What can neural networks do?

Neural network uses the existing data to find the weight relationship between input and output (approximate), and then uses this weight relationship to simulate, such as inputting a set of data to simulate the output results. Of course, your input should be in the same category as the data set used in training. For example, forecast weather: temperature, humidity, air pressure, etc. As the input weather conditions as the output, a neural network is trained by using the historical input-output relationship, and then today's temperature, humidity, air pressure and so on. By using this neural network, the upcoming weather conditions can be obtained. Of course, this example is not accurate enough, but neural networks are usually used. Hope to adopt!