Fuzzy cognitive maps based on text analysis for supporting strategic planning

Abstract

Strategy maps are attracting considerable interest in strategic planning due to their capacity to represent causal-effect relationships among the key concepts. Several studies have used expert estimates to quantify the relationships. However, these evolve dynamically and are context-specific. Therefore, there is a need to develop automatic knowledge acquisition systems. Here, the assumption was used that knowledge can be extracted from strategic documents in order to conduct a detailed analysis of causal strategic concepts. First, latent semantic analysis is employed to obtain an interpretable semantic model. Second, collocated causal concepts are used to model relationships among strategic concepts. This approach to generate fuzzy cognitive maps (FCMs) is semi-automatic, requiring theoretical background literature/domain experts to determine the direction of the causalities. The FCMs can subsequently be used to simulate the effects of strategic management and, thus, provide an effective decision support tool. Several innovation strategies of regions are used as a case study and it is demonstrated that the generated FCMs are consistent with expert opinions and fuzzy ANP method.