AI Viewership Projection Using Partial Data for Broadcast Timing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing viewership rating systems struggle to predict viewership for television programming where prior ratings are unavailable, making it difficult for broadcasters and advertisers to determine optimal broadcast times and advertising rates.
Innovation Solution
A machine learning model is trained to identify relationships between content parameters and viewership ratings, allowing it to project optimal broadcast times for unaired video content to maximize viewership ratings using partial data sources, including social media data for real-time adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sampling techniques are used to measure viewership, then viewership ratings can be obtained for select channels and time slots, but it becomes difficult to predict viewership for programming without prior ratings data
Solution Approach 1:
The system performs preliminary actions by collecting and storing various data sources (social media metrics, content attributes, demographic information) before the broadcast occurs. This advance data gathering enables the machine learning model to make predictions for unaired content without requiring prior viewership ratings, resolving the contradiction between measurement precision and adaptability to new content
Solution Approach 2:
The patent introduces an intermediary machine learning model that bridges the gap between available data sources and viewership prediction. This intermediary system processes alternative data (social media engagement, content metadata) to generate viewership estimates for content without traditional ratings, enabling both precision and versatility simultaneously
2Measurement precision
If more data sources are collected to improve prediction accuracy, then the model can make better projections, but the complexity of the system increases
Solution Approach 1:
The machine learning model is designed with multi-functionality to process diverse data types (social media metrics, content attributes, demographic data) through a single unified system. This universal approach improves prediction accuracy across multiple content types and platforms while avoiding the need for separate complex systems for each data source
Solution Approach 2:
The system dynamically adjusts parameters and weights in the machine learning model based on the specific content being analyzed and the available data sources. This parameter adaptation allows the system to optimize prediction accuracy for different scenarios without requiring fundamentally different system architectures, thereby managing complexity while maintaining precision
Data Source
AI summary
A computer-implemented method of dynamically determining when to broadcast video content. The method includes providing one or more content parameters corresponding to broadcasted video content to a machine learning (ML) model, iteratively training the ML model to identify relationships between the one or more content parameters and viewership ratings associated with the broadcasted video content, receiving one or more content parameters corresponding to unaired video content, wherein the one or more content parameters include a desired platform for broadcasting the unaired video content, receiving viewership data associated with the unaired video content and/or the desired platform, providing the one or more content parameters and the viewership data to the trained ML model, and determining, via the trained ML model, a target time period for broadcasting the unaired video content on the desired platform that maximizes a projected viewership rating of the unaired video content.