Entity Type Prediction for Search Query Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large online systems, such as CRM systems, face inefficiencies in processing search queries due to the need to evaluate each record individually, leading to high resource consumption and prolonged processing times, which negatively impacts user experience.
Innovation Solution
Implementing a machine learning model that predicts the entity type of a record based on morphological and dictionary features of a search query, allowing the system to limit searches to selected entity types, thereby reducing resource usage and improving processing speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the online system makes an individual determination for each stored record to identify records related to the search query, then the search results will be comprehensive and relevant, but the system resource consumption increases and processing time extends
Solution Approach 1:
The system performs preliminary classification of records into entity types before search queries are executed. By pre-organizing records with entity type labels, the system can quickly filter and retrieve only the relevant entity types when a search query is received, avoiding the need to individually evaluate every record during search operations.
Solution Approach 2:
The invention extracts the entity type information from individual records and creates a separate, aggregated structure that maps search queries to likely entity types. This extraction allows the system to operate at the entity type level rather than the individual record level, significantly reducing processing complexity while maintaining search relevance.
2Measurement precision
If the online system makes an individual determination for each stored record, then comprehensive search coverage is achieved, but system resource consumption increases
Solution Approach 1:
Records are pre-classified into entity types during data ingestion, and this classification is stored as metadata. When a search query arrives, the system uses this pre-established mapping to quickly identify which entity types are relevant, avoiding the need to scan and evaluate every individual record in detail, thus reducing computational resource consumption while maintaining comprehensive search coverage.
Solution Approach 2:
The entity type acts as an intermediary layer between the search query and the actual records. Instead of directly evaluating each record against the search query, the system first determines which entity types match the query, then retrieves only records belonging to those entity types. This intermediary classification layer significantly reduces the computational burden.
3Measurement precision
If the online system processes search queries by evaluating all records, then complete search results are obtained, but the user experience deteriorates due to prolonged processing time
Solution Approach 1:
The system performs preliminary entity type assignment during record creation, establishing a ready-to-use classification index. When users submit search queries, the system leverages this pre-built index to rapidly filter records by relevant entity types, dramatically reducing search response time and providing immediate, useful results that improve user experience while maintaining completeness.
Data Source
AI summary
As part of providing the services to users, an online system stores multiple records that are accessible by users of the online system. When a user provides a search query, the online system extracts morphological and dictionary features from the query. The online system provides the extracted features to a machine learning model as an input. The machine learning model outputs a score for each potential entity type that indicates a likelihood that the search query is for a record associated with the entity type. The output from the machine learning model is used by the online system to select one or more entity types that the user is likely searching for. The online system searches the stored records based on the search query but limits the searching to records associated with at least one of the selected entity types.