改进的量子遗传算法优化BP神经网络的恶性肿瘤诊断

An Improved Quantum Genetic Algorithm Optimized BP Neural Network for Malignant Tumor Diagnosis

乳腺恶性肿瘤对人们的生命构成重大威胁,影像是目前的重要诊断手段,早期诊断与预测模型一般以图像数据作为输入,但在图像转换过程中出错的可能性较大。而人工智能技术的飞速发展可为提高恶性肿瘤识别的准确率提供技术手段和思想创意。为提高乳腺恶性肿瘤识别的准确率,克服传统反向传播(Back Propagation,BP)神经网络和单一遗传算法的缺点,构建一种改进的量子遗传算法(Quantum Genetic Algorithm,QGA)优化BP神经网络的模型,即在QGA算法中引进动态改进旋转角策略和加入遗传算法的交叉变异操作,再利用改进的QGA算法优化多目标BP神经网络的权值和阈值。以美国威斯康辛州女性乳腺癌肿瘤数据集为例,对其进行分析。实验表明,QGA-BP模型在乳腺肿瘤诊断上具有收敛速度快且预测精度高的优点,能够准确诊断恶性肿瘤。

Breast malignant tumors pose a major threat to human life.Imaging is an important diagnostic means at present.Early diagnosis and prediction models generally use image data as input,but there is a high probability of errors in the process of image conversion.The rapid development of artificial intelligence technology provide technical means and ideas for improving the accuracy of malignant tumor identification.In order to improve the accuracy of breast cancer identification and overcome the shortcomings of traditional Back Propagation(BP)neural network and single genetic algorithm,an improved Quantum Genetic Algorithm(QGA)model for optimizing BP neural network is constructed,that is,the dynamic improved rotation angle strategy is introduced into the QGA algorithm and the crossover mutation operation of the genetic algorithm is added,and then the improved QGA algorithm is used to optimize the weights and thresholds of the multi-objective BP neural network.The data set of female breast cancer in Wisconsin,USA was taken as an example to analyze it.The experiment shows that the QGA-BP model has the advantages of fast convergence and high prediction accuracy in breast tumor diagnosis,and can accurately diagnose malignant tumors.