Scholarlink

<p>Predicting financial risks for publicly listed companies in the context of the digital economy is crucial due to the uncertainty as well as dynamism of financial markets. Traditional models were not very capable of handling features and problems that exhibited non-linear and high-dimensional characteristics. To address these challenges, an Efficient Multi-layer Diffusion Sea-horse Kernel Convolutional Spiking Attention Neural Network (EMD-SHK-ConvSpAt-Net) is proposed to enhance risk forecasting capabilities. Using the Time-Series Forecasting of Yahoo Finance Data for stocks such as ‘TSLA’, ‘AAPL’, ‘GOOGL’, ‘MSFT’, and ‘AMZN’, the data is pre-processed with a Z-score Min-Max technique for normalization. Feature extraction is carried out by a Spike-Driven Transformer (SDT) to extract salient trends in financial data, and the subsequent selection of the best features with the aid of the Electric Eel Foraging Optimization Algorithm (EEFOA). To improve the temporal resolution of time series data, the EMD-SHK-ConvSpAt-Net is proposed, utilizing the Sea-horse Optimization Algorithm (SHOA) to enhance predictive performance for financial management trends. Implemented in Python, this model is applied to credit risk prediction, demonstrating robust performance in real-world financial environments. The EMD-SHK-ConvSpAt-Net achieves 99.9% accuracy and 99.8% recall, outperforming existing approaches. This model guarantees not only high accuracy in stock market trend forecasting and a reasonable minimum cost but also competitive performance in credit risk prediction, utilizing a sophisticated neural network that integrates superior feature selection and dimension reduction methods. Hence, it becomes one of the most promising approaches for rational financial management in the complex digital economy.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Risk Forecasting in Financial Management for Public Companies Using Efficient Multi-layer Diffusion Sea-horse Kernel Convolutional Spiking Attention Neural Network in the Digital Economy

  • 期刊论文

  • 发表日期：

    2026年2月17日

查看全文
  • K. L. Meera,
  • Rahul S Chaudhari,
  • Sudhansu Sekhar Nanda,
  • D. Barani

摘要

Predicting financial risks for publicly listed companies in the context of the digital economy is crucial due to the uncertainty as well as dynamism of financial markets. Traditional models were not very capable of handling features and problems that exhibited non-linear and high-dimensional characteristics. To address these challenges, an Efficient Multi-layer Diffusion Sea-horse Kernel Convolutional Spiking Attention Neural Network (EMD-SHK-ConvSpAt-Net) is proposed to enhance risk forecasting capabilities. Using the Time-Series Forecasting of Yahoo Finance Data for stocks such as ‘TSLA’, ‘AAPL’, ‘GOOGL’, ‘MSFT’, and ‘AMZN’, the data is pre-processed with a Z-score Min-Max technique for normalization. Feature extraction is carried out by a Spike-Driven Transformer (SDT) to extract salient trends in financial data, and the subsequent selection of the best features with the aid of the Electric Eel Foraging Optimization Algorithm (EEFOA). To improve the temporal resolution of time series data, the EMD-SHK-ConvSpAt-Net is proposed, utilizing the Sea-horse Optimization Algorithm (SHOA) to enhance predictive performance for financial management trends. Implemented in Python, this model is applied to credit risk prediction, demonstrating robust performance in real-world financial environments. The EMD-SHK-ConvSpAt-Net achieves 99.9% accuracy and 99.8% recall, outperforming existing approaches. This model guarantees not only high accuracy in stock market trend forecasting and a reasonable minimum cost but also competitive performance in credit risk prediction, utilizing a sophisticated neural network that integrates superior feature selection and dimension reduction methods. Hence, it becomes one of the most promising approaches for rational financial management in the complex digital economy.