A DSR Approach to predict liquidity risk using CNN and Sentiment Analysis

The Design Science Research method (DSR) is an approach to provide practical solutions based on scientific principles in order to produce substantiated and inferred results and products, and at the same time, the results can be scientifically evaluated in the form of primary artifacts and practical use in four main stages which ultimately results in their practical efficiency and effectiveness in the outside world. By designing and creating an archetype in the prototyping stage, DSR evaluates real scenarios and then examines the solution in practical cases. From this point of view, in this research, it has been tried to use the DSR method to provide an innovative solution for predicting bank liquidity risk and upcoming scenarios. This study uses semtiment analysis and deep learning algorithm such as deep convolutional network in predicting liquidity risk and presents a simple and effective method to identify dynamic qualitative variables from recent news about a domestic bank in the country. Predicted scenarios are available to banking experts in the real world to facilitate decision-making in risk measures. According to the guidelines of the Basel Committee and other European banking regulatory frameworks, comparing these scenarios with the scenarios occurring in the bank indicates a relatively high accuracy of the proposed method. In the scenarios derived from the Basel Committee and derived from the European Banking Authority, the forecasting accuracy is about 91% and 82%, respectively