FUZZY CLUSTERING AS AN ALTERNATIVECLASSIFICATION METHOD IN INFORMATIONSYSTEMS RESEARCH

Cluster Analysis is a powerful exploratory method that has been used in pattern analysis, grouping, decisionmaking, and pattern classification. Cluster analysis is particularly appropriate for the exploration of interrelationships among the data points to make an assessment of their structure. General tasks in cluster analysis (Jain, Murty, and Flynn 1999) involve pattern representation, definition of pattern proximity, clustering or grouping, and assessment of out. Within the broader scope of information systems research, cluster analysis has been used in the area of end-user computing, assessment of IS strategies, global information systems research and assessment of supply chain configurations. In most cases, researchers have adopted hard or crisp cluster analysis whereby each respondent or analysis unit is assigned a class label, or becomes a part of one and only one cluster (unitary membership). In several cases, unitary membership may not reflect situation realities. An alternative clustering methodology – fuzzy clustering provide an alternative analytical mechanism to unitary membership by allowing gradual membership in different groups, with membership values indicating the probability or degree of membership within a specific group or cluster. We illustrate fuzzy clustering with an example from the field of end-user computing and then develop general guidelines for the broader use of the method.