Scholarlink

With advancements in modern pharmacology, systems biology, and deep learning (DL) technology, uncovering the intricate connections between components of traditional Chinese medicine and their targets has become a growing research focus. The pharmacological effects of traditional Chinese medicines are usually manifested as multi-target and multi-pathway effects, but traditional research methods have difficulty in feature selection, limited expression capabilities, and insufficient target classification and prediction capabilities, making it difficult to fully reveal the interactions between traditional Chinese medicine ingredients and targets. To tackle the issues mentioned above, this paper integrates Convolutional Neural Networks (CNN) and Transformer networks for joint modeling to investigate the multi-target mechanisms of traditional Chinese medicine components. The study leverages data from the public TCMSP (Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform) to optimize target classification prediction. The introduction of a multi-head attention mechanism allows the model to capture global correlations among different targets, thereby fully revealing the complex pharmacological mechanisms of traditional Chinese medicine. Experiments demonstrate that, when compared to the Long Short-Term Memory (LSTM) network, Transformer model, and CNN model, the model presented in this paper outperforms all others across every evaluation metric for target classification prediction tasks, especially the accuracy of classification prediction is 0.92. and the model's F1 score of 0.895 indicates a strong balance between precision and recall. This paper also explored the connection between Chinese medicine ingredients, their targets, and the various target categories, highlighting the multi-target pharmacological properties of Chinese medicine from multiple perspectives. It offers valuable insights for understanding the therapeutic mechanisms of Chinese medicine and aids in the development of new drugs.

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

Data Mining and Traditional Chinese Medicine Pharmacology Based on Computer Algorithms

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Lihan Jiang,
  • Ying He,
  • Yu Cui,
  • Mengyu Guo

摘要

With advancements in modern pharmacology, systems biology, and deep learning (DL) technology, uncovering the intricate connections between components of traditional Chinese medicine and their targets has become a growing research focus. The pharmacological effects of traditional Chinese medicines are usually manifested as multi-target and multi-pathway effects, but traditional research methods have difficulty in feature selection, limited expression capabilities, and insufficient target classification and prediction capabilities, making it difficult to fully reveal the interactions between traditional Chinese medicine ingredients and targets. To tackle the issues mentioned above, this paper integrates Convolutional Neural Networks (CNN) and Transformer networks for joint modeling to investigate the multi-target mechanisms of traditional Chinese medicine components. The study leverages data from the public TCMSP (Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform) to optimize target classification prediction. The introduction of a multi-head attention mechanism allows the model to capture global correlations among different targets, thereby fully revealing the complex pharmacological mechanisms of traditional Chinese medicine. Experiments demonstrate that, when compared to the Long Short-Term Memory (LSTM) network, Transformer model, and CNN model, the model presented in this paper outperforms all others across every evaluation metric for target classification prediction tasks, especially the accuracy of classification prediction is 0.92. and the model's F1 score of 0.895 indicates a strong balance between precision and recall. This paper also explored the connection between Chinese medicine ingredients, their targets, and the various target categories, highlighting the multi-target pharmacological properties of Chinese medicine from multiple perspectives. It offers valuable insights for understanding the therapeutic mechanisms of Chinese medicine and aids in the development of new drugs.