Linear Algebra and Learning from Data

Gilbert Strang, well known MIT professor and author, writes another book on Linear algebra. He put a lot of effort into making the material accessible and not assuming a background in linear algebra (matrices) so aimed at beginners. There is a bit of 'personal commentary' added to the text that is trying to make the public comfortable that wouldn't normally be in a text book but doesn't bother me much here. The added focus is on applications to Machine learning and other data extraction so it focuses on linear algebra that are useful for that purpose and how they are useful.

This has an incredible amount of information about applications of Linear Algebra all over Computer Science and beyond.It is reasonably well explained, however Mike X Cohen's Linear Algebra book "Theory, Intuition, Code" has much better explanations (although covering far fewer topics).I would recommend using both books together.

Prof Strang has been writing intoductory linear algebra books since the mid nineteen seventies. All of them good. In this book he sharply departs from his own and ever other book in introductory presentation and presents the outer product at the same level of detail as the inner product. This clarifies many applications, by allowing discussion of rank one corrections in applications. I really like this. For me it was late undergraduate or early graduate school that the outer and inner product became peer in the type of applications that I was interested in. While this book has an eye towards “machine learning,” it is very clear that Prof Strang sees all applications as data science. This book is his distillation of that. I wish I’d had this at the beginning of my education. This was a wonderful reintroduction to a view I use everyday.

Prof. Strang is just a masterful teacher, a teacher's teacher. I have followed him for years via both his online lectures and his books, and neither this book nor the companion online course disappoint. It is all the more admirable that even late in his career he has stayed current in his field (he is quite the guru of applied linear algebra) and seems to always be striving for innovative ways to teach in both his lectures and his books. For those interested in learning both some of the basics of linear algebra and some of the more advanced topics that are pertinent to ML and data analysis, I highly recommend this book.

While not a full-on linear algebra book (despite the title), this does serve as a perfect undergraduate-level introduction to the Machine Learning galaxy and its many, many applications and increasingly popular methodology. Computer scientists, mathematicians and engineers - as well as math-savvy economists and businesspeople - could benefit from a class using this text or from self-learning if one is not prepared for a classic like Deep Learning.Upsides include a thorough review of linear algebra and a very up-to-date list of data analytic topics that are at the edge of research and recently implemented techniques, including Google's new ranking system, which it continues the recent trend of texts in covering as a side section late in the book. This is a Gil Strang book, which means problems, problems, problems galore (no answers are given within the text). Chapter VI on optimization methodology is the star of this show, and is a really good stepping stone to the higher-level texts on these topics.Downsides include breadth-not-depth coverage and the need for a little bit of organization. There is far more in here for a single class, but it seems like it was written with the purpose to be all or almost all used for a single semester's worth of coursework. This may work for MIT, but not for a second- or third-tier undergraduate department (at least not without significant trouble). Fortunately, an instructor can simply cut some topics, but the students should at least read up on these for the purposes of interest; they are certainly NOT comprehensive introductions - hence, the "breadth-not-depth" critique above; see e.g. the Compressed Sensing section for a prime example of ... well, sparsity of information. It could also stand for some significant clean-up and organization.Glibert Strang is well-known in math cirles as an extraordinary innovator and teacher of linear algebra. I found this book from him simply delightful, and, though my needs already exceeded it in general, was surprised on how much I learned from it still.

I wanted to brush up on linear algebra just because it's been so long since I've really gone through all the details. This is a great refresher course for me. And a lot of the conceptual pedagogy that Strang is so well known for really shines here. Maybe it's because I am already familiar with the material, but Strang makes it feel like meeting an old friend for dinner.

Excellent coverage of Principal Component Analysis. Good high-level overview of Fast Fourier Transformation. It is not a complete chapter on Fourier analysis, but enough to understand how it works.Has a collection of applications of linear algebra. Not a good first text book -- use Strang's Linear Algebra for that, but an excellent book for applications.

I'm preparing a statistics course with a large linear algebra component. I needed a good solid reference that links linear algebra to statistical modeling. So why not go with the guru?