The analysis of job requirements is crucial for several stakeholders. Companies define current and future personnel needs through job requirements and job seekers demand current information on job requirements to plan their careers. So far, job requirements were analyzed by applying hypothesis-testing approaches from the social sciences such as questionnaires or structured interviews. Moreover, as large numbers of data on job requirements are available through online job advertisements, also data-driven approaches were used such as keyword searches by pre-defined keyword lists. However, all of these methods can bias the research results as the researchers’ opinions directly influence the expected outcomes. Therefore, the thesis deals with developing and evaluating an approach that enables independent analyses of job requirements. Hence, the web content mining process is implemented as a hypothesis-generating approach and is derived from the methods of big data analytics and knowledge discovery in databases. It combines methods from data mining, web mining, and natural language processing, and is especially adapted to the automatic analysis of job requirements in large numbers of online job advertisements. The web content mining process is based on the analysis of n-grams, word co-occurrences and their relative context information in online job advertisements. The process is applicable when no prior information on job requirements is available. The results show that the web content mining process helps to discover information on job requirements that other approaches do not provide such as information on non-academic qualification requirements and professional experiences. Thus, it independently discovers patterns in data without requiring any pre-defined research hypotheses. It is a valuable research method that complies with the general quality criteria of objectivity, validity, and reliability. When combined with a matching and recommendation component, the web content mining process can be integrated into an overarching recruiting 4.0 framework. Such a framework uses several software components to support decision-making in the recruiting process based on big data analytics. Future research shall concentrate on adapting the web content mining process to real-time data and to define the interfaces that are necessary to build the recruiting 4.0 framework.
Web Content Mining for Analyzing Job Requirements in Online Job Advertisements
The analysis of job requirements is crucial for companies and job seekers. The thesis deals with developing a web content mining process for analyzing job requirements in online job advertisements. It combines methods from big data analytics, knowledge discovery in databases, data mining, web mining, and natural language processing. In the future, the web content mining process can be integrated into an overarching recruiting 4.0 framework to support decision-making processes.