COLLABORATIVE SYSTEMS - SOLVING THE VOCABULARY PROBLEM

Vocabulary differences have created difficulties for on-line information retrieval systems and are even more of a problem in computer-supported cooperative work, where collaborators with different backgrounds engage in the exchange of ideas and information. In almost all computer applications, users must enter correct words for the desired objects or information. Yet studies have found that in spontaneous word choice for objects in five domains, two people favored the same term with less than 20 percent probability. Can computer technologies and information system be designed to help alleviate the vocabulary problem and foster seamless collaboration? This article examines the vocabulary problem in CSCW and suggests a robust algorithmic solution. Using various information science, statistical, and artificial intelligence techniques, the author has developed an algorithmic approach for creating a vocabulary-rich dictionary/thesaurus called the concept space. This concept space is generated by extracting concepts (terms) automatically from the texts produced during collaboration and then linking them through co-occurrence analysis. In contrast to the previous manual techniques for solving the vocabulary problem (for example, unlimited aliasing), this technique automatically extracts vocabularies from text, identifies vocabulary similarities, and groups similar vocabularies together. This computationally intensive approach helps alleviate some cognitive burden of collaborators while consolidating different vocabularies.