量子神经网络的多发性骨髓瘤预测建模

Predictive modeling of multiple myeloma in quantum neural networks

为了提高多发性骨髓瘤的预测精度,对于多发性骨髓瘤基因表达数据库,采用BP神经网络创建多发性骨髓瘤预测模型,但其算法存在收敛速度慢、网络泛化能力差等缺点,影响预测精度。鉴于此,文中提出采用量子神经网络模型进行多发性骨髓瘤预测建模。结果表明,该预测方法较BP神经网络方法而言,其预测精度提高了8%,达到了很好的预测效果。结论与传统BP神经网络相比,其收敛精度、收敛速度有了显著提高,同时能避免陷入局部极小的缺点。

In order to improve the prediction accuracy ol multiple myeloma, for multiple myeloma geneexpression database to create predictive models ol multiple myeloma, using B P neural network algorithmslow convergence, network generalization poor shortcomings, impact prediction accuracy. In view of this,a quantum neural network model is proposed for the prediction of multiple myeloma. The results show thatthe prediction accuracy of the proposed method is improved by 8 % compared with the B P neural networkmethod. It achieves a good prediction. Compared with the traditional B P neural network, the convergenceprecision, the convergence rate is significantly improved, and at the same time it avoids falling into thelocal m i n i m u m.