Icon

5. Gaussian naive bayes classification

Simple Model Training for Naive Bayes ClassificationThis workflow demonstrates how a naive bayes classifier is built and applied to new data. It also illustrates theuse of KNIME's hiliting capabilities, which allow interactive views to be connected within the same workflow.Task Predict the whether the species has virus or not from demographic attributes of the west nile virus data set Data ReadingRead the new file 'westnilevirus.csv'created after data preprocessing. It contains:1. Trap Type2. Number of Mosquitoes3. Species4. Test Results Data PartitioningCreate two separate on data set1. training set (80%) 2. test set (20%). Train a ModelThis node builds a Naive BayesModel. Most Learner nodes output aPMML model (blue square outputport). Apply the ModelPredictor nodes apply a specificmodel to a data set and append themodel predictions. Score the ModelCompute a confusion matrixbetween real and predicted classvalues and calculate the relatedaccuracy measures. ROC Curve Evaluate predictions on basis ofROC Curve training set test set Reading westnilevirus.csvRandom drawing 80% upper port20% lower portBuild ROC for predictionsTarget Class = ResultApply the trained modelto predict ResultsMatch original vs. predicted Result valuesFile Reader Partitioning ROC Curve (local) Naive Bayes Learner Naive BayesPredictor Scorer Simple Model Training for Naive Bayes ClassificationThis workflow demonstrates how a naive bayes classifier is built and applied to new data. It also illustrates theuse of KNIME's hiliting capabilities, which allow interactive views to be connected within the same workflow.Task Predict the whether the species has virus or not from demographic attributes of the west nile virus data set Data ReadingRead the new file 'westnilevirus.csv'created after data preprocessing. It contains:1. Trap Type2. Number of Mosquitoes3. Species4. Test Results Data PartitioningCreate two separate on data set1. training set (80%) 2. test set (20%). Train a ModelThis node builds a Naive BayesModel. Most Learner nodes output aPMML model (blue square outputport). Apply the ModelPredictor nodes apply a specificmodel to a data set and append themodel predictions. Score the ModelCompute a confusion matrixbetween real and predicted classvalues and calculate the relatedaccuracy measures. ROC Curve Evaluate predictions on basis ofROC Curve training set test set Reading westnilevirus.csvRandom drawing 80% upper port20% lower portBuild ROC for predictionsTarget Class = ResultApply the trained modelto predict ResultsMatch original vs. predicted Result valuesFile Reader Partitioning ROC Curve (local) Naive Bayes Learner Naive BayesPredictor Scorer

Nodes

Extensions

Links