Research On Neural Network Optimization Based On Quantum Genetic Algorithm And Its Application In Tumor Diagnosis

Currently,breast cancer is one of the relatively frequent diseases in the world,and its development greatly endangers the lives and physical health of female compatriots,especially among middle-aged female patients,with a high mortality rate.In medical diagnosis,due to doctors’ misdiagnosis and unclear initial symptoms of malignant breast tumors,patients miss the optimal treatment period.Therefore,it is considered to combine artificial intelligence with medicine to make medical testing of breast tumors more accurate and intelligentThe method for predicting breast malignant tumors based on BP neural network has good self-learning ability and nonlinear mapping characteristics,which can effectively predict breast malignant tumors.However,BP neural networks have the drawbacks of slow convergence speed and easy to fall into local extremum,resulting in low accuracy in predicting breast malignant tumors.Scholars are constantly trying to solve the problem of poor prediction performance of BP neural networks in the prediction process through quantum genetic algorithms,genetic algorithms,and particle swarm optimization algorithms.In recent years,quantum genetic algorithm optimized BP neural networks(QGA-BP)have been applied to the diagnosis of breast malignant tumors,combining the accuracy of local search of BP neural networks and the strong global search ability of QGA.Research has found that although the QGA-BP algorithm has fast convergence speed and strong global search ability,there are drawbacks such as lower chromosome utilization and slower convergence speed in the later stage of algorithm evolution.In view of this,based on the above research,this article considers introducing a dynamically improved quantum rotation angle strategy on the basis of quantum genetic algorithm,and incorporating quantum crossover,mutation,and catastrophic operations to optimize the weights and thresholds of BP neural network using the improved quantum genetic algorithm,and construct an improved QGA-BP model.The selected data is Wisconsin Diagnostic Break Cancer data donated by the Wisconsin Break Cancer Database of the University of Wisconsin for the UCI database.Simulation experiments were conducted to test the diagnostic effectiveness of the QGA-BP model in breast malignant tumors,and its diagnostic rate was significantly higher than machine learning algorithms such as GA-BP algorithm,BP neural network,SVM,LVQ neural network,and decision tree.It can reduce the risk of delayed or undiagnosed diagnosis and assist doctors in early diagnosis.