Artificial intelligence is a subject that explores and studies the way to simulate human intelligence with various machines, so that human intelligence can be materialized and extended. It uses bionic thought for reference and abstracts the knowledge with mathematical language to imitate the biological system and the intelligent mechanism of human beings. The main methods are neural network, evolutionary computing and granularity calculation. Neural network is a kind of parallel processing network which simulates the intuitionistic thinking of human image on the basis of the research of biological neural network. According to the characteristics of biological neuron and neural network, a kind of parallel processing network is summarized by simplification, induction and extraction.

The main functions of neural network are associative memory, classification and clustering, optimization calculation and so on. Although neural network has many shortcomings, such as complex structure, poor interpretability, long training time and so on, because of its high tolerance to noise data and low error rate, as well as the continuous proposal and improvement of various network training algorithms, such as network pruning algorithm and rule extraction algorithm, the application of neural network in data mining is more and more favored by the majority of users. Evolutionary computation is a general problem solving method developed by simulating biological evolution theory. Because it comes from the biological evolution of nature, it has the The very strong adaptability shared by natural organisms makes it possible to solve complex problems that are difficult to be solved by traditional methods.