A General Framework for Information Processing with Application to Quantitative Software Testing

We propose in this paper a general framework for collecting and processing data from various sources of information. An advantage of our framework over traditional data mining frameworks is the ability to collect nuggets of information from sources encoded in different formats. We can handle documents or fragments of knowledge written in natural language (open structure documents), XML-like data (semi-structured documents), and database repositories (structured documents). Our proposed framework leads to an integrated technology for collecting data from these various sources as opposed to a technology in which there is a set of isolated components that deal with format-specific