Machine Learning Algorithms – DenStream Clustering

SAP HANA SPS11 introduces two machine learning algorithms that can be used in streaming projects: Adaptive Hoeffding Tree and DenStream Clustering. Integrating machine learning algorithms with smart data streaming combines supervised learning and unsupervised learning such that one can efficiently train data models in real-time.

This tutorial will walk you through a demo of the DenStream Clustering machine learning algorithm.

Summary of DenStream Clustering:

• Analyzes continuous streams of data in real-time
• Forms data clusters of any shape and handles outliers
  • There are core-micro-clusters and outlier-micro-clusters
• Changes the weight of clusters over time (depending on the weight, outlier-micro-clusters can become core-micro-clusters)
• Updates so only necessary information is kept

Applications of DenStream Clustering:

• Analyzing data streams from sensors on Internet of Things (IoT) devices
• Identifying patterns in spending money (helps detect fraud)
• Grouping information for search engines
• Mapping demographics of consumers to types of purchases

View the tutorial here: Machine Learning Algorithms - DenStream Clustering