To master the mathematical knowledge in machine learning quickly, we need to proceed from three directions. The first is to master the core concepts and master the core concepts. We need to master the core concepts, such as what are the core concepts in linear algebra? It is a measure of linear space, vector matrix and vector matrix, including norm and inner product. These are its core concepts. Therefore, in probability statistics, the difference between frequency school and Bayesian school is a core concept, as well as expected variance, conditional probability and other indicators. Such a concept, joint probability of conditional probability is also a core concept. So in optimization, these algorithms, this gradient descent method, or Newton method, are the core concepts.
Then there is the point to face. Specifically, in the case of limited time, we must concentrate our limited energy on important knowledge. Make these core concepts clear first, and then spread these key issues through these core concepts, and gradually contact other issues. Doing so will help to increase our knowledge of mathematics.
Finally, problem-oriented, that is, combining our actual needs with our actual problems to decide what we want to learn. Because after all, machine learning and mathematics are all about solving problems. If you can't solve the problem, what you have learned is not as valuable as the knowledge that can solve the problem. Of course, it can't be said that it is worthless. When studying, you can try to be problem-oriented. Explore this knowledge with questions and learn knowledge with questions. At that time, we will find that it will be more efficient.
This paper introduces the related contents of mathematics in machine learning. Through these contents, we can better grasp the essentials of machine learning. You should know that mathematical knowledge is a very important knowledge system. Only by learning mathematics well can we lay the foundation for machine learning. I hope this article can help you better.