THE IN-COGNITIVE APP

The InCognitive App provides a GUI and backend code for Fuzzy Cognitive Map (FCM) simulations.

THE IN-COGNITIVE APP FUNCTIONALITIES

• Deployment and graphical representation of a FCMs.

• Selection of FCM-transfer-function's parameter lambda; it is based on the methodology of [1]. The yielded lambda value guarantees that the stand-alone FCM or FCM-MC simulations (see next functionalities) do not yield chaotic or indefinite oucomes.

• Stand-alone-FCM simulation, based on [1] and [2]. It provides the final nodes' state vector, given predifined input-node values and FCM parameters (e.g. FCM-transfer-function's parameter lambda).

• A combination of FCM and Monte Carlo (MC) simulation. This combination, FCM-MC, analyses the uncertainty propagation from input nodes all the way to the output nodes, in case the "weights" of FCM edges and/or the input node values are random variables. It's outcome is the final node value distributions, given that some, or all, of the input nodes and/or the FCM parameters (e.g. weights) are random variables.

EXECUTION

To run the application, the user needs only to execute the main.py module. The GUI provides all necessary interaction between the end-user and backend-code. For further details, see the manual here: https://github.com/ThemisKoutsellis/InCognitive/wiki/Manual

REFERENCES

[1] T. Koutsellis et al., "Parameter analysis for sigmoid and hyperbolic transfer functions of fuzzy cognitive maps," 2022, Oper Res Int J, 22, pp. 5733–5763, https://doi.org/10.1007/s12351-022-00717-x

[2] T. Koutsellis et al., "Normalising the Output of Fuzzy Cognitive Maps," 2022 13th International Conference on Information, Intelligence, Systems & Applications (IISA), 2022, pp. 1-7, doi: 10.1109/IISA56318.2022.9904369.