What is your background and what is your goal for learning the methods?
This is somewhat long and there is a disclaimer towards the end, but hopefully some of this is helpful.
Working through a book can mean reading what’s on the pages and being able to recall names of techniques or methods. Or using pen and paper to work through the examples and be able to solve problems. This could even be deriving what’s in the book from first principles.
This will depend on what you want to do with the material. If you want to apply it using pre-made R packages, you probably don’t need to recreate everything from scratch and you can probably get away with ISL. If you want to be creating new methods or going beyond pre-made R packages, then you probably need to work up to ESL and solve things from first principles.
ISL is used in an undergrad elective course at my uni. The prerequisite stat material covers Devore probability and stats for engineers and intro to linear regression by Douglas Montgomery. ISL would be a third course in stats (see the bottom for the math background 4 courses). There are entire courses dedicated to the topics in ISL, so I really think ISL is most useful to bring previously studied topics together.
ESL is used in a second year MSc course. This assumes knowledge of mathematical statistics (Casella and Berger Statistical Inference + Wasserman All of Statistics), computational statistics (topics: bootstrap, MCMC, EM algorithm, numerical analysis methods, optimization, and matrix decomposition) and courses on linear regression and the general linear model. So it’s a “capstone” of sorts that ties all of the material together. I haven’t taken any of these courses, so I can’t comment on what’s really necessary.
Disclaimers follow: As others have mentioned someone’s background and preparation may be different and more advanced than what is outlined. Above I outlined the course sequences for ISL and ESL at my uni. We do not require a course on real analysis and we do not do measure theoretic probability (PhDs do but ESL is covered in the MSc that is required for PhD admissions). Of course not every chapter in a textbook is covered in each course and I’m sure there is some sort of minimal coverage of topics that will allow you to get to ISL or ESL in a more efficient way. What that is, I am unable to comment on.
Yes there are people admitted to the MSc program without a stats BSc degree. Examples are physics, math, and computer science majors from what I have seen. Usually they have to make up missing BSc math stats courses.
Undergrad level math background assumes calculus to include multi variable calculus (Stewart Calculus omitting the chapters on vector calculus). Partial derivatives, Lagrange multipliers, multiple integrals. Also linear algebra, matrix multiplications, determinants, eigenvalues, trace (linear algebra and its applications by Lay).
This is somewhat long and there is a disclaimer towards the end, but hopefully some of this is helpful.
Working through a book can mean reading what’s on the pages and being able to recall names of techniques or methods. Or using pen and paper to work through the examples and be able to solve problems. This could even be deriving what’s in the book from first principles.
This will depend on what you want to do with the material. If you want to apply it using pre-made R packages, you probably don’t need to recreate everything from scratch and you can probably get away with ISL. If you want to be creating new methods or going beyond pre-made R packages, then you probably need to work up to ESL and solve things from first principles.
ISL is used in an undergrad elective course at my uni. The prerequisite stat material covers Devore probability and stats for engineers and intro to linear regression by Douglas Montgomery. ISL would be a third course in stats (see the bottom for the math background 4 courses). There are entire courses dedicated to the topics in ISL, so I really think ISL is most useful to bring previously studied topics together.
ESL is used in a second year MSc course. This assumes knowledge of mathematical statistics (Casella and Berger Statistical Inference + Wasserman All of Statistics), computational statistics (topics: bootstrap, MCMC, EM algorithm, numerical analysis methods, optimization, and matrix decomposition) and courses on linear regression and the general linear model. So it’s a “capstone” of sorts that ties all of the material together. I haven’t taken any of these courses, so I can’t comment on what’s really necessary.
Disclaimers follow: As others have mentioned someone’s background and preparation may be different and more advanced than what is outlined. Above I outlined the course sequences for ISL and ESL at my uni. We do not require a course on real analysis and we do not do measure theoretic probability (PhDs do but ESL is covered in the MSc that is required for PhD admissions). Of course not every chapter in a textbook is covered in each course and I’m sure there is some sort of minimal coverage of topics that will allow you to get to ISL or ESL in a more efficient way. What that is, I am unable to comment on.
Yes there are people admitted to the MSc program without a stats BSc degree. Examples are physics, math, and computer science majors from what I have seen. Usually they have to make up missing BSc math stats courses.
Undergrad level math background assumes calculus to include multi variable calculus (Stewart Calculus omitting the chapters on vector calculus). Partial derivatives, Lagrange multipliers, multiple integrals. Also linear algebra, matrix multiplications, determinants, eigenvalues, trace (linear algebra and its applications by Lay).