A Syntactic Classification based Web Page Ranking Algorithm

TL;DR

This paper proposes a web page ranking algorithm based on syntactic classification, which improves search result relevance by considering user preferences and page properties without analyzing content meaning.

Contribution

It introduces a novel ranking method using syntactic classification supported by fuzzy c-means and neural network algorithms, emphasizing user demand-based property weighting.

Findings

Ranking varies with page type for the same query

Syntactic classification effectively differentiates web pages

The approach enhances search result relevance

Abstract

The existing search engines sometimes give unsatisfactory search result for lack of any categorization of search result. If there is some means to know the preference of user about the search result and rank pages according to that preference, the result will be more useful and accurate to the user. In the present paper a web page ranking algorithm is being proposed based on syntactic classification of web pages. Syntactic Classification does not bother about the meaning of the content of a web page. The proposed approach mainly consists of three steps: select some properties of web pages based on user's demand, measure them, and give different weightage to each property during ranking for different types of pages. The existence of syntactic classification is supported by running fuzzy c-means algorithm and neural network classification on a set of web pages. The change in ranking for difference in type of pages but for same query string is also being demonstrated.