Data-driven applications are ubiquitous (market analysis, agriculture, healthcare, transport networks, ...) and machine learning algorithms have been developed with the specific purpose of analyzing patterns and leveraging correlation within real-world measurements in order to turn data into applications. The Wolfram Language offers fully automated and highly customizable machine learning functions to perform classification, regression, clustering and many other operations. Classical methods are complemented by powerful, symbolic deep-learning frameworks and specialized pipelines for diverse data types such as image, video, text and audio.
Supervised Machine Learning »
Predict — predict values from data
Classify — classify data into categories
ActivePrediction ▪ SequencePredict ▪ Nearest ▪ FindFit ▪ ...
Unsupervised Machine Learning »
FindClusters — partition data into clusters
FeatureExtraction — find how to extract features from data
FeatureSpacePlot — visualize data in a dimension-reduced feature space
FeatureImpactPlot — visualize the impact of the input features on a model result
AnomalyDetection ▪ DimensionReduction ▪ MissingValueSynthesis ▪ ...
Neural Networks »
NetModel — collection of trained and untrained models
NetTrain — train any neural network on CPUs, GPUs, etc.
NetGraph ▪ LinearLayer ▪ ConvolutionLayer ▪ AttentionLayer ▪ ...
Computer Vision »
ImageIdentify — recognize objects in images
ImageCases ▪ FindFaces ▪ TextRecognize ▪ ImageGraphics ▪ ...
Natural Language Processing »
FindTextualAnswer — find answers to questions from text
LanguageIdentify ▪ TextStructure ▪ TextCases ▪ ...
Speech Computation »
SpeechRecognize — speech to text
AudioIdentify ▪ SpeechCases ▪ SpeakerMatchQ ▪ PitchRecognize