Hybrid Traffic Prediction Model Selecting ARIMA and Historical Average
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic prediction methods, such as ARIMA and Historical Average Models, struggle to accurately predict traffic conditions at the edges of rush hours and for long-term predictions, as they fail to incorporate historical patterns and sudden changes caused by events like accidents or road constructions, leading to reduced prediction accuracy.
Innovation Solution
A hybrid approach called Historical-ARIMA (H-ARIMA) is developed, which selectively uses ARIMA for short-term predictions and Historical Average Model for long-term predictions, and incorporates event data to improve accuracy by analyzing correlations between event attributes and traffic congestion, using a decision-tree model to choose between models based on prediction errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If ARIMA model is used for traffic prediction, then short-term prediction capability is improved, but prediction accuracy deteriorates at rush hour boundaries and for long-term predictions
Solution Approach 1:
The system dynamically selects between ARIMA and Historical Average models based on the prediction time horizon and traffic conditions. For short-term predictions, ARIMA is used to capture immediate trends, while for long-term predictions or at rush hour boundaries, the Historical Average model is selected to account for periodic patterns, thereby resolving the contradiction between response speed and accuracy across different time scales
Solution Approach 2:
The system changes the modeling parameters by switching between different prediction models (ARIMA vs. Historical Average) depending on the prediction horizon and traffic conditions. This parameter change allows the system to optimize for either short-term responsiveness or long-term accuracy as needed
2Measurement precision
If Historical Average Model is used for traffic prediction, then long-term prediction accuracy is improved, but ability to capture sudden changes deteriorates
Solution Approach 1:
The system dynamically switches between Historical Average model and ARIMA model based on traffic conditions. When sudden changes or events are detected, ARIMA is selected to capture the dynamic changes, while Historical Average is used during stable periods for accurate long-term predictions, thus resolving the contradiction between stability and adaptability
Solution Approach 2:
The system incorporates feedback mechanisms that monitor traffic conditions and prediction errors in real-time. When sudden changes are detected, the system adjusts its model selection accordingly, using ARIMA to adapt to new conditions while maintaining the benefits of Historical Average during stable periods
3Device complexity
If traditional prediction methods are used, then model simplicity is maintained, but prediction accuracy in presence of events deteriorates
Solution Approach 1:
The system segments the prediction task into different model selection scenarios: using ARIMA for short-term predictions and during events, Historical Average for long-term predictions, and a hybrid approach for normal conditions. This segmentation allows the system to maintain relative simplicity while improving accuracy during specific conditions by selecting the appropriate model for each scenario
Solution Approach 2:
The system creates a composite prediction approach by combining ARIMA and Historical Average models into a unified framework that selects the appropriate model based on conditions. This composite system leverages the strengths of both models to achieve high accuracy during events while maintaining manageable complexity through automated model selection
Data Source
AI summary
Systems and techniques for enhancing accuracy of traffic prediction include a system of one or more computers operable to receive a request relating to traffic prediction, compare a first prediction error for a first (moving average) traffic prediction model with a second prediction error for a second (historical average) traffic prediction model, calculated using a historical data set selected from previously recorded traffic data in accordance with a day and time associated with the request, select use of the first model or the second model based on the comparison of prediction errors, and provide an output for use in traffic prediction, wherein the output comes from applying the first traffic prediction model when the first prediction error is less than the second prediction error, and the output comes from applying the second traffic prediction model when the first prediction error is not less than the second prediction error.