MicrosoftML: Algorithm Cheat Sheet for R / Python
This cheat sheet helps you choose the best MicrosoftML algorithm for your
predictive analytics solution. Your decision is driven by both the nature of
your data and the question you’re trying to answer.
START
Predicting
categories
TWO-CLASS CLASSIFICATION (type = “binary”) REGRESSION (type = "regression")
Accuracy, fast training, Predicting Accuracy, fast training,
rxFastTrees() / rx_fast_trees() rxFastTrees() / rx_fast_trees()
large memory footprint values large memory footprint
Fast training, Fast training,
rxFastForest() / rx_fast_forest() rxFastForest() / rx_fast_forest()
large memory footprint large memory footprint
Linear model, fast training
rxFastLinear() /rx_fast_linear() Linear model, fast training, rxFastLinear() /rx_fast_linear()
large data sets
large data sets
Accuracy, long training time, Two
rxNeuralNet() / rx_neural_net() Accuracy, long training time, rxNeuralNet() / rx_neural_net()
large data sets
large data sets
rxLogisticRegression() / Linear model, fast training
rx_logistic_regression() large data sets
Finding
rxOneClassSvm() / unusual
Highly unbalanced classes data points
rx_one_class_svm()
MULTI-CLASS CLASSIFICATION (type=”multiClass”) ANOMALY DETECTION
Accuracy, long training time, >100 features, rxOneClassSvm() /
rxNeuralNet() / rx_neural_net()
large data sets aggressive boundary rx_one_class_svm()
rxLogisticRegression() / Linear model, fast training, Three or
rx_logistic_regression() large data sets more
© 2017 Microsoft Corporation. All rights reserved.
