Elastic Machine Learning is a machine learning platform integrated into the Elastic Stack that provides both unsupervised and supervised learning capabilities for data analysis and pattern detection. The platform offers two types of unsupervised machine learning: anomaly detection for time series data that constructs probability models to identify unusual events and forecast future behavior, and outlier detection for non-time series data that identifies unusual data points by analyzing proximity and cluster density. For supervised learning, the platform provides classification capabilities to predict discrete categorical values such as identifying malicious versus benign domains, and regression analysis to predict continuous numerical values like web request response times. Both supervised methods require training data sets and produce trained models that can be deployed for predictions on new data. Feature availability varies by project type within the Elastic ecosystem. Elasticsearch Serverless projects include trained models, Observability projects support anomaly detection jobs, and Elastic Security projects provide access to anomaly detection jobs, data frame analytics jobs, and trained models. The platform includes functionality for exporting and importing machine learning job configurations and datafeed details between environments, though models must be rebuilt for anomaly detection jobs in new environments. Data frame analytics trained models are portable and can be transferred between clusters. Jobs can be managed through APIs or the Kibana interface, with automatic synchronization of saved objects for visualization and management purposes.