How to classify the rules, and analyze what the classification rules are

Human beings for the laws of nature.

The understanding developed with the development of the natural sciences, and in ancient times it was intuitive, and in modern times it has a mechanistic theory.

characteristics. In modern times, human understanding of the laws of nature has not only overcome the one-sidedness and isolation of ancient and modern times, but has also been expanded and deepened.

The laws revealed by modern natural science can be broadly divided into two categories:

The law of mechanical determinism. According to this law, the state of the material system at each moment is determined by the initial state and boundary conditions of the system.

It is decided by a single value. Differential equations from integrables.

The dynamics law of the expression is a typical manifestation of this law, and its solution is determined by the initial conditions and boundary conditions of the system.

Statistical laws. This law is the integral characteristic of a system composed of a large number of elements, and any single element in the system remains subject to the laws of mechanical determinism. Statistical physics equations are typical of this regularity, and their solution depends on the statistical average of the corresponding dynamics of the elements of the system at the initial moment.

on quantum mechanics.

There are two different understandings of statistical characteristics. One view is that the statistics of quantum mechanics are the behavior of quantum systems, and that individual microscopic particles do not have random behavior. Most physicists, on the other hand, hold a different understanding, arguing that although quantum mechanics are microscopic systems of quantum states.

It provides an objectively measurable sum, and in principle, it is not possible to make a single-value prediction about the behavior of each microscopic object, but only the probability of the occurrence of each possible behavior. The statistics of quantum mechanics are expressed by the interaction of individual particles with instruments. It is now found that the differential equations of complex mechanical systems are mostly non-integrable, and these equations themselves have "intrinsic randomness", that is, the behavior of the system they describe cannot be determined by the initial conditions singarily.

As a result, some people believe that this is an inherent law of randomness that is different from the laws of mechanical determinism and statistics.

The process of categorical analysis.

Classification is composed of two parts: rule recognition and **, in which the recognition part is implemented by two algorithms: classification and clustering, and the ** part can be completed by the classification model or manually according to the clustering rules.

Classification algorithms can be classified into two categories: classical algorithms and deep learning algorithms, the classical algorithms mainly include logistic regression, support vector machines (SVMs), and decision trees; Deep learning mainly includes forward neural networks and recurrent neural networks.

There are many evaluation indicators of classification and classification models, such as accuracy, precision, recall, etc., and the specific selection of what indicators need to look at the purpose of the classification task.

Finally, there are several types of clustering algorithms: k-means clustering, Gaussian mixture models, and self-organizing mapping neural networks.

Cluster analysis.

The basic logic of clustering is to group things with various differences into several categories according to their similarities in some aspects according to a certain method, and the differences between classes are relatively large, while the differences between things in the same class are relatively small. Therefore, for the clustering method, we need to pay attention to two core issues: the similarity between things, and the difference between classes.

Different angles of categorizing things will lead to different ways of categorizing. When a group of people is classified, there are adults and children according to age, men and women according to gender, southerners and northerners according to geography, and rich and poor people according to wealth.

The opposite of similarity is the difference of things, how to evaluate the difference or similarity of things? Mathematically this is called "distance". When classifying "people", the difference between people can be seen as the distance between people.

How should we evaluate the similarities and differences between people? It can be studied from various descriptive attributes of people.

The same group of people can be divided into different categories according to their demographic variables, but for specific groups of people, such as customers, people who gather in one place, there are other factors to consider, as shown in the figure below for people who gather in a certain location.

, is a composite number, and the others are prime numbers, so the law is a combination of quality and quality.

A composite number is a number of natural numbers that can be divisible by other numbers (except 0) in addition to 1 and itself.

There are simply too many types to find patterns. Some of them have regularities between the given numbers, and some have regularities between the next number. There is also a certain pattern of yards for the difference between two adjacent numbers. The law may be the same addition, the same subtraction, the same multiplication, a number or a series of numbers, or squares.

The definition of law is (objective and universal. ）

It is the inevitable connection between things and determines the inevitable trend of things. Do not transfer with the will of man. The law cannot be changed. But laws can be used for the benefit of mankind. FYI.

Regularity refers to rules that form naturally. The law summarizes the law of movement, the law of change and the law of development. The law of motion determines the trajectory and speed of things; The law of change determines the change of the nature of things; The law of development determines the direction of development of things.

Regularity manifests itself through phenomena. Phenomena are the guide for people to understand the laws, and the laws are revealed through the analysis of the phenomena. To reveal the law, comprehensive and systematic analysis and argumentation must be carried out, and the evidence must be sufficient, true, and reliable.

Things follow certain laws in the process of movement, change and development, which requires us to discover, grasp and use laws. In practice, people can discover and understand laws through a large number of phenomena. If we grasp the law, we can make correct decisions about the movement and changes of things, and we can give correct explanations for the causes of certain phenomena, and we can also prevent the occurrence of phenomena that are unfavorable to people and promote the emergence of phenomena that are beneficial to people.

If we want to achieve the desired goal in practice, we must recognize and respect the law and act according to the law, otherwise we will be punished by the law.

Human beings are powerless in the face of some laws, but they can do a lot in the face of others. For example, human beings are powerless in the face of the laws of the operation of celestial bodies, and they can make great achievements in the face of the laws of economic operation. Human beings cannot regulate the operation of the earth, but they can regulate the operation of the economy, prevent and resolve various economic crises, curb inflation, eliminate poverty, protect and improve people's livelihood, and so on.

Characteristics of regularity.

Laws have objectivity, inevitability, stability, and repeatability.

1. Laws are objective and can neither be created nor destroyed.

2. The existence, effect and consequences of the law are inevitable.

3. The role of law is stable.

4. Patterns can be repeated.

The kind of regularity.

Laws are divided into: causal laws, laws, rules, effects, axioms, and theorems.