Optimization of Enterprise Management Accounting based on Temporal Convolution Network with Gated Recurrent Unit

In recent years, Enterprise Management Accounting (EMA) has developed rapidly by transforming from traditional cost accounting to extensive range of activities which includes risk management and performance measurement. However, EMA consists several challenges such as slow manual peed and analysis which provides difficulty in processing huge quantity of data. To resolve the above issues, Temporal Convolutional Network (TCN) with Gated Recurrent Unit (GRU) namely (TCN-GRU) is proposed for optimizing the EMA. Initially, an enterprise financial management model is designed for incorporating multiple modules. Then, the proposed TCN-GRU is employed for predicting the financial risk effectively by capturing the high-order features and filters unwanted data from financial information. After that, the Multi-Objective Dragonfly Algorithm (MODA) is applied for optimizing the EMA model by generating various solutions that represents exchanges among minimizing risk and maximizing returns. From the results, the proposed TCN-GRU attained better results in terms of True Positive Rate (TPR) of 80.90%, True Negative Rate (TNR) of 83.67%, Total Rate (TR) of 82.40% when compared to the existing TCN-Long Short-Term Memory (LSTM).