What type of neural network to choose?

Question

Neural networks

There is a certain ecosystem with many parameters. Periodically, crises occur in the system. If a crisis is caused by a sharp change in one of the parameters, then it is easily detected and prevented. If the reason is a slight change in a group of parameters, then this is easily determined after the fact, but it is problematic to notice in real time.
In theory, of course, everything can be described by dependencies and conditions, but it turns out an unrealistic amount of work. I want to try to attach a neural network to all this so that it predicts the onset of a crisis. What type of network is better to use in this case and in which direction (algorithms) to dig? (I would not like to start with something that is radically inappropriate).
I am new to neural networks, but I understand the general idea and principle of operation.

Answer 1

There is such a section of Data Science, called "Anomaly Detection" (Anomaly Detection). Within this section, there are subsections with the search for stationary anomalies, and with the search for anomalies in time series, and the analysis of information in the stream, etc. The task of the fastest detection of anomalies is also set.
Neural networks are also used in this section, but not only. But in any case - it is necessary to dig in this direction. Fortunately, there are more than enough literature and Internet resources.

Answer 2

With such introductory - any.