Media Plan Traffic Prediction via Exponential Dot Product Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting traffic data for advertising campaigns struggle to accurately forecast user interactions over arbitrary time spans, especially when dealing with time periods of different lengths, which hampers effective media planning and budgeting.
Innovation Solution
The method involves using mathematical formulas to predict traffic data by analyzing first and second traffic data from substantially different time spans, employing exponential and dot product functions based on natural logs to estimate unique entities interacting with a location, and applying curve fitting techniques to determine predicted traffic over an arbitrary time span.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traffic data is collected and stored for multiple time periods to enable accurate prediction, then prediction accuracy improves, but data storage requirements and system complexity increase
Solution Approach 1:
The patent extracts only the essential elements needed for prediction (two traffic data points from different time periods) rather than storing and processing all available historical data. This selective extraction maintains prediction accuracy while reducing data storage requirements and system complexity.
Solution Approach 2:
The patent uses mathematical formulas to create a predictive model that replicates traffic patterns based on historical data, eliminating the need to store extensive historical data while maintaining prediction capability. The formula serves as a compact representation that captures essential traffic behavior.
2Adaptability or versatility
If extensive historical traffic data is stored for arbitrary time spans, then prediction coverage improves, but data storage requirements increase
Solution Approach 1:
The patent changes the approach from storing data for arbitrary time spans to using a mathematical formula that can predict any time span based on two reference points. This parameter transformation allows unlimited prediction coverage without proportional increases in data storage, as the formula adapts to different time periods through mathematical scaling.
Data Source
AI summary
Methods, systems, and apparatus for predicting traffic data, including a method comprising: receiving data indicative of traffic data for a mediaplan during first and second time spans, the data representing numbers of unique entities that have interacted with a location during the first and second time spans; and using a function to predict third traffic data during the third time span. The function, when the arbitrary third time span does not exceed the first and second time spans, uses an exponential that includes a difference of dot products using natural logs of combinations of the time spans and first and second traffic data, divided by a difference of natural logs of the first and second time spans. Otherwise, the function uses a difference of dot products using combinations of the time spans and the first and second traffic data, divided by a difference of the first and second time spans.