Relevance Ranking

The most powerful weapon in the PLWeb Turbo arsenal is relevance ranking. Simply put, relevance ranking arranges a set of retrieved records so that those most likely to be relevant to your request are shown to you first. That is, after PLWeb Turbo retrieves all documents that satisfy your search query, it uses relevance ranking to arrange them based on a measurement of similarity between your query and the content of each record.

What determines the likelihood of relevance? PLWeb Turbo performs a content analysis of records in your database by using a combination of the following indicators:

Breadth of Match. The more distinct query terms that appear in a document, the higher the weight of relevance.
Inverse Document Frequency. Rare terms (within the entire database) receive a higher weight of relevance.
Frequency. The number of times a query term occurs in a document.
Density. The comparable length of retrieved documents.

Consideration of these combined criteria produces intelligent on-the-fly evaluation of a record's likelihood of satisfying the intent behind your query.

This allows you to find more relevant information with less effort. Regardless of how many records your search query retrieves, you will have to review relatively few of them, because moving down the ranking means moving toward less relevant records. With relevance ranking, you will spend less time reviewing search results before deciding whether they are satisfactory.

Additionally, you are free from the burden of composing complex logical queries, which are used to reduce the amount of retrieved data to manageable proportions. You don't have to care about how many records are retrieved, as long as you know that the best information floats to the top.

Note: Now you can see why stopwords exist. Because words like the and of are so commonly used in English, their presence could artificially boost the relevance of what are actually irrelevant documents. Stopwords also provide a speed advantage, since a search for the would probably retrieve every record in a database.

[Previous Topic] [Contents] [Next Topic]