Word Decompounder for Compound Keyword Splitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online advertising systems face challenges in effectively identifying and ranking advertisements when search queries include compound words, as existing methods often fail to decompose these words into relevant substrings, leading to suboptimal ad selection and revenue generation.
Innovation Solution
A system and method that splits compound words into substrings, applies morphemes, and scores their relevance based on query log data and properties to select the most appropriate keywords for ad selection, utilizing a splitting engine and scoring engine to determine the highest scoring substring for ad targeting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the entire compound word is used as a keyword, then the ad selection process is simple, but the ad relevance and targeting accuracy deteriorate
Solution Approach 1:
The patent applies segmentation by dividing compound words into multiple substrings or candidate keywords. The system takes a compound word query (e.g., 'Blumenstraeusse') and segments it into potential keyword components, then evaluates each segment to determine the most relevant keyword for ad selection. This resolves the contradiction by maintaining simple processing through automated segmentation while improving ad relevance through precise keyword identification.
2Measurement precision
If compound words are decomposed into substrings, then ad relevance improves, but the processing complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-processing compound words into candidate substrings before the actual ad selection process. The system performs word decomposition and candidate keyword generation in advance, storing these pre-processed results for efficient retrieval during ad selection. This reduces the complexity during real-time processing while maintaining high keyword accuracy through thorough preliminary analysis.
Solution Approach 2:
The patent introduces an intermediary component (the substring generation and scoring module) that bridges the gap between compound word queries and ad keyword selection. This intermediary processes the compound word, generates candidate substrings, scores them based on relevance criteria, and selects the best keyword. This mediator handles the complexity of decomposition internally, presenting a simplified interface to the rest of the ad selection system while achieving high keyword accuracy.
3Measurement precision
If multiple substrings are generated and scored, then keyword selection accuracy improves, but the computational time increases
Solution Approach 1:
The patent applies partial action by generating and scoring only the most promising candidate substrings rather than evaluating all possible combinations. The system uses heuristics and scoring mechanisms to identify and focus on high-probability keyword candidates, processing a limited subset of substrings that are most likely to be relevant. This reduces computational time significantly while maintaining high keyword selection accuracy by concentrating resources on the most promising candidates.
Solution Approach 2:
The patent changes parameters by dynamically adjusting the number of candidate substrings generated, the scoring thresholds, and the selection criteria based on query characteristics and performance metrics. The system can modify these parameters to balance between processing time and accuracy requirements for different scenarios. This allows flexible optimization of the trade-off between computational time and keyword selection accuracy without sacrificing overall system effectiveness.
Data Source
AI summary
A word is split into one or more first substrings. A morpheme is applied to the one or more first substrings to create one or more second substrings. The one or more first and second substrings are selected as one or more splittings, and a score is generated for each of the one or more splittings. One of the one or more splittings with a score higher than the other one or more splittings is selected as a keyword.