• 文档
main content

分类学习器 -凯发k8网页登录

以交互方式训练、验证和调整分类模型

可以选择各种算法来训练和验证二类问题或多类问题的分类模型。训练多个模型后，可以横向比较它们的验证误差，然后选择最佳模型。要帮助您确定使用哪种算法，请参阅train classification models in classification learner app。

此流程图显示在分类学习器中训练分类模型或分类器的常见工作流。

workflow in the classification learner app. step 1: select data and validation. step 2: choose classifier options. step 3: train a classifier. step 4: assess classifier performance. step 5: export the classifier.

如果您要使用您在分类学习器中训练的模型之一来运行试验，您可以将该模型导出至试验管理器。有关详细信息，请参阅。

app

分类学习器使用有监督的机器学习训练模型以对数据进行分类
design and run experiments to train and compare machine learning models

主题

常见工作流

  • train classification models in classification learner app
    workflow for training, comparing and improving classification models, including automated, manual, and parallel training.

  • import data into classification learner from the workspace or files, find example data sets, choose cross-validation or holdout validation options, and set aside data for testing. alternatively, open a previously saved app session.

  • in classification learner, automatically train a selection of models, or compare and tune options in decision tree, discriminant analysis, logistic regression, naive bayes, support vector machine, nearest neighbor, kernel approximation, ensemble, and neural network models.

  • compare model accuracy values, visualize results by plotting class predictions, and check performance per class in the confusion matrix.

  • after training in classification learner, export models to the workspace, generate matlab® code, generate c code for prediction, or export models for deployment to matlab production server™.
  • train decision trees using classification learner app
    create and compare classification trees, and export trained models to make predictions for new data.

  • create and compare discriminant analysis classifiers, and export trained models to make predictions for new data.

  • create and compare binary logistic regression classifiers, and export trained models to make predictions for new data.

  • create and compare naive bayes classifiers, and export trained models to make predictions for new data.

  • create and compare support vector machine (svm) classifiers, and export trained models to make predictions for new data.

  • create and compare nearest neighbor classifiers, and export trained models to make predictions for new data.

  • create and compare kernel approximation classifiers, and export trained models to make predictions for new data.

  • create and compare ensemble classifiers, and export trained models to make predictions for new data.

  • create and compare neural network classifiers, and export trained models to make predictions for new data.

自定义工作流


  • identify useful predictors using plots or feature ranking algorithms, select features to include, and transform features using pca in classification learner.

  • before training any classification models, specify the costs associated with misclassifying the observations of one class into another.

  • create classifiers after specifying misclassification costs, and compare the accuracy and total misclassification cost of the models.
  • hyperparameter optimization in classification learner app
    automatically tune hyperparameters of classification models by using hyperparameter optimization.

  • train a classification support vector machine (svm) model with optimized hyperparameters.

  • import a test set into classification learner, and check the test set metrics for the best-performing trained models.

  • determine how features are used in trained classifiers by using partial dependence plots.

  • export and customize plots created before and after training.

  • train a classification model using the classification learner app, and generate c/c code for prediction.

  • this example shows how to train a binary glm logistic regression model using classification learner, and then generate c code that predicts labels using the exported classification model.

  • train a model in classification learner and export it for deployment to matlab production server.

  • train a binary classification model using classification learner app to detect anomalies in sensor data collected from an industrial manufacturing machine.

试验管理器工作流


  • export a classification model to experiment manager to perform multiple experiments.

  • use different training data sets, hyperparameters, and visualizations to tune an efficient linear classifier in experiment manager.

相关信息

网站地图