If so can we apply the approach for Market Regime Detection?
Market regime detection is an important problem in finance, as it involves identifying shifts in the behavior of financial markets over time. Understanding these shifts can help investors to manage risk and develop more effective trading strategies. The research paper "Causal deconvolution by algorithmic generative models" describes a method for inferring causal relationships between variables using algorithmic generative models. This study proposes to apply this method to the problem of market regime detection, with the aim of identifying causal relationships between different market variables and detecting changes in market regimes.
First of all, what is the most optimal way to represent market information in a cell configuration?
Then, how do we auto-discover and extract features from such space-time diagrams?
I do understand where your question is aiming, but this is a massive chunk of information and research to consume. When time allows, possibly, I will try to develop such approach as you are speaking about.
In the meantime, you might find interesting complex systems entropy measure as useful to classify different operational modes of virtually any complex system in existence. The whole methodology is discussed in detail on prediction of arrhythmia from ECG recordings in the bellow provided research.
Basically, what is done is that a signal is taken, processed by permutation entropy, and classified by machine learning methods. This one sentence is encompassing a lot of understanding and steps that are in depth explained in this research.
https://www.researchgate.net/publication/340460356_Prediction_of_Medical_Drugs-Induced_Arrhythmias_Torsades_de_Pointes_Ventricular_Tachycardia_and_Ventricular_Fibrillation_on_Rabbit_Model_Up_to_One_Hour_Before_Their_Onset_Using_Computational_Method_Ba
I don't think so. Economic organizations and automata are not similar. Using cellular automata to simulate the economy doesn't work.
In complex systems science, computers to simulate biological evolution,likewise impossible, because whether living organisms or economic organization, they are something else, too far from computers and cellular automata to do so.
But I think an economy or biology can compare with computers or cellular automata, propose some concepts by analogy, and then build a theory.
After reading answers, @Yeu Mak, and @Duan Xian Xiang,
I would like to stress out two issues. Automata can be networked or we can work with agents, and use networked or agent-based modeling. These concepts are very powerful in solving societal, biological, and life-like natural phenomena.
The other thing, entropy-based complexity measures, they are cited in my paper on prediction of arrhythmias, are assesing rte behavioral mode of the system without its detailed understanding.
Hopefully, it will help to make a broader overview of assesing thr mode of bevior of this type of models.
After reading the previous answers, one thing seems to be important to mention. Self-Organization and emergence are playing a great role in description of many complex systems observed in the Nature and even in those man-made.
I recommend you to study those processes in social systems. It gives you a completely different perspective.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology