Determining influential factors and challenges in automatic taxonomy generation: A systematic literature review of techniques 1999-2016

Introduction. Taxonomy is an effective mean of managing and accessing a large amount of digital information. Various techniques have been developed to generate taxonomy automatically. The purpose of this study is threefold: (i) review methods and approaches adopted during taxonomy generation, (ii) identify the factors influencing the choice of a particular method or approach, (iii) highlight issues and open challenges. Method. This paper adopts a systematic literature review approach proposed by Kitchenham, and the nature of this review is qualitative. Analysis. A total of thirty techniques were reviewed and categorized into various categories and subcategories. An in-depth analysis of the existing techniques was performed based on this categorization. This ultimately helps in identifying factors influencing the choice of a particular method or approach, and also determines issues and challenges associated with the automatic taxonomy generation. Results. Four major factors influencing the choice of a particular method or approach for generating taxonomy have been identified. Moreover, five major challenges associated with the existing automatic taxonomy generation techniques have also been highlighted. Conclusions. This paper presents a comprehensive review of taxonomy generation so that taxonomy can be used effectively, and it highlights open challenges for future research in the area of taxonomy generation so that new and improved techniques can be developed.