Transaction Vectorization via Word Segmentation and Embeddings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in accurately understanding the context of electronic transactions due to varied syntax and formats, leading to inefficiencies in processing and providing meaningful services to customers, such as merchant recommendations and discounts.
Innovation Solution
A method and system that preprocesses transaction information using segmentation and machine learning models to generate word embedding vectors, identifying transaction categories and determining similar merchants for display to users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If existing systems use various proprietary syntax and formats to record transactions, then each system can maintain its own data structure, but it becomes computationally expensive and difficult to accurately understand transaction context across different systems
Solution Approach 1:
The patent segments transaction information into discrete transaction words by splitting transaction descriptions at delimiter characters (commas, colons, semicolons, hyphens). This segmentation transforms various proprietary formats into a standardized structure of transaction words, enabling consistent processing across different transaction types while reducing computational complexity.
Solution Approach 2:
The patent transforms transaction data from various proprietary formats into a unified parameter structure consisting of transaction words, categories, and contextual attributes. By changing the representation parameters from format-specific to standardized, the system achieves both adaptability to different formats and computational efficiency in processing.
2Loss of information
If existing systems process transactions in various proprietary formats, then they can maintain format-specific data structures, but they fail to accurately extrapolate contextual information and provide meaningful services
Solution Approach 1:
The patent introduces an intermediary processing layer that translates various proprietary transaction formats into a standardized intermediate representation consisting of transaction words and categories. This intermediary structure preserves all contextual information while simplifying the data for downstream processing, enabling accurate context understanding without increasing overall system complexity.
Solution Approach 2:
The patent creates a universal transaction representation structure that can handle multiple proprietary formats through a single standardized framework. The transaction word structure serves multiple functions: preserving context, enabling categorization, and facilitating service generation, thereby reducing the need for format-specific processing logic.
3Productivity
If existing systems aggregate and process transactions for each customer, then they can provide personalized services, but the various syntax and formats make this computationally expensive
Solution Approach 1:
The patent performs preliminary processing of transactions by segmenting them into transaction words and assigning categories before aggregation. This preliminary action standardizes the data structure in advance, enabling faster and more accurate aggregation and service generation while maintaining reliability through consistent contextual understanding.
Data Source
AI summary
Certain aspects of the present disclosure provide techniques for vectorization of transactions including: receiving electronic transaction information of one or more transactions of a user; for each transaction of the one or more transactions: segmenting the electronic transaction information of the transaction into one or more transaction words; generating a second transaction description related to the transaction; and identifying a category of the transaction; generating, based on the corresponding identified categories of the one or more transactions, a set of transaction history data of the user; providing the set of transaction history data of the user as an input to a machine learned model trained to output a set of word embedding vectors; determining, based on an output of the machine learned model comprising a set of word embedding vectors, a set of similar merchants; and providing the set of similar merchants for display to the user.