction to Inductive Statistical Inference: from Parameter Estimation to Decision-Making},
  
  publisher = {arXiv},
  
  year = {2018},
  
  copyright = {arXiv.org perpetual, non-exclusive license}
}

Earlier this week I completed a general revision of these lecture notes and provided a hyperlink to worked examples coded in R and Stan at GitHub - hve1964/stanCodes: Stan codes accompanying online lecture notes on Inductive Statistical Inference. Again, I’d like to express my deep gratitude to the Stan community for sharing their knowledge and material so generously. Cheers.

8 Likes
2

Thanks for sharing, @hve1964. The Stan code looks great. Some of it could be made shorter and more efficient with vectorization, but it might be harder for beginners to understand. Also, some of the sampling would be more efficient with the lupdf forms of densities that drop normalizing constants. There’s usually no reason to implement the full joint density as we only need it up to a proportion for MCMC.

P.S. I love the spelling of “Gauß”!

P.P.S. The story I heard about the invention of Monte Carlo methods was Ulam being sick in the hospital and trying to estimate the probability of winning at solitaire by simulation.

2 Likes
3

Dear @Bob_Carpenter,

many thanks for your very kind comments. I appreciate this very much, foremost since they motivate me to try and dig deeper.

Actually, as regards programming skills (in Stan and in general) I consider myself still an absolute beginner. In particular, I am anxious to learn more about further possibilities of vectorising (and thereby improving efficiency of) Stan code. Also, I am totally ignorant of lupdfs: Could you please point me to relevant references? (I suppose I would have to work through the Stan User’s Guide, etc. in more detail, wouldn’t I?)

All the best.

PS: Well, by now, LaTeX offers the possibility of employing (and, thus, printing) special characters such as “\ss” for at least three decades. Unfortunately, this information appears not to have spread as much in the community(ies) as one would hope for.

PSS: Thanks for sharing the anecdote on Ulam, which was unknown to me so far.