Parallel architecture of CNN-bidirectional LSTMs for implied volatility forecast

Mathematics

• Parallel Architectures 100%
• Implied Volatility 99%
• Forecast 79%
• Memory Term 75%
• Neural Networks 63%
• Convolution 55%
• Ensemble 38%
• Long Memory 34%
• Batch 31%
• Standard deviation 29%
• Normalization 28%
• Neural Network Ensemble 21%
• Performance 17%
• Moving Block Bootstrap 15%
• Serial Dependence 13%
• Artificial Neural Network 12%
• Autoregressive Model 10%
• Standards 5%

Business & Economics

• Convolution 94%
• Volatility Forecasts 85%
• Neural Networks 80%
• Implied Volatility 74%
• Forecast Performance 36%
• Batch Size 36%
• Long Memory 32%
• Normalization 31%
• Standard Deviation 27%
• Forecast Comparisons 15%
• Block Bootstrap 14%
• Serial Dependence 13%
• Global Standards 12%
• Autoregressive Model 10%
• Out-of-sample Forecasting 10%
• Artificial Neural Network 10%

Engineering & Materials Science

• Parallel architectures 77%
• Convolution 66%
• Long short-term memory 58%
• Neural networks 40%