Relevance Ranking Using LIKER and LIKEP

In addition to the Metamorph searches listed above there is another type of search based on Metamorph. This will return rows in order of relevance, with the most relevant record first. LIKER calculates a relevance based solely on the presence or absence of the terms in the document. LIKEP uses this same information, but also uses the proximity of the terms to calculate relevance.

There are several restrictions and points to note about LIKER and LIKEP. The conditions that must be met to obtain a relevancy search are that a Metamorph index exists on the field in question. LIKER and LIKEP can only work with an index. The other condition is that the query must consist of word terms only. None of the other pattern matchers are available with LIKER or LIKEP.

The query is a list of terms to be searched for. The words are weighted by their uniqueness in the document set being searched. This means that infrequent words are weighted more than common words.

The weight that was calculated for the record is available by selecting the generated field $rank, which will contain the rank value. The rank value for LIKER is unscaled. With LIKEP the number will range between 0 and 1000.

For fuller details on using the logic operators in LIKE see Chapter here



Copyright © Thunderstone Software     Last updated: Tue Feb 21 13:46:30 EST 2017
Copyright © 2017 Thunderstone Software LLC. All rights reserved.