A Taxonomy Management System Supporting Crowd-based Taxonomy Generation, Evolution, and Management

Information overload continues to be a challenge. With the growth of the internet and the birth of social media, the volume of information increases exponentially. By dividing the material into many small subsets, data, information, and content classification based on a taxonomy makes information exploration and retrieval faster and more accurate. Instead of having to know the exact keywords that describe the knowledge resource, users can browse and search for them by selecting the descriptive categories that the resource most likely belong to. Furthermore, a taxonomy supports consistency in classification and navigation in complex websites. Nevertheless, developing taxonomies is not an easy task. It requires authors to have a certain amount of knowledge in the domain. The workload will always increase as any new taxonomy needs to be frequently updated to remain relevant and useful. The collaboration process, where many people work together in the development, evolution, and management of taxonomy, will potentially generate problems as working in a group requires soft skills and patience. While growing, a taxonomy needs more computing time, memory, and persistent storage space. Without an effective and efficient method, suitable representation models, and a good user interface, a taxonomy cannot be easily modified, quickly processed, and well represented. To combat these problems, many approaches were introduced in several scientific disciplines. A common approach is to use social tagging, document keywords, or words that are used frequently in documents, to form a term corpus. Then applying some types of hierarchical clustering algorithms or machine learning to organize this corpus into hierarchical subgroups. While these methods provide many advantages and good results in building a taxonomy, they also have their disadvantages and prerequisites in order to work properly. In this dissertation, a new method using collaboration, crowdsourcing, and crowdvoting has been proposed. With support of crowdsourcing, not only experts and knowledge workers but also normal users can participate in the development, evolution, and management of taxonomy. This helps reducing the maintenance cost of taxonomy. Furthermore, a version control component has been developed to support users in these processes. The user-centered design with four steps has been applied to design a conceptual system. From user requirements and use case diagrams, five components have been introduced. Their purposes are to enables the crowd to create, edit, upgrade, and vote for taxonomies in an easy-to-use, effective, and efficient manner. A prototype has been developed as an open-source web application based on client-server architecture and integrated into the Content and Knowledge Management Ecosystem Portal. Finally, two evaluations have been organized in the context of EU-funded R&D projects. The first evaluation asked experts and the crowd to do similar tasks. The results then were compared to validate if the crowd has really done a similar good or even better job in taxonomy development than the experts. The second evaluation asked two groups of participants to complete an assignment and answer a questionnaire. Both evaluations produced a good result, which, on one hand, demonstrated the feasibility of the approach and the usability of the initial prototype, and on the other hand, validated the quality and effectiveness of the chosen method. Furthermore, the implemented prototype has been productively applied in different application domains with different requirements in the context of several scientific and technical projects, which have been funded by the European Commission. This proofed the generality and adaptability of the developed system prototype to new application and knowledge domains.