Unstructured Data Framework for Early Warning of Disruptive Events
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for planning and risk assessment in various domains fail to account for unforeseen events that can disrupt operations, leading to delays and disruptions in supply chains, manufacturing, and potential losses due to the inability to detect these events from existing data logs.
Innovation Solution
A framework that processes unstructured data using machine learning models to identify and rank activities of interest, generating alerts for potentially disruptive events, allowing for timely warnings and adaptation to different domains with minimal user feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional techniques analyze large volumes of structured data for planning and risk assessment, then operational efficiency is improved, but the ability to detect unforeseen disruptive events deteriorates because these events are not found in traditional data logs
Solution Approach 1:
The patent introduces unstructured data sources (news articles, social media posts, weather reports, geopolitical data) as intermediary information channels that capture disruptive events before they impact operations. These external data sources act as mediators between emerging risks and the company's risk assessment system, enabling detection of events not present in internal operational logs
Solution Approach 2:
The system performs preliminary analysis of unstructured data to identify potential disruptive events before they materialize into operational impacts. By continuously monitoring external data sources and applying machine learning models to detect patterns and anomalies, the system alerts companies to emerging risks in advance, allowing proactive rather than reactive risk management
2Measurement precision
If machine learning models analyze the entire corpus of unstructured data to identify disruptive events, then detection accuracy is improved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent segments the analysis process into distinct stages: data collection from multiple unstructured sources, preprocessing and filtering to remove irrelevant information, extraction of key features and entities, application of machine learning models to identified patterns, and generation of risk alerts. This segmentation allows the system to process large volumes of unstructured data efficiently by focusing computational resources on the most relevant analysis steps
Solution Approach 2:
The system applies different processing techniques and machine learning models to different types of unstructured data based on their specific characteristics. For example, natural language processing techniques are applied to text data, while other methods are used for structured unstructured data. This localized approach optimizes processing efficiency while maintaining high detection accuracy for each data type
Data Source
AI summary
One embodiment of the present invention sets forth a technique for processing unstructured data. The technique includes applying one or more machine learning models to a set of candidate topics extracted from the unstructured data to determine a set of activities of interest included in the set of candidate topics. The technique also includes generating a set of scores for the activities of interest, wherein each score included in the set of scores represents an estimated impact of a corresponding activity of interest on operations within a domain. The technique further includes determining one or more activities included in the set of activities of interest based on a ranking of the activities of interest by the scores, and causing one or more alerts to be outputted in a user interface, wherein each of the alerts is associated with a potential event related to the one or more activities.