HomeSEO TheoryBlogContact

Analyzing Keyword Match Types for Document Ranking

Keyword matches come in various forms and are an important metric for determining the relevance of a search query in relation to a document or webpage. This page will analyze the various match types and determine an effective way of giving value to strings of keywords that match a user's query in different ways.

Page Content
Metadata and Page Details
Metadata and Page Details
Creator: Devin Peterson
Date: Created 02/04/2014
Subject: Keywords, Word Matching, SEO
Publisher: DNM Int'l
Contributors:
Peer Review:
Resources:
Citation: Peterson, D. (2014), "Keyword Matching: Document Ranking Based on Word Match Conditions", Retrieved (date), from http://seowrit.com/keyword-match-types

4 Types of Keyword Matches

Broad Match

This is a broadly matched search query.

Phrase Match

This is a search query that matched in part.

Exact Match

This is an exact matched query.

Why Keyword Matches Matter for Determining a Document's Relevance.

Coming Soon...

The Keyword Match Ranking Function

An anecdotal analysis can provide some insight as to why various types of matches matter more or less than others. The exact degree will always be subjective and probably be tweaked from now til the end of time. Perhaps an algorithm can compute different analysis of keyword matches based on the actual search query topic. The problem in measuring match types and assigning relevancy values is not in the lack of computing abilities, but rather in the fact that often times a searcher isn't quite sure what they are searching for and may rely on the search engine for delivering results that are unexpected but satisfactory.

The creation of a ranking function based on keyword matches relies heavily on vector analysis. A dose of matrix operations are implemented to determine how much particular keywords are valued based on the match type. An understanding of dot products and cross products are required to fully grasp the keyword match type ranking function.

Let's start with a simple example where documents are ranked based on one concept. The degree to which the search query matches the text in the documents. In this example, we shall ignore keyword frequency and all other potential ranking factors. We shall ignore the case of a tie, where 2 results share the same match.