Extending WHIRL with background knowledge for improved text classification

Abstract

Intelligent use of the many diverse forms of data available on the Internet requires new tools for managing and manipulating heterogeneous forms of information. This paper uses WHIRL, an extension of relational databases that can manipulate textual data using statistical similarity measures developed by the information retrieval community. We show that although WHIRL is designed for more general similarity-based reasoning tasks, it is competitive with mature systems designed explicitly for inductive classification. In particular, WHIRL is well suited for combining different sources of knowledge in the classification process. We show on a diverse set of tasks that the use of appropriate sets of unlabeled background knowledge often decreases error rates, particularly if the number of examples or the size of the strings in the training set is small. This is especially useful when labeling text is a labor-intensive job and when there is a large amount of information available about a particular problem on the World Wide Web.

Determining if two documents are written by the same author

“…Because we have seen that we have a reasonably effective solution to the many‐candidates problem, we can use impostors to reduce the verification problem to the many‐candidates problem. The use of impostors as a background set is a well‐established practice in the speaker‐identification community (e.g., Reynolds, ) and has also been applied to information‐retrieval problems (Zelikovitz, Cohen, & Hirsh, ), but, as far as we know, has not been previously used for authorship attribution.…”

An unsupervised transfer learning approach to discover topics for online reputation management

“…The semantics could be derived from a collection of much longer documents in a similar domain as the short texts [14], by sending the input short texts as queries to a search engine to retrieve a set of most relevant results [11] , or by exploiting external resources such as Wikipedia 1 [9] and WordNet 2 [5]. Once the auxiliary data is obtained, it is often used to expand the original texts, which are then processed by traditional text mining models.…”

Short text classification using very few words

“…Most existing approaches try to enrich the representation of a short text using additional semantics. The semantics could be derived internally from the short text collection [3], externally from a collection of much longer documents in a similar domain as the short texts [5], or from much larger external sources such as Wikipedia and WordNet [1,3,4]. In [1,4], the classification accuracy is significantly improved by enriching short text feature vector with relevant hidden topics derived from Wikipedia pages using topic model.…”