Semi-automatic Term Extraction with Simplified Term-clips Method: Development and Applications of a JavaScript Component

Humanists often have their own digitalized texts. Sometimes they want to extract as many terms of specific type as possible from the texts. Term extraction methods are computational algorithms to extract meaningful terms from a large corpus of digitized texts. Term-clips method is a semi-automatic term extraction algorithm that highly demands human-computer interaction to get comfortable extraction results. The algorithm of term clips was developed years ago, and it was shown to be powerful in extracting categorical terms from Chinese ancient novels. However, perhaps due to the complicated original algorithm, or maybe due to the lack of user-friendly interface, this method does not used widely by humanists. In this paper, we shall propose a simplified term-clips method. We emphasize the importance of human interaction to select effective clips and desirable terms. We adopt this algorithm to implement a JavaScript component, and use this component to develop a real system that allows people to extract meaningful terms from their own texts. Encapsulating algorithms into components is helpful to the development of more complicated systems, since system developers do not need to implement the algorithms themselves. Developers can simply include the components and reuse the functions provided by the components. This term-clip component has been used by MARKUS, an open source system developed for taking markups on Chinese ancient texts. In addition, we shall discuss considerations in the design and implementation of the JavaScript component. We show how to use this component to make a simple application, and develop a complete system that helps people extract meaningful terms from their own texts. We illustrate how to use the system by the extractions of personal names and clothing terms from the Chinese ancient novel Dream of the Red Chamber.