No abstract

This paper presents a framework system, called ProcessSearch, which is used to search and identify business process descriptions stored on the web. Some metrics were proposed in order to identify the web documents describing a business process. Identifying descriptions of business processes is a difficult task because the web stores "bag-of-words" documents. Hence, the ProcessSearch framework has been constructed and implemented, using known technologies in information retrieval and computational linguistics. The result was a satisfactory list of web documents ranked (top-k) by relevance for some given keywords. For validation, a "precision at k" analysis was used to confirm the accuracy of the proposed metrics.