Mining Stream Data using k-Means clustering Algorithm

Abstract

Time-stamped sequences of data comprise what is known as stream data. Sources with fluctuating update rates and high dimensionality may be many and varied. There are instances when it is not feasible to process all of the data in a stream, and storage is also a common problem. Random sampling, sliding windows, and histograms are just a few of the ways stream data may be processed. Analysis of traffic, telecommunications, and stock market data may all be done using stream datasets. Stream data analysis often makes use of data mining methods including association analysis, classification, and clustering. The k-means clustering technique is employed in this study to mine data from a city's road traffic stream. Sliding window method is used to manage the data in the stream. Python's visualisation tools are used to depict the clusters visually. People may follow the movement of traffic with the use of real-time updates to the clusters. This paper details the findings.