Big Data Clustering Analysis Algorithm for Internet of Things Based on K-Means

Introduction

The “Internet of things” refers to that a variety of hardware devices and the Internet are combined through the information sensing device to form a huge network of connected objects. The sensing device includes RFID (Radio Frequency IDentification) device, infrared sensor, laser scanner, two-dimensional code and so on. Among them, the RFID sensor technology is widely used in various application fields. The analysis of Internet of things big data can start from the formed complex event stream of RFID, mainly using the composite events on the formation of RFID complex CEP technology. Based on the data source, we establish a certain relationship to its model, and then the clustering analysis of the formed event relationship model is carried out.

Clustering analysis is a commonly used technology for analyzing and processing data in data mining and machine learning algorithms. The main principle is that, in accordance with the relevant characteristics of different objects in data sets, the similar objects are grouped into the same class, while the greatly different objects are divided into different classes. By this method, we can find the inner relations among data, and find some useful information related to each other in different types of data, so as to provide support for the relevant decision-making. For the Internet of things, the processing of big data clustering analysis can be divided into two steps: complex event flow data cleaning and data analysis. In the process of data cleaning, we can remove label data which do not conform to the corresponding specifications, and at the same time, we can simulate and generate data to make up.

According to the data clustering analysis, based on the current heterogeneous, multi-dimensional data networking, in solving the problems of big data technology, the industry is less involved in data analysis content, form and other aspects. However, the long-term development of big data will depend on the depth of data analysis, and the ability of data analysis and data processing will determine how far the big data can go. Therefore, the proposal of CEP (Complex Event Processing) technology has greatly improved the probability of a fast and effective solution to this problem. Nevertheless, when the event flow forms big data, there are many shortcomings in the current effective method and technology to obtain the event pattern relationship.