Channel Recommendation Device Using Named Entity Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital TV channel recommendation systems fail to consider user usability and often experience intent analysis failures due to insufficient channel name entities in dictionaries, leading to unsuitable channel recommendations.
Innovation Solution
A channel recommendation device that uses named entity recognition (NER) on user log information and a dictionary to extract and recommend channels based on viewing frequency, updating the dictionary with channels that fail NER to improve intent analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If random selection logic is used to recommend channels, then the system operation is simple, but the usability for users deteriorates and suitable channels are not recommended
Solution Approach 1:
The system collects user log information including channel viewing frequency and uses this feedback to dynamically generate personalized channel recommendations, transforming the random selection into an adaptive recommendation system that improves suitability while maintaining operational simplicity
Solution Approach 2:
The system automatically analyzes user viewing habits and generates recommendations without requiring manual user input or complex configuration, enabling the system to serve itself in optimizing recommendations based on accumulated user data
2Device complexity
If the entity name dictionary is insufficient for channel names, then the dictionary structure is simple, but intent analysis accuracy deteriorates and channel recommendation fails
Solution Approach 1:
The system performs named entity recognition on channel names in advance and proactively adds unrecognized channel names to the dictionary, preparing the dictionary for future intent analysis tasks and preventing analysis failures before they occur
Solution Approach 2:
The system automatically expands the entity name dictionary by identifying and adding channel names that fail NER, enabling the dictionary to self-improve and adapt to new channel names without manual intervention
Data Source
AI summary
A channel recommendation device includes a memory, a communication interface configured to communicate with a display device, and a processor configured to extract N channels from electronic program guide (EPG) information, perform named entity recognition (NER) on the extracted N channels, obtain one or more of channels successful for the NER as a recommended channel, and transmit information about the recommended channel through the communication interface to the display device.


