Resilience in the Decision-Making of an Artificial Autonomous System on the Stock Market

• Main Authors: Daniel Cabrera, Rolando Rubilar, Claudio Cubillos
• Format: Article
• Language: English
• Published: IEEE 2019-01-01
• Series: IEEE Access
• Subjects: Resilience; artificial autonomous system; stock market
• Online Access: https://ieeexplore.ieee.org/document/8856190/

This paper presents the design of a resilience mechanism for supporting investment decision-making processes performed by artificial autonomous systems. In the field of Psychology, resilience is understood as the capacity of people to overcome adversity. Resilience has been determined to be a permanent necessary element for the life of an individual. In addition, different levels of intelligence, analysis capacities, and degrees of autonomy have been progressively incorporated within information systems that are oriented to support decision-making processes, such as those for stock markets. Particularly, the inclusion of affective criteria or variables within decision-making systems represents a promising line of action. However, to the best of our knowledge, there are no proposals that suggest the inclusion of a psychological approach to resilience within an autonomous decision-making system for stock markets. Specifically, the incorporation of a psychological approach to resilience allows the autonomous system to face special difficult investment scenarios (e.g., an economic shock) and prevent the system from achieving a permanent negative performance. Thus, psychological resilience can enable an artificial autonomous system to adapt its decision-making processes according to uncertain investment environments. Our proposal conducts experiments using official data from the Standard & Poor's 500 Index. The results are promising and are based on a second-order autoregressive model. The test results suggest that the use of a resilience mechanism within an artificial autonomous system can contain and recover the affective dimensions of the system when it faces adverse decision scenarios.