Esophageal Cancer Risk Prediction Based on Machine Learning

Using machine learning technology to predict disease is one of the hot spots in modern medicine. Esophageal cancer is the 1 serious digestive system disease. About 220000 people die of esophageal cancer every year in the world. Early detection and diagnosis are the key to treatment. In order to diagnose esophageal cancer and precancerous lesions in the early stage, eight machine learning models based on logical regression, random forest, support vector machine, XGBoost, decision tree, gradient promotion decision tree, KNN and simple Bayesian are constructed. The model is continuously optimized by parameters and cross-validation, and the performance of the model is evaluated by using four indicators: accuracy, precision, recall and the harmonic average of the two. The results show that the classification effect of the four models of random forest, XGBoost, Decision Tree and Gradient Boosting is better than that of other models. The accuracy rate, precision rate, recall rate and F1 score of the test set are all 100, and the accuracy rate, precision rate, recall rate and F1 score of the logistic regression and SVM test set are also 85% and above, K-Nearest the classification effect of Neighbors and Naive Bayes is relatively weak. By comparing the training and test results of various machine learning algorithms, the performance of these algorithms in esophageal cancer detection is analyzed and studied, which is of great significance for the diagnosis and treatment of esophageal cancer.