Moral Analyzer for Text Data Ethics Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Companies face challenges in quantitatively and objectively evaluating moral expressions related to social responsibility, corporate ethics, and human rights in large amounts of text data, which is essential for timely and appropriate measures, but current techniques lack the capability to analyze these aspects effectively.
Innovation Solution
A moral analyzer is developed, comprising an extraction unit that identifies morality-related words from text data using a dictionary and an analysis unit that evaluates moral values based on these words, employing a moral foundations theory to classify expressions as virtues or vices, enabling objective analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If text data is collected using keyword-based acquisition techniques, then the quantity of text data increases, but the capability to evaluate moral expressions remains insufficient
Solution Approach 1:
The patent segments the analysis process into distinct functional units: an extraction unit that identifies moral expression words from text data using dictionary matching, and an analysis unit that evaluates the moral value of extracted expressions. This segmentation enables specialized processing for moral evaluation while handling large volumes of text data.
Solution Approach 2:
The patent introduces dictionary data containing moral expression words and their categories as an intermediary between the raw text data and the evaluation process. This intermediary enables systematic identification and classification of moral expressions, bridging the gap between unstructured text and quantitative moral evaluation.
2Measurement precision
If manual analysis is performed by individuals with specialized knowledge, then the accuracy of moral evaluation improves, but the efficiency and scalability deteriorate
Solution Approach 1:
The patent enables the system to perform moral evaluation automatically without requiring continuous human intervention. The extraction unit and analysis unit work autonomously to identify and evaluate moral expressions in text data, allowing organizations to process large volumes of data efficiently while maintaining consistent evaluation standards.
Solution Approach 2:
The patent transforms the abstract concept of moral evaluation into quantifiable parameters by categorizing moral expressions into distinct types (e.g., care, fairness, authority, purity, ingroup) and measuring their frequencies and intensities in text data. This parameterization enables objective comparison and analysis of moral values across different datasets.
Data Source
AI summary
A moral analysis unit extracts, as a moral expression word, a word matching a morality-related word related to morality from acquired data which is text data on the basis of a moral foundations dictionary which is dictionary data defining the morality-related word. A processed data display unit analyzes the moral value of the acquired data by using the moral expression word.