Scholarlink

<p>Understanding customer preferences is crucial for improving marketing strategies and increasing sales. Detecting these preferences through electroencephalography (EEG) signals requires robust preprocessing, feature extraction, and machine learning. Many existing methods overlook noise handling and classification accuracy. We propose a new approach that combines tailored preprocessing, advanced feature extraction, and a deep learning model for neuromarketing. This approach requires high-performance computing (HPC) capabilities to process large-scale EEG data in real-time. Band-pass and Savitzky–Golay filters (4–45&#xa0;Hz) enhance signal quality, while discrete wavelet transform (DWT) and power spectral density (PSD) extract reliable features. Our convolutional neural network (CNN) long short-term memory (LSTM) model, integrated with an attention mechanism, captures spatial–temporal EEG patterns effectively. This model achieves 98.98% accuracy, 99.34% precision, 99.64% recall, and 98.68% F1-score, outperforming traditional CNN–LSTM and deep neural networks (DNN). The proposed method identifies subconscious consumer behavior, offering valuable insights into decision-making, and sets a benchmark for scalable, EEG-based marketing strategies. HPC and parallel processing are essential for real-time preference classification, enabling scalable neuromarketing strategies.</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

Robust consumer preference classification through EEG analysis with CNN-LSTM-attention

  • 期刊论文

  • 发表日期：

    2026年6月15日

查看全文
  • Hojjat Azadravesh,
  • Ali Mousavi

摘要

Understanding customer preferences is crucial for improving marketing strategies and increasing sales. Detecting these preferences through electroencephalography (EEG) signals requires robust preprocessing, feature extraction, and machine learning. Many existing methods overlook noise handling and classification accuracy. We propose a new approach that combines tailored preprocessing, advanced feature extraction, and a deep learning model for neuromarketing. This approach requires high-performance computing (HPC) capabilities to process large-scale EEG data in real-time. Band-pass and Savitzky–Golay filters (4–45 Hz) enhance signal quality, while discrete wavelet transform (DWT) and power spectral density (PSD) extract reliable features. Our convolutional neural network (CNN) long short-term memory (LSTM) model, integrated with an attention mechanism, captures spatial–temporal EEG patterns effectively. This model achieves 98.98% accuracy, 99.34% precision, 99.64% recall, and 98.68% F1-score, outperforming traditional CNN–LSTM and deep neural networks (DNN). The proposed method identifies subconscious consumer behavior, offering valuable insights into decision-making, and sets a benchmark for scalable, EEG-based marketing strategies. HPC and parallel processing are essential for real-time preference classification, enabling scalable neuromarketing strategies.