Traffic Data Replacement Interface Using AI Model Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic data analysis is hindered by abnormal data, such as anomaly and missing data, which are generated due to environmental factors or network issues, leading to analysis errors.
Innovation Solution
A user interface is provided for efficiently replacing abnormal traffic data using artificial intelligence models, allowing users to select and verify replacement results, and incorporating performance information to optimize the replacement process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If abnormal traffic data is used in analysis, then data processing is simple, but analysis accuracy deteriorates due to errors from anomaly and missing data
Solution Approach 1:
The patent extracts and removes abnormal traffic data (anomaly data and missing data) from the overall traffic dataset before analysis. By identifying and separating problematic data points, the system ensures that only normal, reliable data is used for analysis, thereby improving accuracy without requiring complex processing of abnormal data patterns
Solution Approach 2:
The patent performs preliminary data cleaning and validation before the main analysis process. By pre-identifying and flagging abnormal data points using validation rules, the system prepares the dataset in advance, preventing analysis errors before they occur and simplifying the subsequent analysis workflow
2Reliability
If manual verification of replacement data is implemented, then data reliability is improved, but processing time increases
Solution Approach 1:
The patent implements partial manual verification by providing a user interface that displays candidate replacement data for reviewer examination. Instead of requiring verification of all replacement data, the system presents only the generated replacement candidates for selective review, balancing reliability improvement with time efficiency through targeted rather than comprehensive verification
Data Source
AI summary
A method for providing a user interface to replace abnormal traffic data on a computing device is disclosed. The method involves displaying target traffic data for a site where a replacement task will occur. Upon user input specifying a replacement period, the interface shows performance data for AI models mapped to the site and candidate inference results generated by these models. The user selects a target inference result or AI model, and the interface displays the replacement result. The interface includes a verification screen with a chart highlighting the replacement period and a list detailing performance data and replacement results for sub-periods. Each sub-period replacement is linked to corresponding AI model performance, using models pre-stored and mapped to the site. This may ensure accurate and detailed data replacement guided by AI model performance and user input.