5. MODELLING NON LINEAR DYNAMICAL SYSTEMS
• Stochastic noise-driven linear and nonlinear dynamical systems.
• Another approach involves a state space where time series observations are
transformed to the phase space vectors. They model the dynamics of a system by
modeling the dynamics of the corresponding points in the phase space using a
Mapping Function
• The Mapping Function can be used for Prediction of future states.
7. ALGORITHMIC METHODS
• As soon as you introduce self-loops in your model structure, you introduce time,
and therefore the system becomes dynamical.
• In many statistical models, we don't have a closed form solution for inference on
particular variables, so we must use iterative/algorithmic methods to approximate
/ sample values.
• Neural Networks play an important role for modelling Dynamic non linear
systems.
8. ARTIFICIAL NEURAL NETWORKS
• Consist of Connected local processing elements (neurons)
• These accept weighted inputs from other such elements
• They use these weighted inputs to give a single output
9. FREEMAN 1991
• Artificial neural systems were designed to capture some of the useful brain
functions by modeling the features of the brain.
• Chaos "may be the chief property that makes the brain different from an artificial-
intelligence machine“.
10. CHAOS
• Chaos is statistically indistinguishable from randomness, and yet it is deterministic
and not random at all.
• A chaotic system will produce the same results if given the same inputs
• But you can not predict in what way the system's behavior will change for any
change in the input to that system
• It is random appearing, and yet has a large degree of underlying order.
11. BUTTERFLY EFFECT
• Affects Initial Conditions
• Initial Condition inaccuracy grows exponentially.
• We get extra ordinary and Counter Intuitive Results
12. CHAOS ENGINEERING
• How Emergent behavior from Component Interactions can cause the system to
devove into Chaotic state,
• Controlled Experiments on Distributed Systems by introducing failure.