Analysis Of Breast Cancer Pathological Image Based On Ensemble Learning

In recent years,the incidence of breast cancer has increased rapidly,breast cancer has now become the most common cancer in the world.If breast cancer is detected at an early stage,it is usually treatable.If breast cancer is not detected and treated as early as possible,it will get worse.The pathological image of the breast reflects the information of the breast,which helps doctors analyze the patient’s condition and plays an important role in the diagnosis and treatment of breast cancer.The analysis of breast pathological images requires experienced doctors to combine various information in the images to observe and study repeatedly,which is time-consuming and labor-intensive.There are few experienced doctors,and the manual observation of pathological images is time-consuming.If the doctors are tired,they may miss important indicators and cause misdiagnosis.With the development of computer technology,the accumulation of medical data and the improvement of medical technology,artificial intelligence have gradually been applied to the medical field,assisting doctors in analyzing medical data,providing doctors with diagnostic opinions and improving the efficiency of diagnosis.Ensemble learning combines the results of multiple base learners,can predict data more accurately and improve model performance,and has demonstrated its satisfactory results and potential in many aspects.In the medical field,the application of ensemble learning is currently in the research and exploration stage.Most of the existing models for breast cancer pathological image analysis are based on a single model,and there are few studies on the application of ensemble learning to breast cancer pathological images.This paper comprehensively considers the diversity of images and the diversity of models,constructs a breast cancer pathological image analysis model based on ensemble learning,fully integrates the characteristics of different models,and enhances the diversity and accuracy of the model.In the analysis of breast cancer pathological images,this paper uses single machine learning and deep learning methods to build models,and then uses ensemble learning methods to fuse information from multiple models and improve them to obtain a more accurate ensemble learning model.First,this paper constructs five different machine learning models:SVM,K-nearest neighbor,decision tree,random forest,and Adaboost.Then,use the ensemble learning method to integrate five machine learning models,build an ensemble learning model,and compare it with the five models.The results show that the ensemble learning model has a higher accuracy.In order to improve the accuracy of the analysis of breast cancer pathological images,this paper also ensemble convolutional neural networks and constructs an ensemble learning model.Firstly,VGG16,ResNet50 and DenseNet121 convolutional neural networks are constructed based on transfer learning.The preprocessed images are input into three neural networks for training and testing.Then use the ensemble learning method to integrate the three networks to improve the accuracy of the model’s analysis of breast cancer pathological images.After that,the above-mentioned ensemble neural network model was improved by using ensemble learning ideas,and an improved ensemble learning model with better effect was obtained.Use ensemble learning ideas to improve the VGG16 network,build an ensemble VGG16 network,which improve its accuracy while enhancing its generalization ability.Then,combine the improved ensemble VGG 16 network with ResNet50 and DenseNet 121 networks to construct a new ensemble learning model,and compare and analyze it with the initially constructed ensemble learning model.The results prove that its performance has indeed improved.This paper uses the ensemble method to optimize the single machine learning and deep learning models respectively,and builds the ensemble learning model.The experimental results prove that the accuracy of the ensemble learning model and its generalization ability have been effectively improved.The analysis of pathological images of breast cancer based on ensemble learning can assist doctors in diagnosis,improve their work efficiency,and reduce cases of misdiagnosis or missed diagnosis due to the lack of experienced doctors.