Breast Cancer Analysis And Predictive Diagnosis Based On Data Mining

The incidence rate and mortality of breast cancer are gradually increasing in today's society.Due to its difficult to ascertain the nosogenesis and disease is difficult to discover,so early breast cancer is easily neglected by patients.Many primary breast cancer patients' symptoms worsen because they miss the diagnosis time,and the hospital examination process has the disadvantages of low efficiency and strong subjectivity.In recent years,with the development of computing science and medical technology,the data level has also increased rapidly,big data statistics has gradually become a hot topic in today's research,but it also adds a lot of pressure and difficulties to the work of medical personnel.Medical data is different from general data.It has a large scale and exaggerated growth rate.It has a wide variety and contains a lot of repeated information or irrelevant information.This paper first describes the current situation of data mining application in breast cancer research,the data comes from UCI open database,and the classification supervised learning algorithm is selected according to the data type,Logistic Regression model was used to predict the clinical treatment of breast cancer,the factor analysis method is integrated to screen the factors with large proportion or influence from many attributes,as the test attribute of the model.The experimental results show that the logistic regression model has good performance,however,the model is prone to fitting and instability,and the classification effect needs to be further strengthened.In order to further enhance the precision rate of prediction,come up with a support vector machine model based on RFECV feature selection.Firstly,the algorithm uses RFECV feature selection to screen the sample attributes,select the optimal number of characteristic variables and characteristic variables.After standardization,select the optimal kernel function and parameters.Experiments show that SVM can better predict the nature of breast cancer.