Database Querying System Using Semantic Segmentation

The patent segments the search query into multiple contiguous sequences of words (n-grams) and processes each sequence independently through the semantic network and classification system. This segmentation allows the system to evaluate partial matches and combine results, achieving comprehensive coverage without querying all possible permutations of the entire query, thus reducing resource consumption while maintaining completeness.

The patent performs preliminary classification of contiguous word sequences against a semantic network before executing full database queries. By pre-filtering sequences that match known semantic categories and setting classification flags, the system identifies promising search paths in advance, avoiding unnecessary full database scans and reducing overall resource consumption while ensuring relevant results are captured.

The patent divides the query into overlapping contiguous sequences of words and processes each segment through the semantic network independently. This segmentation enables the system to capture local semantic meanings (such as compound terms like 'red bull' in different contexts) without requiring complex global analysis of all possible word aggregations, thus maintaining accuracy while reducing processing complexity.

The patent introduces a semantic network as an intermediary layer between the raw query and the database. This semantic network acts as a mediator that pre-classifies word sequences and establishes semantic relationships, allowing the system to interpret word aggregations accurately without directly computing all possible permutations. The intermediary simplifies the processing by providing pre-computed semantic knowledge.