Saturday, September 9, 2017

Frederik Michel Dekking "A Modern Introduction to Probability and Statistics, Understanding Why and How"

What is it?

As you can see from the title this is a book about probability theory and math statistics. About one you studied (or studying now) if you've ever studied at technical university. It describes how to estimate your chances in a situation when you don't have enough info to be 100% sure and how to verify the quality of info that you have.

The first half of this book I can remember from my university probability theory class, but the second half is a bit deeper. First insight that I got is that we can actually check the quality of our data. So usually when we work with data from our website or app analytics, for example, we make a conclusion based on data we have. Suppose we got an info that our users spend in average 2 minutes on the product profile screen. How can we say that this info is true? We can test our hypothesis about 2 minutes against the hypothesis that users actually spend on user profile page less than 2 minutes. The book gives us a lot of methods of doing this in different situations so we can get more precise results from data that we have. Which is very helpful!

There are a lot of interesting historical examples in the book as well, i.e. like how probability theory gave more precise information about German production during Second World War than British intelligence agency. It gives you understanding about real world application of your knowledge which gives you a lot of motivation to continue reading. 

Unfortunately, I had not so much time to complete the exercises, but I solved all the Quick Exercises (which are very easy, honestly, but still). You meet Quick Exercises among the text of the lesson and it helps you to understand what you've just read. Since I've read the book in transport I solved everything in my mind without even the piece of paper in my hands. I hope I will find time to get back to real exercises because otherwise, I think, I will waste my knowledge very fast

Who should read the book?

If you work in IT and want to be more effective for your company, probably this book will be a good choice. It will give you a knowledge about how to treat the data, how to work with it, how to make the right conclusions based on it. IT full of data now, all business is data driven, so you have to be effective with it. 

This book will be very good for machine learning and data scientist beginners because it will help you to understand that you can do a lot of things without machine learning and it will help you to understand how machine learning works internally, so you can use it more efficient.

This book will be awesome for students of different departments. It will help you to get probability theory right and it will definitely help you in your scientific research (because all the experiments and theory are based on probability theory because we don't have enough data to be 100% sure in something)

Is it complicated? Should I know maths?

It is not complicated. Authors made a great job to put so difficult material in so easy-to-understand way. But maths is still necessary. You should know at least how derivatives work, integral calculus and some math analysis. I read all the book in a bus (on my way to the office and back), sometimes it took me almost 10 minutes to get how does the thing work, but it is because I've done all the calculus in my mind without paper and pen. So if I can do it in my head, and I am not a very good mathematician, I guess it is not complicated at all. 


This is a very good book to get probability theory. It explains everything very clear and now I can tell that I understand probability theory and maths stat basics at least. I think it is enough for me, cause it covered all my questions, it can be not enough for one who works with probability theory every day and wants to get some deeper concepts. But if you want to learn it or recall what you've studied in university and a bit more this is a great book that will answer all your question and give the good vision of what is the probability theory now.

No comments:

Post a Comment