A sequential pattern mining approach to identifying potential areas for business diversification

Although many quantitative models have been presented to identify potential areas for business diversification, most have focused on the assessment of technological capabilities and/or similarities using patent information. New data sources and scientific methods have thus seldom been addressed. We propose a sequential pattern mining approach to identifying potential areas for business diversification using the historical business segment data. Our approach includes (1) sequential pattern mining to identify potential areas for business diversification by extracting the significant changing patterns of firms’ business segments; and (2) index analysis to assess the market and financial characteristics of the areas identified. Taken together, three diversification strategy maps are developed to provide comprehensive views of analysis results. An empirical analysis of 25,126 unique firms with 1320 business segments confirms that the proposed approach enables a wide-ranging search for potential areas for business diversification and the quick assessment of their characteristics.