Research On Data Mining Technology Of Internet Of Things Based On Cloud Computing

Internet of thing(IoT)data mining has been a paramount part of IoT technology.Traditional data mining can not meet the requirements of IoT technology yet.However,cloud computing provides new opportunities for IoT data mining which has powerful computing power and the ability to store large-scale data.In this context,a cloud computing-based IoT data mining technology has emerged,which has attracted worldwide attention.This paper researches on the K-means algorithm in data mining to overcome the instability during clustering and disability in large-scale Internet of Things data.The main research contents of this paper include as follows.Firstly,the DLCK-means algorithm is proposed for solving the influence of noise point and the randomness over clustering center initialization.The algorithm combines the advantages of LCD calculation,removes noise points and determines the initial cluster center points.The experiment validates the feasibility of the algorithm.Then,in terms of k value to be specified in advance,the AK-means algorithm is proposed.The effectiveness of the algorithm is proved in experiment.Furthermore,the ALCDK-means algorithm is proposed which combines DLCK-means and AK-means algorithms,therefore it can solve the shortcomings of randomness during center initialization,noise point influence,and k-value pre-specification.The algorithm is employed on a single machine to be tested in different datasets.The comparison results show that the ALCDK-means algorithm is significantly improved compared with the K-means algorithm in respect to the clustering effect and the accuracy rate.Nonetheless,the ALCDK-means algorithm still can not deal with large-scale data sets effectively.Hence the algorithm is parallelized based on Hadoop platform,which solves the problem that traditional data mining can not handle large-scale data.Finally,the experiment shows that the ALCDK-means algorithm solves low efficiency of traditional data mining to a certain extent,which is beneficial to large-scale data mining.