Text Level Classification Model for CSR Report Material Recommendation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Enterprises face challenges in compiling high-quality Corporate Social Responsibility (CSR) reports due to difficulties in understanding how to collect and present data effectively, leading to inconsistencies and errors in reporting sustainable development information.
Innovation Solution
A method and system for recommending report material that involves obtaining evaluated reports, extracting reference text materials, training a text-level classification model based on these materials and their actual rating levels, and using this model to determine the predicted level information and recommended order of text materials for generating a report.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If CSR reports are written and revised based on experiences of expert consultants, then the content meets ESG requirements and is recognized by various enterprises, but unexpected situations and errors may occur and high-quality CSR reports cannot be achieved
Solution Approach 1:
The system copies and analyzes high-rated CSR reports to extract effective content patterns and structures. By training the classification model on evaluated reports with known rating levels, the system learns from successful examples and replicates their qualities in new report recommendations, reducing reliance on subjective expert experience.
Solution Approach 2:
The system implements feedback by using actual rating levels of evaluated reports as training data to improve the classification model. The model continuously learns from the correlation between report content and rating outcomes, enabling it to provide increasingly accurate recommendations that reflect what actually leads to high ratings.
2Productivity
If multiple text materials are collected for CSR report compilation, then comprehensive coverage is achieved, but it becomes difficult to select and organize the most appropriate materials efficiently
Solution Approach 1:
The classification model performs self-service by automatically evaluating and ranking text materials based on their relevance to rating criteria. Instead of requiring manual assessment of each material, the system autonomously determines the recommended order, significantly reducing the time and expertise needed for material selection while maintaining comprehensive coverage.
Solution Approach 2:
The system transforms the complex qualitative task of material selection into a quantitative parameter-based evaluation. By assigning predicted level information and recommended orders to each text material based on multiple parameters learned from training data, the system enables efficient automated sorting and selection of the most appropriate materials.
3Measurement precision
If enterprises use manual methods to collect and write CSR report content, then flexibility in customization is maintained, but inconsistencies and errors occur and rating scores are limited
Solution Approach 1:
The classification model acts as an intermediary between the raw text materials and the final report compilation process. It mediates by evaluating materials against learned rating criteria and providing recommended orders, ensuring that the selected content aligns with what leads to high rating scores while maintaining the flexibility of manual report writing.
Data Source
AI summary
The disclosure provides a method and a system for recommending report material. The method includes the following steps. A plurality of evaluated reports and an actual rating level of each of the evaluated reports are obtained. A plurality of reference text materials related to a rating topic are extracted from the evaluated reports. A classification model training is performing based on the reference text materials and the actual rating levels of the evaluated reports to establish a text level classification model. Predicted level information for each of text materials to be evaluated is determined by using the text level classification model, to obtain recommended order for each of the text materials to be evaluated. A report is generated based on the recommended order of each of the text materials to be evaluated.