Governance Risk Scoring Using NLP Knowledge Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The ESG marketplace focuses on environmental and social data analytics, neglecting governance-focused analytics, and existing solutions lack automated methods for governance risk assessment using Natural Language Processing and scoring processes.
Innovation Solution
A proprietary Scoring Algorithm utilizing Natural Language Processing, graph theory, and neural networks to analyze unstructured data from regulatory filings, converting qualitative text into quantitative governance risk scores, and providing risk quotients/scores for companies, with features like predictive models and benchmarking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If automated methods utilizing Natural Language Processing and scoring processes are implemented, then governance risk assessment capability is improved, but existing solutions lack such automation resulting in manual analysis limitations
Solution Approach 1:
The patent replaces manual mechanical analysis of governance documents with automated Natural Language Processing systems. The NLP pipeline automatically extracts, analyzes, and scores governance risks from regulatory filings, replacing the need for manual review while maintaining or improving reliability through consistent algorithmic application across all analyzed documents.
2Loss of information
If quantitative analytics techniques are applied to governance data, then decision-making insights are improved, but the ESG marketplace neglects governance analytics focusing only on environmental and social data
Solution Approach 1:
The patent creates a universal analytics platform that processes all three ESG pillars (Environmental, Social, and Governance) through the same NLP-based quantitative framework. This multi-functional system applies consistent text analysis methods across diverse ESG data types, enabling comprehensive governance analytics while maintaining adaptability to environmental and social data as well.
Solution Approach 2:
The patent transforms qualitative governance text data into quantitative risk scores through parameter changes in the analysis framework. By converting unstructured textual information about governance practices into structured numerical metrics, the system enables governance analytics to be treated with the same quantitative rigor as environmental and social metrics, improving information utility while maintaining market versatility.
3Measurement precision
If neural network models and iterative graph construction are used to analyze unstructured input data, then risk quotient precision is improved, but computational complexity increases
Solution Approach 1:
The patent segments the complex analysis process into distinct modular components: data collection, text preprocessing, entity extraction, relationship mapping, graph construction, and risk scoring. Each module performs a specific function with well-defined inputs and outputs, making the overall complex system manageable and maintainable while achieving high precision through the cumulative effect of each segmentation step.
Solution Approach 2:
The patent performs preliminary actions by pre-processing input documents through tokenization, lemmatization, and entity recognition before main analysis. It also pre-constructs knowledge graphs with predefined relationships and schemas before applying the neural network models. These preliminary steps organize raw data into structured formats that reduce computational complexity during the main risk assessment phase while improving measurement precision.
Data Source
AI summary
Embodiments are directed to computer-implemented methods of governance risk analysis and scoring of documents that includes: extracting raw text from documents; transforming individual words from the raw text into nodes and edges that create networks of nodes and edges; analyzing certain words pertaining to governance, and graphing the words to a network of nodes and edges; assessing the certain words in the network using mathematics, graph theory, or NLP; determining a semantic score for each node using the number and weight of edges surrounding each node; determining a saliency score based on a word's reoccurrence in the raw text; determining a risk sentiment score determined by the seed dictionary with the list of words exhibiting risk; determining a node score for a certain node based on the semantic score, the saliency score, and the risk sentiment score; and iterating the scores automatically every time a new filing is uploaded.