Topic-Model Recommendation Training from Word-Importance Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI systems for item recommendation require large amounts of learning data and time, leading to increased complexity and inefficiency.
Innovation Solution
An electronic device and method that utilize text data analysis to derive importance and preference data through a topic model, training a recommendation model to provide item recommendations based on user preferences and item relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a deep neural network-based AI system is used for item recommendation, then recommendation accuracy is improved, but learning data requirements and time consumption increase exponentially
Solution Approach 1:
The patent extracts key information from review text data using NLP techniques to create feature vectors that capture user preferences and item characteristics. This extraction process transforms unstructured text into structured features that can be efficiently processed, reducing the need for large volumes of structured learning data while maintaining recommendation accuracy.
Solution Approach 2:
The patent introduces an intermediary processing layer that includes NLP modules and feature extraction mechanisms between the raw data and the recommendation model. This intermediary layer processes and transforms text data into meaningful features, acting as a bridge that reduces the direct dependency on large training datasets by pre-processing and extracting essential information.
2Measurement precision
If a deep neural network-based AI system is used for item recommendation, then recommendation accuracy is improved, but learning time and system complexity increase
Solution Approach 1:
The patent segments the recommendation system into distinct functional modules: NLP processing module, feature extraction module, and recommendation model module. This segmentation allows each component to be optimized independently and enables parallel processing, significantly reducing the overall learning time and computational complexity while maintaining high recommendation accuracy.
Solution Approach 2:
The patent performs preliminary action by pre-processing review text data to extract features and create feature vectors before feeding them into the recommendation model. This pre-processing step includes tokenization, stopword removal, and feature extraction, which prepares the data in advance and reduces the computational burden during the actual learning and recommendation phases.
3Quantity of substance
If text data analysis is used to reduce learning data requirements, then data efficiency is improved, but processing complexity increases
Solution Approach 1:
The patent changes the parameter representation by transforming text data into numerical feature vectors using NLP techniques. This parameter transformation converts unstructured text into a format that can be efficiently processed by machine learning models, reducing the need for large datasets while managing processing complexity through standardized transformation pipelines.
Data Source
AI summary
An electronic device and a controlling method thereof are provided. The electronic device includes a memory and a processor configured to obtain importance data representing an importance of each of a plurality of words included in the text data using text data related to a plurality of items written by a plurality of users, obtain a topic model representing a relationship between a topic and a word by applying a topic modelling algorithm to the obtained importance data, obtain preference data representing preference of each of the plurality of users for the topic and relationship data representing a relationship between the plurality of items and the topic, based on the obtained topic model, and train a recommendation model to output result data including an estimated preference for the plurality of items of the plurality of users, based on the preference data and the relationship data.