streamDM 0.0.1

albert bifet, silviu maniu, Jianfeng Qian, Guangjian Tian — Tue, 28 Apr 2015 12:34:00 -0000

Due to the large amount of data that is created -- and needs to be processed -- in real-time streams, methods on such streams need to be extremely time-efficient while using very small amounts memory. streamDM includes advanced stream mining algorithms, and is intended to be the gathering point of practical implementation and deployments for large-scale data streams.

This new library will contain methods for classification, regression, clustering and frequent pattern mining. In its current iteration, it contains Stochastic Gradient Descent, Perceptron, Naive Bayes for classification, and CluStream for clustering streams.

mloss.org streamDM

streamDM 0.0.1