篇名 | A Fast Clustering Method for Real-Time IoT Data Streams |
---|---|
卷期 | 32:1 |
作者 | Jing Sun 、 Xin Yao |
頁次 | 083-094 |
關鍵字 | PML 、 Bayesian network model 、 data streams clustering 、 dynamic sliding window 、 EI 、 MEDLINE 、 Scopus |
出刊日期 | 202102 |
DOI | 10.3966/199115992021023201007 |
As an effective way of data analysis, clustering is widely applied in the IoT based applications. By studying the related existing proposals of data clustering, a new clustering method for IoT Data streams is proposed in the present work. Firstly, the characteristics of PML documents in the process of data acquisition and identification are introduced and a hybrid PML document similarity calculation method based on the Bayesian network is developed and expected to assist in data streams clustering. Secondly, a PML data streams clustering method based on a dynamic sliding window is proposed. Finally, we evaluate the performance of our clustering method and the related methods with respect to Running time, Similarity, Purity, Entropy, and F-measure. Experimental results exhibit that the innovative clustering approach can adaptively learn from data streams that change over time, while still maintains comparable accuracy and speed.