Search Tree Storage Unit for Chinese Character Recommendation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search engine recommendation systems face challenges in efficiently processing polyphones during phoneticization, leading to increased search noise and decreased user experience due to the need for traversing large datasets and incorrect keyword suggestions.
Innovation Solution
A recommendation system and method that utilizes a search tree storage unit with each data node recording address information of recommended words, and a recommended word database that uses Pinyin to correspond to Chinese phrases, improving search speed and accuracy by avoiding polyphonic noise through a Double Array Trie structure and polyphonic word database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a hashmap is used for index searching in the suggest service, then the searching process can be implemented, but the performance deteriorates due to frequent cell calls during user input
Solution Approach 1:
The patent replaces the traditional hashmap data structure with a Trie tree structure. This substitution eliminates the need for frequent cell calls and hash calculations during user input, significantly improving searching performance while maintaining the suggest service functionality. The Trie tree allows for more efficient prefix-based searching without the performance penalties of hashmap collisions and cell management.
2Adaptability or versatility
If polyphones are processed by enumerating all pronunciations, then comprehensive coverage is achieved, but search noise increases and correct results become confused
Solution Approach 1:
The patent extracts and removes polyphonic characters from the search index. Instead of including all possible pronunciations of polyphonic characters, the system identifies and excludes these ambiguous characters from the indexing process. This extraction eliminates the source of search noise while maintaining the ability to handle polyphone-related queries through alternative means, such as contextual analysis or user feedback.
Solution Approach 2:
The patent converts the harmful effect of polyphone ambiguity into a beneficial filtering mechanism. By detecting polyphonic characters and excluding them from the traditional indexing approach, the system actually benefits from this limitation by avoiding the generation of spurious search results. The polyphone issue becomes a trigger for more sophisticated query processing that focuses on disambiguation rather than exhaustive enumeration.
3Quantity of substance
If a larger dataset is used in the dictionary, then more comprehensive suggestions are available, but the searching duration increases
Solution Approach 1:
The patent segments the large dictionary dataset into a structured Trie tree format. This segmentation allows the system to maintain comprehensive coverage of the entire dictionary while enabling efficient searching through the tree structure. Users can navigate the segmented data more quickly by following prefix paths through the Trie tree, reducing searching duration even as the overall dataset size increases.
Data Source
AI summary
The present invention discloses a recommendation system and method for search input. It relates to the field of search engine. The system comprises: a keyword acquisition unit configured to obtain a keyword according to a user input; a search tree storage unit configured to store Chinese characters in a tree data structure, wherein each data node in the tree stores one Chinese character and the address information of the recommended word(s) containing the Chinese character; a recommended word database configured to store the recommended words; an address acquisition unit configured to query the search tree storage unit according to the search keyword to acquire the address information of the recommended word(s); and a suggesting unit configured to query the recommended word database according to the address information to acquire the recommended word(s) and then suggest the recommended word(s) to the user.