Fuzzy search: smart search with synonym handling and user error correction

What this solution does

The module extends the standard Bitrix Search functionality by adding support for handling typos, input errors, and synonymous queries. Its efficiency is achieved through a combination of fuzzy string search algorithms and computational linguistics methods.

The fuzzy search module provides intelligent search with synonym support and correction of user input errors.

Expansion method. A set of variants containing potential errors is generated from the original query, followed by an exact search in the dictionary for each variant.

The fuzzy search module provides similarity-based search with synonym support and user input error handling. The system accounts for input variations to improve result accuracy.

The fuzzy search module enables searching through a synonym dictionary while accounting for potential input errors and semantic proximity of queries.

The fuzzy search module processes user queries, accounting for typos and synonyms. The system analyzes search query statistics and provides solutions for cases where no exact matches are found.

Installation and basic configuration of the module do not require specialized knowledge. The parameter system allows flexible adaptation of fuzzy search for complex scenarios, including typo handling and synonym recognition. All available options are described in detail in the documentation.

The module includes components catalog.search, search.title, and search.page. In fuzzy search mode, they retain the functionality and templates of standard components with the same names.

Module advantages

The module combines multiple fuzzy search algorithms to enable efficient searching with synonym support and typo correction.

The fuzzy search module provides intelligent search with synonym support and resilience to input errors. The system analyzes search query statistics to improve result accuracy.

The fuzzy search module handles user typos and supports synonym matching. It is fully compatible with Sphinx.

The Tega fuzzy search module features a flexible configuration system. This allows tuning the search algorithms for specific tasks, ensuring accurate synonym recognition and proper handling of user typos without requiring deep core reconfiguration.

The module allows restricting the scope of fuzzy search. This enables using search with synonym handling and input error correction only in selected system sections.

Supports both UTF-8 and Windows-1251 encodings.

The fuzzy search module provides intelligent search with synonym support and user input error handling. The built-in dictionary contains over 80,000 synonyms.

The fuzzy search module provides intelligent search with synonym support and correction of input errors. The documentation includes detailed instructions for installation and configuration.

The fuzzy search module provides intelligent search with synonym support and resilience to user input errors. The solution is ready for installation.

The Fuzzy Search module settings are located in the "Product Settings" > "Module Settings" section.

Initiate the reindexing process.

When similarity search mode is enabled, replace the standard components bitrix:catalog.search, bitrix:search.title, and bitrix:search.page with the corresponding components from the Tega module.

Software requirements

The fuzzy search module from Tega provides intelligent search with synonym support and correction for user input errors. It requires PHP version 5.6.0 or higher to operate.

The fuzzy search module by Tega provides intelligent search that accounts for typos and synonyms. The required MySQL database version is 5.6 or higher. This version represents a tested configuration. If your hosting environment does not meet these requirements, contact technical support to verify the module's compatibility with your specific software.

The mechanism works by artificially expanding the standard Bitrix search index. The system generates a set of words containing common input errors and optionally includes synonyms. Consequently, if a user enters a query with a typo, the system automatically returns relevant results provided that variation was included in the expanded index. This mode operates without requiring modifications to standard components.

The module's functionality is illustrated using the indexed word 'dress' as an example.

The module implements similarity search to handle erroneous queries. When a string is entered, the system sequentially compares it against indexed data to find the closest matches based on Levenshtein distance. Additional comparison methods include stemming (morphological analysis) and the Metaphone algorithm (phonetic pronunciation). This mode operates using both the standard Bitrix index and the module's own indexing tables. Activating this feature requires replacing the standard components bitrix:catalog.search, bitrix:search.title, and bitrix:search.page with the corresponding module components.

The Fuzzy Search module extends the search query statistics collection functionality available when the 'Collect search phrase statistics' option is enabled in the Search module. Collected data is accessible on the Settings > Search > Statistics > Phrase List tab. The module analyzes queries that yielded no matches. Queries are split into individual words, and their morphological roots are identified. During site data indexing, these roots are compared with the roots of existing indexed words. If the roots are similar, they are added to the search index.

Example of operation: a user enters a query with a typo, 'ftstbolka' (transliterated Russian). Since no exact matches are found, the query is sent to the module's special table. The word 'futbolka' (transliterated Russian) is already indexed on the site. Because the morphological root of 'futbolka' is similar to the root of 'ftstbolka', the index is expanded. Consequently, when 'ftstbolka' is entered again, the system returns data corresponding to the index for 'futbolka'.

Fuzzy search functionality is enabled via the 'Add search phrases to the index' option in the 'General settings' tab of the 'Fuzzy search' module. The feature requires the 'Collect statistics on search phrases' option in the 'Search' module to be enabled.

The system operates in Mode I 'Sample expansion' and Mode II 'Similarity search'. Comparison of morphological stems follows the general rules defined in the 'Fuzzy search' module settings.

Processed search phrases can be viewed and managed on the 'Settings > Fuzzy search > Processed search phrases' tab. Users may deactivate specific phrases; deactivated data is excluded from indexing.

Synonym dictionary

The 'Fuzzy search' module supports synonym search. Synonyms are added to the search index if the morphological stem of a word in the index matches the stem of a word in the synonym table.

The Fuzzy Search module includes an extensive synonym dictionary that users can extend with custom entries. The full list of available synonyms is displayed in the Settings menu under Fuzzy Search > Synonym Dictionary.

The functionality operates as follows: if the search index contains the word futbolki (the plural form of the Russian word for T-shirts), and the dictionary has an active entry for futbolka (the singular form) with synonyms rubashka (shirt) and mayka (tank top), the search results will expand. This occurs because the morphological bases of futbolki and futbolka match, allowing the system to include results for the synonyms.

Synonym processing is controlled by the Add synonyms to index option in the General Settings tab of the Fuzzy Search module. Enabling or disabling this option activates or deactivates synonym handling.

Each dictionary entry has an activity flag. Synonyms with this flag disabled are excluded from the search process.

Administrators can configure the system to use only custom synonyms by excluding the default module dictionary. To do this, enable the Use user synonym dictionary option and disable the Use default synonym dictionary option in the General Settings tab of the Fuzzy Search module. Conversely, disabling the user dictionary option while keeping the default active excludes custom synonyms.

The module handles input errors caused by incorrect keyboard layouts. This functionality is inherited from standard Bitrix components and is enabled via the 'Enable automatic keyboard layout detection' parameter. This allows the system to correctly process queries typed with the wrong layout, such as converting 'gkfnmt' to 'dress'. Additionally, the module settings include a phonetic similarity comparison mode. This feature ensures correct processing of phonetically similar queries across different languages, such as recognizing 'platie' as 'dress'.

When needed, General iT can implement “Fuzzy search: smart search with synonym handling and user error correction”, configure the module and verify it against the current website setup.