Financial Crisis Prediction using Feature Subset Selection with Quantum Deep Neural Network

In the process, financial decisions are mostly dependent upon the classification system that is employed for allocating a group of observations into stable groups. A different group of data classification systems can be projected to forecast the financial crisis of institutions utilizing the previous data. An important procedure concern the design of precise financial crisis prediction (FCP) method containing the best of proper variables (features) which are connected to the problems at hand. It is called a feature selection (FS) problem which helps for improvising the classification outcomes. Also, computational intelligence systems are utilized as classifier methods for determining the financial crisis of organizations. Therefore, this study develops an automated FCP using FS with quantum deep neural network (FCPFS-QDNN) technique. The FCPFS-QDNN technique intends to predict the financial crisis via the choice of FS and ML models. Initially, the FCPFS -QDNN technique normalizes the input financial data into a scalar format. For FS process, the FCPFS-QDNN technique uses interactive search algorithm based FS (ISA-FS) technique to choose feature subsets. Finally, QDNN model is applied to the predictive process in the financial sector. The experimental output exhibit that the inclusion of FS and ML has promising influence on enhancing the predictive results of the FCPFS-QDNN technique in terms of different measures.