Multi-RNN Media Prediction System for Audience Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional media generation and distribution systems face challenges in accurately, flexibly, and efficiently targeting specific consumers with media content, as they often rely on aggregated panel data and rule-based models that fail to account for individual user trends and are inflexible and resource-intensive.
Innovation Solution
The use of a multi-RNN prediction system that trains a plurality of long short-term memory neural networks based on historical media consumption data for individual users or small groups, allowing for dynamic and granular media consumption predictions across various target audiences by selecting and combining subsets of trained neural networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional systems use aggregated panel data and fixed audience groups to generate predictions, then the system structure is simple, but the prediction accuracy deteriorates for smaller groups and non-average audiences
Solution Approach 1:
The patent segments the audience prediction problem by training separate neural networks for different audience segments or individual users rather than using a single aggregated model. This allows the system to capture nuanced patterns in specific groups while maintaining overall system functionality through modular network structures.
Solution Approach 2:
The patent applies local quality by allowing different neural networks to have specialized parameters and architectures optimized for specific audience characteristics. Each network learns local patterns relevant to its target segment, improving prediction accuracy for that specific group without requiring the entire system to be redesigned.
2Measurement precision
If conventional systems use rule-based models to account for sequential trends, then the system is easier to implement, but the ability to accurately account for sequential trends deteriorates
Solution Approach 1:
The patent replaces rule-based mechanical models with neural network-based learning systems. The neural networks automatically learn sequential trends and patterns from historical data without requiring explicit programming of rules, thereby capturing complex temporal dependencies that rule-based systems miss.
Solution Approach 2:
The patent utilizes parameter changes in neural network weights and biases that adapt over time based on incoming data. This allows the model to dynamically adjust to changing sequential trends in media consumption patterns, providing accurate predictions even as user behaviors evolve.
3Productivity
If conventional systems use a single complex model to generate predictions, then the system architecture is unified, but the computing resources and time required deteriorate
Solution Approach 1:
The patent segments the prediction task by dividing it into multiple smaller neural networks that can process different aspects of the prediction problem independently. This segmentation allows for parallel processing and reduces the computational burden on any single model, improving overall prediction speed and resource efficiency.
Solution Approach 2:
The patent applies partial action by using ensembles of simpler models that each handle specific portions of the prediction task rather than one comprehensive complex model. This approach achieves sufficient prediction accuracy through combined simpler models, reducing overall computational requirements and training time.
Data Source
AI summary
Methods, systems, and non-transitory computer readable storage media are disclosed for utilizing a plurality of recurrent neural networks to generate media consumption predictions and providing media content to a target audience. For example, the disclosed system can train a plurality of long short-term memory neural networks for a plurality of users based on historical media consumption data over a plurality of time periods. In one or more embodiments, the disclosed system identifies a target audience including a subset of users and the corresponding neural networks. The disclosed system can then utilize the neural networks of the subset of users to generate a plurality of predictions for a future time period for the users. In some embodiments, the disclosed system then combines the predictions for the users to generate a media consumption prediction for the target audience for the future time period.