Order Search Using Logistics Data for Higher Recall Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional order searching methods in e-commerce have low recall rates and inaccurate recalls, primarily due to reliance on product names and store names without incorporating logistics data.
Innovation Solution
An order searching method that utilizes both a transaction database and a logistics database, associating them through a common order identifier, segments and processes logistics data to extract relevant information, and matches search queries against both databases for accurate retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional order searching methods using only product names and store names are used, then the system simplicity is maintained, but the recall rate and search accuracy deteriorate
Solution Approach 1:
The patent segments the search system into two distinct databases: a transaction database storing order data (product names, store names, order identifiers) and a logistics database storing logistics data (tracking numbers, courier information, delivery status). This segmentation allows each database to be optimized for its specific data type while improving overall search accuracy through comprehensive data coverage.
Solution Approach 2:
The patent adds a new dimension to the search system by introducing logistics data as an additional search dimension beyond traditional product names and store names. Users can now search for orders using logistics tracking numbers, courier information, or delivery status, significantly expanding the search capabilities and accuracy.
2Reliability
If logistics data is integrated into the search system, then the recall rate improves, but the data processing complexity increases
Solution Approach 1:
The patent segments logistics data into structured fields including tracking numbers, courier information, delivery status, and time information. This segmentation enables efficient indexing and retrieval operations, reducing the processing complexity despite the increased data volume from integrating logistics information.
Solution Approach 2:
The patent introduces an order identifier as an intermediary that links transaction data and logistics data across the two databases. This intermediary enables efficient joining and correlation operations without requiring complex many-to-many relationships, thereby improving recall rate while controlling processing complexity.
3Measurement precision
If multiple databases are used for searching, then the search coverage and accuracy improve, but the query processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing and indexing logistics data during data ingestion, including extracting key fields like tracking numbers and courier information. This pre-indexing enables fast retrieval operations when queries are executed, reducing the actual query processing time despite using multiple databases.
Solution Approach 2:
The patent implements a multi-path query processing mechanism that can skip to the most appropriate database based on the search input type. For example, if a tracking number is provided, the system directly queries the logistics database without unnecessarily accessing the transaction database, thereby reducing query processing time.
Data Source
AI summary
Input search information comprising keyword information representing an intention of an order query is obtained; and order data matching the search information is queried based on a transaction database and a logistics database. The transaction database comprises order data stored during order generation, and the logistics database comprises dynamically obtained logistics data of an order. The transaction data and the logistics data have a same order identifier for a same order. The queried order data is presented.