Skip to content

Neural network that has been trained to detect temporal correlation and distinguish chaotic from stochastic signals

Notifications You must be signed in to change notification settings

brunorrboaretto/chaos_detection_ANN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 

Repository files navigation

chaos_detection_ANN

A neural network that has been trained to detect temporal correlation and distinguish chaotic from stochastic signals.

The '/auto_ANN_Omega/' directory depicts the fully automatic code with the necessary libraries.

The main file 'chaos_detection_ANN.py' contains all the information.

The files: W1.dat', 'W2.dat', 'B1.dat', 'B2.dat' are the weights of the ANN. 'colorednoise.py' is the library to generate the flicker noise (colored noise).

Instructions for running the code:

python chaos_detection_ANN.py serie.dat

'serie.dat' is the time series to be analyzed. The code compares the time-series with 1 flicker-noise time-series with the same correlation coefficient (predicted by the ANN) and the same length. For small time-series length<1000 points we suggest the command:

python chaos_detection_ANN.py serie.dat 10

In this case, the code compares the time-series with 10 flicker-noise time-series.

The 'tests' directory presents an autorun of the Figure 3 for a practical use, with fewer points (101 initial conditions insted of 1000) and less precision (length 2^16 instead of 2^20).

Instructions for running the code: python autorun3.py

After a few minutes the figure 'test_fig3.png' is generated.

Releases

No releases published

Packages

No packages published