Simulating Language Dynamics by Means of Concept Reasoning

Abstract

Abstract.A problem in the phenomenological reconstruction of Complex Systems (CS) is the extraction of the knowledge that elements playing in CS use during its evolution. This problem is important because such a knowledge would allow the researcher to understand the global behavior of the system [1,2]. In this paper an approach to partially solve this problem by means of Formal Concept Analysis (FCA) is described in a particular case, namely Language Dynamics. The main idea lies in the fact that global knowledge in CS is naturally built by local interactions among agents, and FCA could be useful to represent their own knowledge. In this way it is possible to represent the effect of interactions on individual knowledge as well as the dynamics of global knowledge. Experiments in order to show this approach are given using WordNet.

Synthetizing Qualitative (Logical) Patterns for Pedestrian Simulation from Data

“…Since our model works with knowledge-based agents, it could be extended, by joining knowledge from distinct agents, in order to work in the problem of collaborative localization [18]. Stem (and Luxenburger) basis of different agents can be re-interpreted in order to fuse their knowledge base by conciliation (as in [6]) or by considering a common (global) Knowledge [4]. In general terms, the integration of different knowledge elements within FCA can pose some issues on the extensibility and scalability of the model, which could be solved by means of discretization methods, multi-valuation and scales from FCA [12].…”