Examining concepts of author disambiguation: co-authorship as a disambiguation feature in EconBiz

Name ambiguity of authors is a long-standing challenge in Digital Libraries. Simple string searches for authors often have unsatisfactory results: publications by the author in which their name is written differently cannot be found, and publications by other authors of the same name get included. The authors can be distinguished with the use of persistent identifers. This improves the search function and also contributes to the data linking process. But many catalogs of libraries are not fully disambiguated and it is not feasible for big libraries like the ZBW - Leibniz Information Centre for Economics to disambiguate the bibliographic records manually. The goal of this work is to help the ZBW with their disambiguation task. For this purpose I analysed the datasets of the ZBW and identifed the challenges that need to be overcome. I proposed an approach which uses already disambiguated records to assign persistent identifer to ambiguous author references. The approach could disambiguate ambiguous author references, though the quality of the method needs to be further evaluated.:List of Figures List of Tables Listings List of Algorithms 1 Introduction 1.1 Aim of the work 1.2 Structure of the thesis 2 Author disambiguation 2.1 Formal defnition 2.2 Types of disambiguation methods 2.3 Types of disambiguation features 2.4 Co-authorship as a disambiguation feature 3 The datasets of the ZBW 3.1 Methodology 3.2 Analysing the Econis data 3.3 Disambiguating names from other sources 4 Current approaches 4.1 Solving the common name problem 4.2 An incremental approach 4.3 Approaches in related fields 5 Proposal of an application-relevant approach 5.1 Disambiguation approach 5.2 Test datasets 5.3 Evaluation of the results 5.3.1 Partly disambiguated bibliographic records 5.3.2 Ambiguous bibliographic records with two or more authors 5.3.3 Ambiguous bibliograpic records with one author 6 Conclusion A Full figures B Extended tables C Full results of the tests of the proposed method C.1 Partly disambiguated bibliographic records C.2 Ambiguous bibliographic records Bibliograph