the google similarity distance（谷歌的相似性距离）.pdfVIP

The Google Similarity Distance Rudi L. Cilibrasi and Paul M.B. Vitan´ yi Abstract— Words and phrases acquire meaning from the normalized information distance ( NID ), normalized Google way they are used in society, from their relative semantics distance ( NGD ), meaning of words and phrases extracted to other words and phrases. For computers the equivalent from the web, parameter-free data-mining, universal similar- of ‘society’ is ‘database,’ and the equivalent of ‘use’ is ity metric ‘way to search the database.’ We present a new theory of similarity between words and phrases based on information distance and Kolmogorov complexity. To fix thoughts we I. Introduction use the world-wide-web as database, and Google as search Objects can be given literally, like the literal four-letter engine. The method is also applicable to other search engines and databases. This theory is then applied to construct genome of a mouse, or the literal text of War and Peace a method to automatically extract similarity, the Google by Tolstoy. For simplicity we take it that all meaning similarity distance, of words and phrases from the world-wide- of the object is represented by the literal object itself. web using Google page counts. The world-wide-web is the Objects can also be given by name, like “the four-letter largest database on earth, and the context information entered genome of a mouse,” or “the text of War and Peace by millions of independent users averages out to provide automatic semantics of useful qual