Enhancement of fraud detection for narratives in annual reports

Therefore, this work integrates natural language processing (NLP), queen genetic algorithm (QGA) and support vector machine (SVM) to develop a fraud detection method for narratives in annual reports, such as reports to shareholders, and thereby enhance the fraud detection accuracy and reduce investors' investment risks. To achieve the above-mentioned objective, a process of fraud detection for narratives in annual reports is first designed. Techniques related to fraud detection for the narratives in annual reports are then developed. Finally, the proposed fraud detection method is demonstrated and evaluated.