Construction Of NB-IoT Terminal Date Management Platform And Data Quality Analysis

With the rapid development of the Internet of Things technology,the narrow-band Internet of Things(NB-IoT)technology,which has the advantages of low energy consumption,low cost,high coverage,and strong connection,is widely used in various terminal sensors.Due to the surge in the number of NB-IoT terminal devices,the amount of data it transmits will inevitably show explosive growth.However,due to the unevenness of the data quality of the terminal equipment,it will cause a lot of inconvenience to future data analysis,so it is particularly important to analyze the data quality of the terminal data.Classify the data according to the quality of the data,analyze the data with low data quality to find out the terminal equipment that may be abnormal,and analyze the data with high data quality to mine its huge commercial value.The main content of this paper is the quality analysis of NB-IoT terminal data.Therefore,this paper selects the decision tree C4.5 algorithm to analyze the data quality.According to the characteristics of terminal data,the decision tree C4.5 algorithm has been optimized from the perspective of simplification of calculation and selection of conditional attributes,so that it can more accurately classify data quality.This article mainly completed the following parts:(1)Introduces the methods and steps used in the data quality analysis of this article.(2)Analyze the characteristics of sensor data on the NB-IoT terminal data management platform and the deficiencies of the classic C4.5 algorithm,and propose two improvements to the classic C4.5 algorithm:?logarithm operation when calculating the information entropy gain It is converted into linear operation to reduce the complexity of the operation.?The conditional attribute that is not useful for the classification result is removed by calculating the Kendall correlation coefficient between the conditional attribute and the decision attribute.And the temperature sensor data is selected to compare the time of building decision tree and the accuracy of decision tree classification to verify that the improved C4.5 algorithm is superior to the classic decision tree C4.5 algorithm.(3)Implemented the entire NB-IoT terminal data management platform.Firstly,the platform requirements were analyzed,and then the terminal and operator docking module,platform and operator docking module,platform data codec module,platform equipment data quality analysis module,platform equipment data classification module,platform and the design and implementation of six modules including three-party service docking module.