Modified Breadth-first order-based link categorization for finding financial statement documents

With the exponential growth of information on the World Wide Web, there is a challenge to find for a document that contains specific information on the Internet. There are many statistical documents that are available on the Web. However, to search, recognize and take these kinds of documents need much effort and also require much time. One of the solutions that can be used to do that is a web crawler. In this paper, we develop a method to search statistical documents, specifically on finance domain using a web crawler based on link categorization. The method is a modified form of Breadth-first ordering strategy, where every link that is found will be categorized into three groups: Positive link, Negative link and Neutral link. The objective is to identify whether the link is likely pointing into relevant documents or not. In addition, we also utilize keywords related financial domain to recognize a relevant document. Based on our experiments, the value of precision and F-measure score of the proposed method is higher than its baseline value.