Parameter Estimation via Analysis of Fuzzy Clusters (PEAF): An Algorithm to Estimate Parameters of Agent-Based Models

Biologically focused, agent-based models need many parameters in order to simulate system dynamics. It is often essential to explore the consequences of many parameter vectors before satisfactorily representing phenomena. In this work we propose a simple algorithm based on fuzzy clustering to estimate model parameter values for new situations utilizing the characteristics of previously simulated conditions. The estimated parameters can be used to predict the behavior of the system in a new situation. Using limited data, we successfully applied the algorithm to estimate parameter values of an agent-based model of hepatocytes (liver cells). Predictions provide acceptable correlations with observed values (p < 0.05, R2 = 0.65).

• Type: Conference
• Year: 2007
• Where: CAINE
• Authors: Shahab Sheikh-Bahaei, C. Anthony Hunt
• Keywords: Agent-based Model, CAINE 2007, Computer Science, Parameter, Parameter Values