الگوریتم خوشه بندی مبتنی بر همبستگی داده های چندمودی در اینترنت اشیا
Abstract
I- Introductio
II- System Model and Problem Statement
III- Device-Level Multimodal Data Correlation Mining and Clustering Model
IV- Device Clustering Algorithm Based on Multimodal Data Correlation
V- Simulation and Result
VI- Conclusion
References
Abstract
With the development of information network, the popularity of Internet of Things (IoT) is an irreversible trend, and the intelligent demands for IoT is becoming more and more urgent. How to improve the cognitive ability of IoT is a new challenge and therefore has given rise to the emergence of cognitive IoT (CIoT). In this paper, a device-level multimodal data correlation mining model is first designed based on the canonical correlation analysis to transform the data feature into a subspace and analyze the data correlation. The correlation of the device is obtained based on the comprehensive of data correlation and the location information of the device. Then a heterogeneous clustering model (heterogeneous device clustering) is proposed by using the result of the correlation analysis to classify the device. Finally, we propose a device clustering algorithm based on multimodal data correlation for CIoT, which combines the functions of multimodal data correlation analyze with device clustering. Extensive simulations are carried out and our results show that the proposed algorithm can effectively improve the quality of data transmission and the intelligent service.
Introduction
The concept of Internet of Things (IoT) is proposed since 1999 [1], which is a technological revolution that brings us into an era of ubiquitous computing and communication. Meanwhile, cognitive IoT (CIoT) emerges to meet the current application requirements and becomes the development trend of IoT. And the center of IoT is shifted from connective to cognitive. The main idea of CIoT enables the traditional IoT to possess the features of self-sensing, decision-making, selflearning and self-adjusting intelligent. [2] proposed a view that CIoT has the ability to combine the physical world (such as goals, sources, etc.) with the social world (social behavior, user needs, etc.) to enhance the relation among intelligent resources allocation, automatic network operation and intelligent service provision. The research on CIoT is still in the development stage compared with IoT. [3] [4] proposed a cognitive management framework which could support the development of the sustainable intelligent city better than before. CIoT is regarded as an advance direction which is able to improve the performance and realize intellectualization to the current IoT [5] [6]. The data of CIoT is collected from multiple heterogeneous devices and different domains, such as numerical observations, the measurements from different devices or text from social media stream [7]. In order to meet the social enterprise needs and extract more valuable data information by mining the data correlation, some algorithms about data correlation and data clustering are studied to solve a practical application problem. [8] proposed a novel fusion learning framework which pays attention to cross-retrieval. The aim of the framework is retrieving the similar data from other types data by regarding a type of data as a query. For example, user retrieves relevant text and video by using a single picture. [9] described three clustering algorithms to analyze the data correlation of the user online behavior, which could solve the problem among clustering, person query, and social network prediction. [10] designed a novel CCA framework which combines CCA algorithm and norm-one regularization technology [11] [12], the CCA framework can extract relevant sensing data and cluster them into different clusters. [13] proposed a mobility prediction-based clustering scheme to solve the high mobility of nodes in ad hoc networks, which consists of two parts: the initial clustering stage and the cluster maintenance stage. [14] proposed an incremental clustering algorithm (ICFSKM) based on K-medoids, which can quickly find and discover the nodes with the density peak. [15] proposed a new heuristic clustering algorithm for numerical data, which aims to maximize DI (Dunn Index) [16] [17] or CHI (Calinshi Harabasz Index) [18]. [7] proposed an adaptive clustering method to design dynamic IoT data stream, the method is suitable for the underlying data drift of the data stream and can determine the number of clusters based on the data distribution, then an online clustering mechanism is used to cluster the input data stream. However, the above researches exist the defect that the processed data do not contain cognitive components and can not handle the data generated with high mobility.