天然產物尤其是海洋天然產物是目前國內外藥物學家研究的重點。一維核磁數據（1H
NMR 和13C NMR）在天然產物結構確定中具有重要地位。本文首次將反向傳播(BP)神經網絡理論應用于13C NMR 對1H NMR 化學位移值的預測。通過對初始數據進行預處理，使其適應BP 神經網絡學習的要求，然后建立基于BP 神經網絡的1H NMR 的數據預測模型。采用實際數據對模型進行驗證。計算結果表明人工神經網絡的BP 算法可以用于天然產物一維核磁數據的預測。
關鍵詞： 1H NMR 和13C NMR；神經網絡；BP 算法；預測模型
Abreast: Natural products, especially marine natural products, are paid more attention by
pharmacolist. The data of 1H NMR and 13C NMR are the most important to determination of
structure s of natural compounds. The Back Propagation (BP) neural network theory is first used to predict the relation between the data of 1H NMR and 13C NMR. During the investigation, the original data are preprocessed to meet the requirements of the study in BP neural work, and a forecasting model based on BP neural network is established. The proposed model is then verified by used actual data. Through the results from the model, the BP neural network can be used to forecasting the data of 1D NMR.
Key words: 1H NMR and 13C NMR; neural network; BP neural network theory; forecasting model