This book is excerpted from books. Python itself has many third-party libraries for machine learning, but in most cases, this book only uses Numpy, the basic scientific computing library, to implement algorithmic code. The goal is to give readers a better understanding of the details of machine learning algorithms and to understand Numpy's various applications. However, as a supplement, this book will apply the model in the mature third-party library of scikit-learn when appropriate.
"Machine learning" may not be as well known to everyone in recent times, but it is also a very hot word. Machine learning is a literal translation of the English word "Machine Learning" (ML for short), which literally shows that this technology is a technique for the machine to "learn". However, we know that the machine is dead after all. The so-called "learning" is, in the final analysis, only a series of operations that humans "give" to the machine. This "giving" process can be implemented in a variety of ways, and Python is one of the languages ​​that is relatively easy to use and performs quite well. This article intends to talk about some of the broader knowledge of machine learning, and then introduces and explains why Python is used as a tool for machine learning. Finally, we will provide an easy-to-understand, practical example to give you an intuitive experience.
Specifically, the main points covered in this chapter are:
The definition and importance of machine learning;
The excellence of Python in the field of machine learning;
How to configure Python machine learning environment on your computer;
Machine learning general steps.
Introduction to Machine Learning As mentioned in the introduction, “machine learning†has become a very popular term due to recent recent achievements. The excellent performance of machine learning in various fields (the master of Go is the most representative existence), making people in all walks of life more or less interested and awe in machine learning. At the same time, however, groups that misunderstand machine learning are growing stronger; they may think of machine learning too mysteriously, or think of it too versatile. This section is intended to provide a general introduction to machine learning, as well as some common terms in machine learning to facilitate the description of the following chapters.
What is machine learning?
In the morning, "The weather is really good today", the chill between friends "You just went to eat", after the exam, I sigh "Review for so long, there are gains"... These words can be seen everywhere in daily life, Behind it is the idea of ​​“learningâ€â€”they are all making effective decisions about past new experiences using unknown experiences. And the process of giving this decision to the computer can be said to be the most succinct definition of "machine learning."
We may first talk about how machine learning differs from previous computer work styles. If a traditional computer wants to get a certain result, it needs humans to give it a series of real instructions, and then the computer executes it step by step according to the series of instructions. The causal relationship in this process is very clear, as long as human understanding does not deviate, the results of the operation can be accurately predicted. But in machine learning, this traditional style is broken: the computer does still need humans to give it a bunch of instructions, but this series of instructions often can't directly get the result; instead, it is a series of instructions that give the machine "learning ability" . On this basis, the computer needs to further accept the “data†and “learn†the final result based on the “learning ability†that humans have given it before. This result is often not just straightforward programming. So here is a slightly deeper definition of machine learning: it's a way for computers to use data rather than instructions to do everything. Behind this, the most crucial thing is the idea of ​​"statistics". The concept of "relevant rather than causal" that it advocates is the theoretical foundation of machine learning. On this basis, machine learning can be said that the computer uses the data input to it, and uses the algorithm given by human to obtain a certain model. The ultimate goal is to use the model to predict the information of unknown future data.
Since statistics are mentioned, certain mathematical theories are indispensable. Relevant, short definitions are given in Chapter 4 (PAC framework). Here we will only describe the profound nature of machine learning under statistical theory: it pursues a rational hypothesis space (Hypothesis Space). The selection and generalization of the model (GeneralizaTIon) capabilities. Some special terms appear in the sentence, and detailed definitions are mentioned in the introduction of terms. Here we provide an intuitive understanding:
The so-called hypothesis space is the "applicable occasion" of our model in mathematics.
The so-called generalization ability is the performance of our model on unknown data.
Note: The above is strictly speaking, it should be the essence of PAC Learning; in the rest of the theoretical framework, machine learning can have different cores.
As can be seen from the above discussion, the process of machine learning and human thinking is more or less similar. In fact, we have the corresponding theoretical background of neuroscience behind the neural network (NN) and the ConvoluTIonal Neural Network (CNN) in Chapters 6 and 7. However, at the same time, it is necessary to know that machine learning is not a "learning robot" and "artificial man with learning ability". This can be clarified from the above discussion. (Oh, the author is in the first When I heard the words "machine learning", it was the image of a "smart robot" that appeared in my mind, and even imagined the scene in which it lived with humans. Instead, it is a tool used by humans to discover the information behind the data.
Incremental encoders provide speed, direction and relative position feedback by generating a stream of binary pulses proportional to the rotation of a motor or driven shaft. Lander offers both optical and magnetic incremental encoders in 4 mounting options: shafted with coupling, hollow-shaft, hub-shaft or bearingless. Single channel incremental encoders can measure speed which dual channel or quadrature encoders (AB) can interpret direction based on the phase relationship between the 2 channels. Indexed quadrature encoders (ABZ) are also available for homing location are startup.
Incremental Encoder,6Mm Solid Shaft Encoder,Hollow Rotary Encoder,Elevator Door Encoder
Jilin Lander Intelligent Technology Co., Ltd , https://www.jilinlandertech.com