Word Embedding Model for Sentiment Analysis Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current social networking systems lack effective methods to analyze public insights and sentiments from user-generated content, failing to provide accurate and timely reflections of contemporary public opinions.
Innovation Solution
A social networking system utilizes a word embedding model trained on a large corpus of text from user-generated content to represent n-grams as vectors in a d-dimensional embedding space, enabling the calculation of similarity metrics and generation of insights that reflect public sentiments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional text analysis methods are used, then system complexity is low, but measurement precision of public sentiments is insufficient
Solution Approach 1:
The patent replaces traditional mechanical text analysis methods with a neural network-based word embedding system. The neural network automatically learns semantic representations of words and n-grams from user-generated content, substituting manual or rule-based analysis with an automated machine learning approach that achieves superior measurement precision in sentiment analysis.
Solution Approach 2:
The patent transforms text data into vector representations in a multi-dimensional embedding space, changing the parameter representation from discrete text tokens to continuous numerical vectors. This parameter transformation enables the application of geometric operations and similarity metrics to capture semantic relationships, significantly improving sentiment analysis accuracy.
2Measurement precision
If word embedding models are trained on user-generated content, then sentiment representation accuracy is improved, but loss of time for model training increases
Solution Approach 1:
The patent performs preliminary action by pre-training the word embedding model on a large corpus of user-generated content before deployment. This preliminary training phase captures contemporary language usage and sentiments, creating a ready-to-use embedding space that can quickly adapt to new queries without requiring retraining, thus reducing operational time loss.
Solution Approach 2:
The patent creates a compressed vector representation (copy) of the semantic meaning of n-grams from user-generated content. Instead of storing and processing the original text data, the system uses compact vector embeddings that capture the essential semantic information, enabling fast similarity computations and sentiment analysis without the time cost of processing raw text.
3Loss of information
If n-grams are represented as vectors in embedding space, then semantic context understanding is improved, but device complexity increases
Solution Approach 1:
The patent maps n-grams from discrete text space into a continuous multi-dimensional embedding space. This dimensionality change allows the representation of semantic context through geometric relationships in the vector space, where similar meanings are positioned closer together. The d-dimensional vectors capture nuanced semantic relationships that cannot be represented in traditional text form.
Data Source
AI summary
In one embodiment, a method includes receiving a request to identify public sentiments for one or more entities from a second computing device, where each of the attribute n-grams represents an attribute among a plurality of pre-determined attributes, accessing a table of word vector relationships, looking up entity word vectors corresponding to each of the names for the one or more entities and attribute word vectors corresponding to each of the plurality of attribute n-grams using the table, calculating, for each of the entity word vectors, a similarity metric to each of the attribute word vectors, and sending a response message to the second computing device, where the response message includes calculated similarity metrics corresponding to all the pairs of an entity word vector and an attribute word vector.