In this paper, the incremental recu

In this paper, the incremental recursive least squares regression parameter estimation algorithm is combined with the generalized likelihood ratio change point detection algorithm. As a result, a data stream real-time trend extraction algorithm is proposed. This algorithm uses the incremental mechanism to determine the data for the continuously arriving data stream elements. Sequence regression model parameters and segmentation points, real-time extraction of data stream trend characteristics. The algorithm not only has faster calculation speed, but also has higher precision. Storm cluster is similar to the Hadoop cluster. There are two kinds of nodes in the cluster: master node and worker node. In summary, through the introduction of real-time computing and related technologies of distributed systems, the discussion of the basic concepts, core concepts, operating mechanisms, and programming models of storm makes us have a more comprehensive understanding of storm's stream data processing system. The data analysis and processing system based on Storm is more advantageous than the data analysis and processing system based on the traditional solution in terms of efficiency, real-time performance, scalability, and availability. Using the framework provided by storm to write real-time data processing applications greatly reduces the complexity of development and deployment. Developers no longer need to build their own message queues and message processing mechanisms and compose real-time processing networks, but only they need to focus on business data.