Finding High-Quality Web Pages Using Cohesiveness

For a document, cohesiveness is a measure of how closely the concepts in it are related to each other. Previous studies in linguistics have shown that a high quality document is likely to be very cohesive. Similarly, since a web page is a type of document, a high quality web page is expected to be very cohesive as well. Using an ontology constructed from the Yahoo! directory and the web pages linked to it as an underlying reference, we define a distance metric to measure how close two nodes (or concepts) are in the ontology. This metric is used to calculate the cohesiveness of a web page as the total distances of all the concepts in it. Users can use web page cohesiveness to more easily find high quality (cohesive) web pages.