Identifying and modeling the dynamic evolution of niche preferences

Abstract

• We propose the methods to identify and model the dynamic evolution of niche preferences. • Our methods introduce a binary variable to automatically distinguish the general preferences and niche preferences from user-generated content. • Extensive experiments indicate that the proposed methods can effectively identify the niche preferences, model the evolutionary trends, and discover targeted users. The niche market is an appropriate place for the nascent small and medium-sized enterprises (SMEs) to acquire a small market share to expand their operations. Niche preferences serve as the cornerstone for discovering niche markets. However, current research about niche markets ignores the meanings and dynamic evolution of niche preferences, which does not benefit the SMEs in choosing the suitable niche markets and the best timing of entry. In this study, we propose three methods to model and identify niche preference. The niche preference (NP) method introduces a binary variable to automatically distinguish the general preferences and niche preferences from user-generated content. The dynamic niche preference (DNP) method is able to model and identify the evolution of niche preferences incorporating the time element. The online DNP method can handle large text streams of data compared to the DNP. The findings then are used to identify the meaning of niche preferences for the niche market and to model the evolutionary trends. Experimental results in the real-world dataset indicate that the proposed methods effectively identify the niche preferences, model the evolutionary trends, and discover targeted users.