Demand Forecasting Model Selection for Web Content
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Websites face challenges in predicting and managing fluctuations in data demand, leading to inadequate provisioning of computing resources, resulting in either underutilization or overutilization during events like sporting tournaments.
Innovation Solution
A method and system that generate multiple models to predict demand spikes and cyclic demand by combining historical and online information, selecting the model with the lowest error, and allocating computing resources based on a predicted demand curve.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If computing resources are provisioned based on average data demand, then resource utilization is stable, but the system cannot handle demand spikes during events
Solution Approach 1:
The system performs preliminary actions by generating multiple demand prediction models in advance and selecting the best model before demand fluctuations occur. The system also provisions computing resources based on predicted demand curves before actual demand spikes happen, allowing the system to prepare adequate resources ahead of time rather than reacting to demand changes as they occur.
Solution Approach 2:
The system dynamically adapts to changing demand conditions by generating multiple different prediction models that can handle various demand patterns (spikes, cyclic variations). The system selects the most appropriate model based on current conditions and continuously adjusts resource provisioning based on predicted demand curves, making the resource allocation flexible and adaptive rather than static.
2Reliability
If computing resources are increased to handle peak demand, then demand spikes can be met, but resources are underutilized during low demand periods
Solution Approach 1:
The system changes the parameter of resource provisioning from fixed to variable by using prediction models to estimate future demand. Instead of provisioning resources at a constant level to handle peak demand, the system adjusts resource allocation based on predicted demand curves generated from historical and real-time data, allowing resource levels to match actual demand patterns and avoiding waste during low-demand periods.
Solution Approach 2:
The system uses feedback mechanisms by continuously monitoring actual demand and comparing it with predicted demand from multiple models. This feedback loop allows the system to refine its predictions and adjust resource provisioning dynamically, ensuring resources are available when needed while minimizing waste during lower demand periods through continuous optimization based on actual performance data.
3Measurement precision
If multiple prediction models are generated and evaluated, then prediction accuracy improves, but processing time and computational overhead increase
Solution Approach 1:
The system applies partial action by generating multiple prediction models but not necessarily evaluating all of them equally or completely. Instead, the system can use a subset of models or employ efficient evaluation methods that identify the best model without exhaustive analysis of every possible model, balancing the need for accuracy with the constraint of available time and computational resources.
Data Source
AI summary
Approaches for predicting demand spikes and cyclic demand for web-related content is provided. An approach includes generating multiple models for predicting demand. The approach further includes selecting a model from the multiple models. The approach further includes inputting information into the model. The approach further includes generating, by at least one computing device, a predicted demand spike using the model and the information. The approach further includes generating a total predicted demand curve by combining the predicted demand spike with predicted cyclical demand.