摘要：
Introduction: For a broad variety of innovation intelligence tasks, like finding the patent agent being specialised in a technology, specialisation profiles of entities are of particular interest. Such a profile indicates for each entity in which activities it is markedly specialised. Conventional approaches suffer two shortcomings, though. First, they tend to consider only the activity profiles of the respective entities, while neglecting information about all other entities' activities. Second, they often lack in considering entity specific characteristics, when assigning an activity to an entity's specialisation profile or not. Method: To address these shortcomings, we introduce the RSI-specialisation, a novel method for retrieving dichotomous and idiotypic specialisation profiles. Operationalising relative specialisation, the novel method is rooted in the theory of comparative advantage. In order to contrast it with the approaches based on absolute specialisation, it is compared with the baseline method ENF-specialisation, which has its theoretical roots in the concept of effective number of components. Analysis: Both methods are demonstrated by applying them to the case of the specialisation of patent agent firms. RSI-specialisation and ENF-specialisation are applied to a data set containing all EPO patent applications in 2014 and 2015. Results: Compared to the baseline approach, the RSI-specialisation reduces noise from market effects to a greater satisfaction. Besides being less dependent on agent size and market structure, it reduces interpretation to the most essential question for an applicant, i.e. why he should opt for one rather than another provider. In addition, it also guarantees that a potential applicant can find in any field a specialised patent agent firm. Conclusion: We find that the novel RSI-specialisation promises to be a robust and reliable method for retrieving dichotomous and idiotypic specialisation profiles. Patent agent firms aside, it
欢迎来到三峡大学图书馆！
详细信息
中文科技期刊
