Financial Distress Prediction by using Long Short-Term Memory based Adaptive Whale Optimization Algorithm

In recent days, financial distress had been attributed to both internal and external factors, the internal causes included poor financial management while external influences economic downturns, heightened competition, and regulatory shifts. Traditional approaches for predicting financial distress had faced several challenges which include high sensitivity to outliers and inability to process sequential data. Therefore, this research proposes Long Short-Term Memory based Adaptive Whale Optimization Algorithm (LSTM-AWOA) for predicting financial distress, as it is effective at handling sequential data. Initially, data is collected from polish company dataset which consists of 10,503 samples with 64 financial attributes. Then, this dataset is preprocessed for handling missing values, and removing redundant rows which improves data quality, leading to better model performance and also resolves overfitting issues. After that, prediction is done by using LSTM and to improve its performance AWOA optimization is done. The proposed LSTM-AWOA achieved better results in terms of accuracy (0.9765), precision (1.0000), recall (0.9678) and F1-score (0.9769) when compared to existing AWOA-Deep Learning (AWOADL).