MultiView-Systeme zur explorativen Analyse unstrukturierter Information

Information research tasks often require condensed information from different information sources. Patent analysis, for instance, requires the combination of structured data like patent classes or company names with unstructured text documents like patent abstracts or claims. Due to missing IT support, the combined analysis of structured and unstructured information often remains a manual and intellectual task. This work motivates, develops and evaluates an approach for interactive text and data analytics. The goal of this approach is to offer concise exploration of the task-relevant information and relationships in a single system. Information from text documents, categories and relational data is presented in graphical views which are interactively coupled so that information relationships might be used for navigation among views. The system SWAPit implements the system concept. One conceptional-technical challenge has been the adaptability of the reference technology for different application domains, business processes, and tasks. Therefore, the methods for tailoring the system and for measuring usefulness are considered important contributions by themselves. SWAPit has been evaluated and optimized in a broad spectrum of industrial case studies from the fields of Business Intelligence (BI), Customer Relationship Management (CRM), and Cooperative Work (CSCW).