Forecasting Economic Data with Neural Networks
Computational Economics
| Hi-index | 0.00 |
This preliminary work proposes the use of a multi-level Multi-Agent System (MAS) to analyze realistic human social interaction. One of the most relevant innovations of the proposed research would focus on the simultaneous modelling of more than one level of agency, with a MAS conformed of agents that are in turn made of agents. We plan to apply this perspective to the Mexican microfinance industry, which has recently been under scrutiny because of the high interest rates it charges. The set up allows for an appropriate description of individual borrowers as agents in their own right, while at the same time considering groups of borrowers---who officially receive the micro-loans---as agents.