Intelligent Analysis Of Internet Of Things Energy Consumption And Application Platform Design

With the advent of data information age, current building energy consumption have raised much attention, energy consumption data began to have a sharp growth trend with diversify forms, it is becoming more and more important to analysis and research mass energy consumption data. Data mining technology has the analysis ability of processing mass data and discovering potential useful knowledge, which provides a way for people to understand the hidden information in the data and knowledge. This paper proposed using data mining technology to analyze buildings energy consumption data, designing Internet of things buildings energy consumption data intelligently analysis platform, in order to improve the effectiveness of energy saving decisions, the main research contents as following:Firstly, association rules data mining method find interesting association or correlation from large dataset, building energy consumption sub item, building area and building person exist some connection. Apriori algorithm is a typical association rules mining algorithm,which need to generate a large number of candidate item sets and scan database many times when calculates the frequent item sets, so time and space overhead is very high. Due to sort index through index number skip to search item sets, which can effectively accelerate the information retrieval process, proposed a algorithm which based on Boolean matrix and sorting index association rule, deal with delete unused transactions and items at Boolean matrix in advance, then according to the pruning Boolean matrix and mark sequence produce second frequent item sets, frequently combined with sorting index Matrix generate other frequent item sets. Effectively improved the efficiency of mining frequent item sets, and reduce the memory usage.Secondly, due to clustering mining method as data mining tool can get the data distribution rule, through clustering algorithm clustering analysis the internal equipment data,can find the abnormal energy consuming equipment and by clustering analysis to formulate reasonable electricity plan. Density clustering algorithm can discover arbitrary shape clusters,and can identify the outliers, can effectively find the internal rules of building energy consumption, but DBSCAN algorithm at the choice of the global parameter require manual intervention and the process of regional query is complex and such query easily lose objects,improved adaptive parameters and fast regional query density clustering algorithm, improved the clustering efficiency. Affinity propagation algorithm is a fast and effective partition clustering method, which use the transfer characteristics of nearest neighbor relationship to discover clusters with lower error results in a short time, it is quite feasible to deal with massive energy consumption data. But in the absence of prior knowledge exists bias parameter choice problem and data information overlap problem in processing complex structure or high dimensional data, proposed manifold structure neighborhood selection locality preserving projections and affinity propagation algorithm, under the conditions of effectively keeping the data inner nonlinear structure, delete the redundant information in the data space, effectively improved clustering precision.Finally, design and realize internet of things building energy consumption data intelligentanalysis system, verify the feasibility of association rules and clustering data mining theory.The system consists of central server and client. The center server realized data acquisition,data mining, data analysis, data storage and data communication interface and other functions,among them data mining model contains typical data mining algorithm(association rules method and clustering method) and the improved algorithms, through the restful communication interface, which is generated by center server, the client acquired the mining data, realized the data parsing, data visualization, to provide decision support for managers,implementation of the building energy consumption data for prediction and the purpose of energy saving.