We had a reading party for Introduction to Statistical Modeling for Data Analysis, which is famous as "Green Book". At that time, I made a reading memo and share it. I uploaded it as Jupyter Notebook on GitHub. The R and WinBUGS codes in the book are written in Python and Stan as much as possible.
Create a statistical model to understand Chap1 data Chap2 Probability Distribution and Statistical Model Maximum Likelihood Estimate Chap3 Generalized Linear Model (GLM) -Poisson Regression Chap4 GLM Model Selection-AIC and Model Prediction Chap5 GLM Likelihood Ratio Test and Test Asymmetry Expanding the range of applications of Chap6 GLM-logistic regression, etc. Chap7 Generalized Linear Mixed Models (GLMM) -Modeling of Individual Differences Chap8 Markov Chain Monte Carlo (MCMC) Method and Bayesian Statistical Model Bayesian modeling of Chap9 GLM and estimation of posterior distribution Chap10 Hierarchical Bayesian Model-GLMM Bayesian Modeling Chap11 Hierarchical Bayesian model with spatial structure
Official name, Introduction to Statistical Modeling for Data Analysis. The cover is green, so it's called a green book. An introductory book on statistical modeling (a method of creating a model and applying it to observation data to understand the phenomenon). I think that it is often introduced in TJO's blog article. According to the preface, the reader is supposed to be "a person who has not received basic training to" express and explain mathematical models with phenomena "."
I usually touch the data, scatter it in Excel and draw an approximate curve, but I don't know what I'm doing, so I thought it would be good to read it. Personally, I felt a difficult atmosphere from the cover and the title of the book, so I hesitated to read it, but as I read it, as the name "Introduction" suggests, it was written in a very easy-to-understand manner. .. Rather than starting studying from Introduction to Statistics at the University of Tokyo (Akamoto), I feel that it may be better to read after understanding the flow of statistical modeling in the green book. I am.
In the book, the explanation is advanced using R, but there are many codes written in Python on the net, and I think that it is possible to run it in Python at hand by referring to that. (Thank you for referring to many sites.) If the sample data is in RData format, I think it is necessary to convert it to CSV on R. You can also see code written in Stan instead of WinBUGS code.
It may get difficult from around chapters 5 and 6, but I think it's a good book, so please read it!
Recommended Posts