Emoji Event Sequences for Machine Learning Data Preparation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unstructured data in event content, such as patient charts or claims data, lacks a pre-defined data model, making it difficult to apply analytical methods like statistical and machine learning techniques without losing information.
Innovation Solution
Convert unstructured data into sequences of electronic characters, like emojis, to create a structured format that can be analyzed using machine learning techniques, enabling pattern recognition and visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If unstructured data is stored in textual form without a pre-defined data model, then information completeness is preserved, but applicability of analytical methods deteriorates
Solution Approach 1:
The patent transforms unstructured textual data into structured sequences by changing the parameter of data organization from free-form text to ordered sequences of events with specific attributes (event type, timestamp, entity). This transformation enables the data to be processed by analytical methods while preserving the original information content through systematic encoding.
Solution Approach 2:
The patent introduces an intermediary processing layer that converts unstructured textual event content into a structured intermediate representation (sequence of events with defined attributes). This intermediary form serves as a bridge between the original unstructured data and the requirements of analytical methods, allowing both information preservation and method applicability.
2Adaptability or versatility
If unstructured data is converted to structured format for analysis, then applicability of machine learning techniques improves, but complexity of data preparation increases
Solution Approach 1:
The patent segments unstructured textual data into discrete event units with specific attributes (event type, timestamp, entity). By dividing the continuous text into standardized segments, the system reduces the complexity of data preparation while enabling machine learning techniques to process each segment systematically.
Solution Approach 2:
The patent changes the parameters of data representation from unstructured text to structured events with defined attributes. This parameter transformation simplifies the data preparation process by establishing consistent formats for time ordering, event classification, and entity extraction, making the data ready for machine learning analysis.
3Loss of information
If event content is stored as unstructured text, then information detail is maintained, but measurement and analysis difficulty increases
Solution Approach 1:
The patent creates an intermediary structured representation that maintains the detailed information from unstructured text while organizing it into measurable units. Each event is captured with specific attributes (type, time, entity) that preserve informational detail while enabling systematic detection and measurement through standardized fields.
Solution Approach 2:
The patent transforms the parameters of event content from free-text descriptions to structured parameters including event type categories, timestamp values, and entity identifiers. This parameter transformation maintains the detailed information while making it accessible to analytical methods through defined measurement fields.
Data Source
AI summary
Systems and methods are described for preparing unstructured data for machine learning analysis. An example method may include: receiving data representing a plurality of processes; analyzing the data to identify, for each process of the plurality of processes, a time-ordered sequence of events that occurred during the process; generating a plurality of emoji sequences by, for each process of the plurality of processes, generating an emoji sequence, each emoji in the emoji sequence representing an event of the events that occurred during the process, and the emoji sequence ordered in accordance with the time-ordered sequence; generating a plurality of feature vectors corresponding to the respective plurality of emoji sequences; and applying a machine learning technique to the plurality of feature vectors.