Recommendation System Using Feature Word Embeddings for Cold Start Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional recommendation systems are limited in generating recommendations for user profiles without historical interaction data and fail to account for implicit interactions between users and items, leading to irrelevant suggestions.
Innovation Solution
A digital medium environment uses feature word embeddings to generate vectors representing user profiles, items, and interactions, distinguishing between explicit and implicit interactions, and employing a word embedding model to create a feature similarity matrix that accounts for both types of interactions, enabling personalized recommendations even for 'cold' user profiles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional recommendation systems use only explicit interaction data, then the system is simple to implement, but the recommendation precision deteriorates due to inability to account for implicit interactions
Solution Approach 1:
The patent combines explicit interaction data and implicit interaction data into a unified recommendation framework. The system merges multiple data sources (view history, purchase history, viewing duration, replay counts) into a single recommendation model, allowing both types of interactions to contribute to recommendation precision without requiring separate complex systems.
Solution Approach 2:
The patent introduces an intermediary processing layer that converts implicit interaction data (viewing duration, replay counts) into meaningful features that can be integrated with explicit interaction data. This intermediary layer processes raw implicit data into usable features before feeding them into the recommendation model, bridging the gap between different data types.
2Adaptability or versatility
If conventional recommendation systems require historical explicit interaction data, then the recommendation model can be trained effectively, but the system fails to provide recommendations for cold start user profiles
Solution Approach 1:
The patent performs preliminary actions by collecting and processing implicit interaction data (viewing durations, replay counts) before explicit interaction data becomes available. This preliminary processing of implicit data creates initial user profiles and item representations that enable the system to provide recommendations for cold start users, even before they have explicit interaction history.
Solution Approach 2:
The patent changes the parameters used for user profiling by incorporating implicit interaction parameters (viewing duration, replay counts) alongside or instead of traditional explicit interaction parameters. This parameter change allows the system to create reliable user profiles and generate recommendations even when explicit interaction data is unavailable.
3Measurement precision
If the system processes both explicit and implicit interaction data, then the recommendation precision improves, but the data processing complexity increases
Solution Approach 1:
The patent segments the data processing into distinct modules: one module processes explicit interaction data (view history, purchase history) while another module processes implicit interaction data (viewing duration, replay counts). This segmentation allows each module to handle its specific data type with appropriate processing logic, reducing overall system complexity while maintaining high recommendation precision.
Data Source
AI summary
A digital medium environment is described to facilitate recommendations based on vectors generated using feature word embeddings. A recommendation system receives data that describes at least one attribute for a user profile, at least one item, and an interaction between the user profile and the at least one item. The recommendation system associates each user profile attribute, each item, and each interaction between a user profile and an item as a word, using natural language processing, and combines the words into sentences. The sentences are input to a word embedding model to determine feature vector representations describing relationships between the user profile attributes, items, and explicit and implicit interactions. From the feature vector representations, the recommendation system ascertains a similarity between different features. Thus, the recommendation system can provide customized recommendations based on implicit interactions, even for a user profile that is not associated with any historical interaction data.