| Since the 20 th National Congress of the Communist Party of China,China has continuously strengthened its emphasis on environmental protection,proposed the "Five in One" strategic development layout,and clearly listed accelerating the construction of ecological civilization as a major strategy for national economic and social development.Strengthening enterprise environmental governance is currently an important task.In the context of big data in the new era,stakeholders’ requirements for environmental information disclosure are also keeping up with the times.They not only pay attention to the monetary disclosure content in environmental reports,but also pay more attention to the internal quality of environmental information disclosure.The disclosure of corporate environmental information is the main way for stakeholders to understand the performance of corporate environmental responsibilities,and it is also a medium for enterprises to establish a good image.The construction of an environmental information quality evaluation system with textual features compensates for the subjective bias of traditional manual scoring,and the objective evaluation of the quality of environmental information disclosure is also conducive to rational and effective decisionmaking by national environmental governance supervision departments and stakeholders.This thesis selects Sichuan Kelun Pharmaceutical Co.,Ltd.as a case company,and applies the construction of an environmental information disclosure quality evaluation system with text characteristics to Kelun Pharmaceutical.It sorts and analyzes the comprehensive evaluation results of the environmental information disclosure quality of the company from 2011 to 2020,and then proposes improvement strategies based on the analysis of the evaluation results.The specific construction process of the evaluation system is based on text analysis.It analyzes the text characteristics of enterprise environmental information disclosure from four aspects: text semantic network analysis,text emotional orientation,text similarity,and LDA theme model,and constructs five evaluation indicators: understandability,reliability,text emotional orientation,text similarity,and integrity through sorting out relevant literature,After standardizing the results of each evaluation index,the entropy weight method is used to assign the weight of the index to obtain the results of comprehensive evaluation of the quality of environmental information disclosure.Finally,the construction of the environmental information disclosure quality evaluation system based on text analysis is completed.This thesis is mainly divided into the following six parts: The first part is introduction.Put forward the research background and significance,sort out the domestic and foreign relevant literature on the quality of environmental information disclosure and text analysis,and elaborate the research background and significance of this thesis.The second part is the quality evaluation of environmental information disclosure and the overview of text mining theory.This thesis expounds the standards and significance of environmental information disclosure at home and abroad,summarizes and sorts out the methods and relevant theoretical basis for the quality evaluation of environmental information disclosure,introduces the concept,process and methods of text mining,and gives the necessity and feasibility of building the evaluation system.The third part is the construction of environmental information evaluation system based on text analysis.This part mainly elaborates the specific operation process of text feature extraction and evaluation index construction,as well as the principle and use of entropy weight method.The fourth part is the environmental information disclosure quality evaluation of Kelun Pharmaceutical.Apply the above evaluation system to the 2011 – 2020 environmental report of Kelun Pharmaceutical,draw specific text characteristics and indicator results,and then analyze the application effect of the evaluation system.The fifth part is the evaluation results and improvement strategies of the environmental information disclosure quality of Kelun Pharmaceutical.Analyze the evaluation results of five dimensions,reveal the problems existing in the quality of environmental information disclosure of enterprises,and then give corresponding strategies to improve the quality of environmental information.The sixth part is conclusion and enlightenment.This thesis mainly draws the following conclusions.According to the comparative analysis of the application effect of the evaluation system and the environmental information disclosure index EDI,the application effect of the system is good.Among the five evaluation dimensions,Kelun Pharmaceutical did not perform satisfactorily in terms of text similarity and text emotional orientation,which revealed that the company’s internal supervision responsibilities were not fulfilled in place,and the quality of environmental information disclosure needed to be improved.According to the research in this thesis,we also draw the following inspiration.In order to improve the quality of environmental information disclosure,enterprises can build an evaluation system with text characteristics,build a joint internal supervision system,and refine the rules for environmental report text disclosure.The quality evaluation system of environmental information disclosure in this thesis is to construct five text-based evaluation dimensions and use Python language and text analysis method to quantify each indicator.Compared with the traditional manual scoring method,it enhances the objectivity of the evaluation.The evaluation system is easy to operate and easy to promote,and text mining using computer program language can extract available and effective information from a large number of text data,which has reference significance for enterprise unstructured information evaluation,enriches the theoretical system of enterprise environmental information disclosure research,and has long-term practical significance for the development of environmental information. |