Forensic accounting, as a process of legal inspection of corporate accounting practices, has become an important strategy to prevent fraud and financial misconduct, creating information transparency for investors to make financial decisions. Among the capital market companies, oil companies can strengthen the forensic accounting paradigm to provide documentation and transparency of decision-making to prevent market distrust caused by investor risks and contribute to the sustainable development of its presence in advancing competitive strategies. This study aims to develop an effective forensic accounting paradigm based on the risks of investors in oil companies. In this study, the components (dimensions of forensic accounting) and research propositions (investor risk themes) were used to from a combined analysis with 15 accounting experts at the university level. In the quantitative part, the components and propositions identified in the form of matrix questionnaires were evaluated by the interpretive ranking process (IRP) by 20 financial managers of oil companies in the capital market. The results showed that the statements of inflation and credit risk as the most influential themes threaten investors in oil companies, strengthening the focus on legal mechanisms as a component of the forensic accounting paradigm. This result shows that, in the presence of inflationary and credit risks for oil companies’ investors, the importance of legal mechanisms in judicial accounting can lead to increased information transparency to protect the interests of oil companies’ investors
در این تحقیق شاخص کل سهام بورس اوراق بهادار تهران با استفاده از مدل های مختلف شبکه های عصبی پیش بینی شده است. تحقیق از نوع کاربردی است و دوره زمانی انجام تحقیق از ابتدای سال 81 تا پایان سال 90 است. گردآوری اطلاعات از طریق آمار و داده های موجود در پایگاه اطلاعاتی در بورس اوراق بهادار تهران صورت گرفته است. برای ایجاد مدل WDBP از موجک db5 برای نویززدایی داده ها و تا پنج مرحله صورت گرفته است. جذر میانگین مربعات خطا (RMSE) معیارِ ارزیابی برای سنجش خطای پیش بینی است. نتایج این تحقیق نشان می دهد، عملکرد شبکه عصبی موجکی در پیش بینی شاخص سهام سطح خطای کمتری دارد و از شبکه عصبی بهتر است.