Research On Diagnosis Of Breast Cancer Based On Machine Learning

Breast cancer has become the first killer of female health hazards.At present,there is a lack of the first level prediction method for breast cancer.The key to improve the cure rate and reduce the mortality of breast cancer is early detection,early diagnosis and early treatment.So early detecting,early diagnosing,early treating has become the only way.The traditional breast tumor diagnosis method of medical images is the focus part.This method relies on artificial experience for the diagnosis is subjective and lower accuraly.The diagnosis results are unreliable extremely.According to the experience,experts determined the image of the patient's tumor to classification classify.In order to improve the efficiency of diagnosis,the research methods have been developing in the direction of intelligence and tool in recent years.Artificial neural network(ANN)is a kind of self adaptively intelligent algorithm,which is widely used in the diagnosis of breast cancer.It also provides a method for the auxiliary diagnosis of doctors.With the development of artificial intelligence technology and neural network technology is mature.its classification ability is very strong and intelligence.It provides a new method for the identification of breast tumors because of the differences in the microscopic images of the breast lesions and normal tissues.According to the two kinds of images,the algorithm can be used to diagnose breast tumors.In this paper,the diagnosis and experiment of breast tumor are studied,which is based on machine learning artificial neural network.Several methods have been used to carry out simulation experiments.It can improve the accuracy of fault diagnosis.it is an effective and correct method for breast tumor diagnosis.And it has high value in medical application.The main contents of this paper include:Firstly,this paper describes the background and significance of this study.In this paper,the method of breast recognition is briefly described.It also briefly describes the main problems of the research and introduces the organization structure of this paper.Secondly,three methods of statistical analysis were used to detect breast tumor data.Three methods of statistical analysis are fisher discriminant analysis,distance discriminant analysis and bias discriminant analysis.Three methods are compared.The simulation results show that the accuracy of Fisher is 97.1%.The correct rate of distance discrimination is 84.1%.Bias discriminant accuracy was 88.41%.Compared with the three methods,fisher discriminant analysis has higher accuracy.And the probability of misclassification is the lowest.Therefore,Fisher discriminant analysis has better experimental results.Then,two methods of K-means neural network algorithm and self-organizing neural network algorithm are used to test the data of breast cancer.The correct rate of K-means neural network and self-organizing neural network is 80% and 81.58% respectively.The data used for the identification of breast tumors are 10 quantitative features of the microscopic image of the breast tumor.The data of each group include 30 data of the average value and standard deviation and the worst value of each of the nuclei in the sampled tissue.So the input dimension is larger.The redundant information is more.And the running time is longer.Using PCA to reduce the dimension of data,it is down to 8.The cumulative contribution rate has reached 99.91%.Then the experiment is carried out with K-means and self-organizing neural network.The running time is shortened.After the principal component analysis,the correct rate of K-means and self-organizing neural network is 88.95% and 88.42% respectively.Finally,genetic algorithm is used to optimize LVQ because that the LVQ algorithm is sensitive to the initial weights,in this paper.According to the application of LVQ in the diagnosis of breast cancer,an improved LVQ algorithm based on genetic algorithm is proposed for LVQ.It is improve recognition accuracy.The experimental results show that the accuracy of the improved LVQ in the diagnosis of breast tumors was 91.3%,which was 4.4%.higher than that of LVQ.