Result Relevance and Ranks

See Also: Adjust Query Results

By default, IDOL Server returns query results in relevance order. It uses a number of methods to calculate this relevance.

Relevancy Calculation

The core relevancy calculation takes into account the APCM weight (statistical weight) assigned to each of the matched terms, and the number of times those terms occur in the result document.

Matching more of the terms generally results in a higher ranking than matching a few of the terms multiple times. Additional factors such as capitalization, stemming, and proximity weighting also adjust the weighting (but again, normally less than matching an additional term).

The advanced search modes (AdvancedSearch, AdvancedCaseSearch, or AdvancedPlus) change the default query operator between terms to use WNEAR where no query operator is specified (rather than OR in the normal search mode). In this case, documents matching the query terms in proximity to each other receive a weighting boost, with a larger boost if consecutive query terms appear as a phrase in a document. This process is distinct from an exact phrase search (quoting the query terms as a phrase), which returns documents only if they match the exact phrase specified.

Field Weights

You can configure a specific weight to apply an automatic boost factor to all fields with that assigned property.

You might choose to weight matches in the title field higher than other index field matches.

You configure the boost by setting the Weight configuration parameter as a field property. For more information, refer to the IDOL Server Reference.

You do not need to reindex data for this configuration to take effect. This method always applies to every query, so its effect is not dynamically adjustable.

Normally, if a term occurs more than a certain number of times in a result, the weight is saturated. For example, a term occurring 100 times in one result weights as much as the same term occurring 110 times in another result. If this saturation is undesirable then you can set a fractional weight for all fields (for example, 0.1) to raise the effective saturation limit.

NOTE:

Take care to avoid saturating the weights when using explicitly configured weights.

Explicit Term Weights

Explicit term weights give you the most control over relevancy ranking, because you can adjust them dynamically for each query (at the expense that the query Text must either have the explicit weights present or have them injected by application code). As with the index Weight field property, you must take some care to avoid saturating the weights.

To use term weights, you add the weight in brackets at the end of the term, without a space. For example dog[50] applies a weight of 50 to the term dog.

Custom Term Weight Files

A custom term weight file defines a set of terms and weights to use during relevance weight calculations at query time. By default, Content calculates the term weights at index time, and stores the list of weights.

In some cases, you might want to use a custom term weight file for extra control over the weighting. In particular, you might want to use a custom term weight file in a distributed system to ensure that all child servers use the same weighting, to ensure consistent results and relevance weighting across the indexes.

You can add, modify, and delete a custom term weight file by using the DREMODIFYTERMWEIGHT index action. For more information, refer to the IDOL Server Administration Guide.

AutnRank

You can give documents an associated AutnRank that arbitrates between two documents that would otherwise return the same relevancy for a particular query. The document with the higher AutnRank value is given a higher relevancy score. You can use AutnRank to assign a measure of absolute importance to a document.

You might give a corporate home page a higher AutnRank than a page several clicks into the site.

You should limit AutnRank values to the range 0 to 4095.

You can adjust the AutnRank by using either the DRECHANGEMETA or DREREPLACE index actions, allowing you to use an application to alter the ranking over the lifetime of the document (for example, to reflect an average user rating system).

BIAS

You can use the BIAS family of FieldText operators to adjust result ranking at query time, based on the values in the specified fields. Picking the right amount to boost relevancy by requires some experimentation, and in general relevancy scores are still dominated by the conceptual matching.

If a BIAS operator references multiple fields, IDOL Server uses the field value that results in the largest boost to calculate the final relevancy adjustment.


_FT_HTML5_bannerTitle.htm