Relevance Ranking Configuration
By default, ISIS databases return results in MFN order (insertion order). The Relevance Module allows you to sort results based on "weights", ensuring that significant matches appear at the top.
Script: edit_relevance.php
File Created: bases/[db]/opac/relevance.def
1. How Ranking Works
The system calculates a score for each record based on where the search term appears.
- A match in the Title is worth more than a match in the Notes.
To set this up, you don't assign numbers directly. Instead, you map your FDT fields to pre-defined Relevance Categories.
2. Configuration Categories
The interface lists all fields from your FDT. You must map them to one of the following:
| Category | Typical Weight | Usage |
|---|---|---|
| Title | High (e.g., 100) | Use for Main Title, Translated Title, Series Title. |
| Author | Medium-High (e.g., 80) | Use for Main Author, Co-authors. |
| Subject | Medium (e.g., 60) | Use for Keywords, Descriptors, Topographic terms. |
| General | Low (e.g., 20) | Use for Abstracts, Notes, and other full-text fields. |
| -- Not Used -- | 0 | Fields ignored by the relevance algorithm. |
3. Setup Strategy
- Load: The script automatically parses the
[base].fdtfile to list your fields. - Map: Select the appropriate category from the dropdown for key fields.
- Save: The system writes a
relevance.deffile that the search engine reads during execution.
Strategy
Do not map every single field. Focus on the 4 or 5 most important fields (Title, Author, Subject) to keep the ranking algorithm efficient and meaningful.