A lot of companies possess valuable data that sits there for years with no action taken on it. Machine learning (ML) technologies offer powerful algorithms and tools that enable technologists and data scientists to convert automated analysis of large-scale data into actionable business decisions. The powerful aspect of ML algorithms is that they offer automated analysis on big data to discover insights from the data. In this talk, we will give an overview of the machine learning (ML) capabilities in the Elastic Stack while focusing on common use cases. First, we will demonstrate a fundamental feature called Transforms that is used to apply transformations and analysis on continuous and static data to perform behavioral analytics. Once the data is properly structured and processed in an Elasticsearch cluster, it is possible to apply a wide variety of unsupervised and supervised ML techniques such as regression analysis, outlier detection, anomaly detection, and binary classification. These techniques are applied via several easy-to-use APIs and interfaces in Kibana to perform different tasks such as sentiment analysis, anomaly hunt, image analysis, price prediction, among others.
Presenter: Elvis Saravia, Education Engineer at Elastic