Search Algorithm
AddSearch scores each document using multiple factors to determine its ranking. You can manually adjust ranking using available tools.
Ranking Factors
The following elements influence how documents rank in search results:
- Self-learning algorithm: Boosts scores for documents with higher click-through rates based on user behavior.
- Sorting criteria: Ranks documents according to defined categories such as date, relevance, or custom fields.
- Pinned results: Allows manual placement of specific documents for given keywords to ensure they appear prominently.
- Site areas: Applies positive or negative boosts to specific sections of your website to emphasize or de-emphasize content.
- Phrase match: Gives higher scores to documents containing exact matches of the search phrase, while assigning lower priority to partial or weaker matches.
- Field weights: Increases the score of documents with keyword matches in title or H1 headings compared to matches found only in the main content. You can adjust these weight settings manually.
Search Features
The AddSearch algorithm includes these capabilities to improve search experience:
- Typo tolerance: Corrects misspelled keywords using fuzzy matching (e.g., a search for "contat information" returns results with correctly spelled "contact").
- Exact matches: Enclosing a search phrase in quotation marks returns only documents containing that exact phrase.
- Logical operators: Use operators like
AND,OR, andNOTto refine queries. For example,green shirt OR jacketreturns results containing either green shirts or green jackets. - Bigram matching and tokenization: Matches compound keywords and variations regardless of spacing or punctuation. For instance, the URL
/contact-information/matches queries for "contact-information," "contact information," and "contactinformation." - Stemming and plurals: When language-specific stemming is enabled, singular and plural forms match. For example, "phones" matches "phone," and "filters" matches "filtering."
- Synonyms: Define synonyms globally or for specific documents so that related terms (e.g., "coat" and "jacket") match searches interchangeably.
- Partial Match / Wildcards: Returns results that partially match a search term, such as "auto" matching "automatic."