A TEXT MINING ANALYSIS FOR THE EFFECTIVENESS OF GREEN INFORMATION TECHNOLOGY IN THE TAIWANESE CORPORATE SOCIAL RESPONSIBILITY REPORT

摘要

With the increasing importance of the environment protection, socially responsible investing growth potential is extraordinary. Many companies incorporate green information technology into the report of corporate social responsibility to consider both financial return and social good. However, it also brings various voluntary disclosure and review problems that lie on the path toward success. One of the problem is the stakeholders generally lack of sound assurance that whether their corporate social report are effective and efficient, or information announced to stakeholders are reliable and integrity. Thus, for the sake of disclosure and assurance, a lot of design and implement integrative platforms have been developed in the field of green information technology assurance. Text mining is a statistical pattern learning method, and has been adopt to verify text categorization and clustering, sentiment analysis, and document summarization, and entity relation modeling. In this research, we try to develop an automatic mechanism to detect the structure and content of the corporate social responsibility reports, and the inconsistency between actual corporate social responsibility with green information technology and announced content of corporate social responsibility reports. The mechanism adopts text mining as verification tool, and we apply this mechanism to audit the content of corporate social responsibility to reach automatic analyzing the content of corporate social responsibility report. A prototype has been developed to evaluate feasibility of mechanism, and we evaluated the verification and validation through a real case. The result shows that it is useful for assurance of the case company's corporate social responsibility report, and the system can provide information effectiveness and efficient for them to detect possible difference.