Partial Query Ad Serving via Probability Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search systems face challenges in providing relevant advertisements for partial queries, as they often result in ambiguous or incomplete queries, leading to irrelevant advertisements being displayed.
Innovation Solution
The method involves receiving a request for advertisements based on query suggestions for a partial query, ranking these suggestions by their probability of selection and commercial value, and selecting a proper subset to identify relevant advertisements for presentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If advertisements are provided for all query suggestions of a partial query, then the quantity of advertisements increases, but the relevance of advertisements to user intent deteriorates
Solution Approach 1:
The patent segments the set of query suggestions into a proper subset based on their probability of being selected by the user. Instead of treating all query suggestions equally, the system divides them into relevant (high probability) and irrelevant (low probability) groups, and only serves advertisements for the relevant subset. This segmentation resolves the contradiction by maintaining advertisement quantity for important queries while eliminating irrelevant advertisements.
Solution Approach 2:
The patent changes the parameter of query suggestion selection from binary (include/exclude) to probabilistic (based on selection probability). By introducing a probability threshold or ranking system, the system dynamically determines which query suggestions warrant advertisement display. This parameter change allows the system to optimize both advertisement quantity and relevance simultaneously.
2Reliability
If a proper subset of query suggestions is selected based on probability ranking, then the relevance of advertisements improves, but the complexity of the advertisement selection process increases
Solution Approach 1:
The patent performs preliminary action by pre-calculating and storing the probability rankings of query suggestions before the actual advertisement selection process. The system maintains a ranked list of query suggestions based on historical data and user behavior patterns. When a partial query is entered, the system simply retrieves the pre-ranked suggestions rather than performing complex real-time analysis, thus reducing the complexity of the advertisement selection process while maintaining high relevance.
3Measurement precision
If query suggestions are ranked by probability of selection and commercial value, then the precision of advertisement matching improves, but the time required for processing increases
Solution Approach 1:
The patent applies partial action by selecting only the top-ranked query suggestions (a proper subset) for advertisement display rather than processing all possible suggestions. The system ranks query suggestions by probability and commercial value, then selects a limited number of top suggestions to serve advertisements for. This partial processing approach maintains high precision in advertisement matching while significantly reducing the time required compared to evaluating all query suggestions.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for serving advertisements based on partial queries. In one aspect, a method includes receiving a request for advertisements that specifies a set of query suggestions. The query suggestions are ranked based, at least in part on a probability of each query suggestion being selected by a user that input the partial query. A proper subset of the query suggestions is selected and includes a highest ranked query suggestion based on the ranking and one or more query suggestions of the set of query suggestions that is at least one of a revision of the highest ranked query suggestion or refers to a same topic as the highest ranked query suggestion. One or more advertisements are identified for each query suggestion of the proper subset. Data specifying the identified advertisements are provided.