SOMKE: Kernel Density Estimation Over Data Streams by Sequences of Self-Organizing Maps

PROJECT TITLE :

SOMKE: Kernel Density Estimation Over Data Streams by Sequences of Self-Organizing Maps

ABSTRACT:

In this paper, we propose a completely unique technique SOMKE, for kernel density estimation (KDE) over data streams based mostly on sequences of self-organizing map (SOM). In several stream Data Mining applications, the ancient KDE strategies are infeasible because of the high computational cost, processing time, and memory demand. To scale back the time and area complexity, we propose a SOM structure in this paper to obtain well-outlined information clusters to estimate the underlying chance distributions of incoming data streams. The most idea of this paper is to make a series of SOMs over the data streams via two operations, that is, making and merging the SOM sequences. The creation phase produces the SOM sequence entries for windows of the info, which obtains clustering information of the incoming information streams. The size of the SOM sequences can be further reduced by combining the consecutive entries within the sequence primarily based on the measure of Kullback–Leibler divergence. Finally, the chance density functions over arbitrary time periods along the information streams will be estimated using such SOM sequences. We have a tendency to compare SOMKE with 2 different KDE methods for information streams, the M-kernel approach and the cluster kernel approach, in terms of accuracy and processing time for various stationary knowledge streams. Furthermore, we also investigate the use of SOMKE over nonstationary (evolving) knowledge streams, including a artificial nonstationary knowledge stream, a true-world monetary information stream and a group of network traffic data streams. The simulation results illustrate the effectiveness and efficiency of the proposed approach.

Did you like this research project?

To get this research project Guidelines, Training and Code... Click Here