An intelligent model with chaos and causality

The current machine learning method has the disadvantages of relying on data and statistics, and the model is uninterpretable. So, we propose a model based on chaos. The model reconstructs the original system's attractor equivalently and explores the causality so that it can achieve accurate prediction of complex systems without statistics. At the same time, causality can explain the model. By analyzing the properties of chaos, such a model is sensitive and adaptable to the change of conditions, and its learning process is more like humans'. This method works well and makes precise predictions in complex nonlinear systems. We think causality and chaotic theory play important parts in the proposed method and it may provide a new idea for current research.