摘要随着物联网技术的不断发展和成熟，物联网技术被逐渐应用到了城市管理当中，逐渐诞生了智慧型城市的概念。在智慧型城市系统中，传感器网络的规模得到了空前的扩展，网络所要检测的范围扩展至整个城市，检测的属性内容相当丰富。在系统运行的过程中，传感器网络时刻都在提交着大量的传感器数据。这些原始数据当中不可避免的存在着各种误差甚至是错误。如果将这些数据直接提交给数据库作为系统高层应用运行的依据必将造成很多错误的决策。因此必须对传感器提交的数据进行清洗，从中得到能够准确反映城市状态的准确的数据集合，从根本上保证系统的有效性。本毕业论文充分分析了智慧型城市应用中对于传感器所提交数据的要求，对此应用背景下被监测属性的变化特点，也就是传感器数据的变化特点进行详细分析和总结。结合应用背景下的数据特征设计了满足系统实时性要求的基于回归方程和时间窗口的数据清洗算法。该算法在清洗过程中充分利用了智慧型城市应用中传感器数据采样频率高，数据密度大连续变化的特征。除此之外还对智慧城市中传感器数据清洗的类型层次进行了细致的分析设计了两级层次化的传感器数据清洗过程。在课题研究过程中通过理论分析和编写实际的清洗程序在实际的应用场景中对研究成果进行真实环境的下的验证这两种方式对研究成果进行了评估。结果证明本文设计的清洗方式在智慧型城市中能够达到很好地数据清洗效果。8408
关键词  数据清洗 数据拟合 线性回归 时间窗口 前向预测
毕业设计说明书（论文）外文摘要
Title      Collecting and Cleaning of Sensor Data in The Context of Intelligent Cities                                       
Abstract
Presently, as the development and maturity of Internet of Things grows. It is gradually applied to urban management, which has gradually gave birth to the concept of intel- ligent cities. In the intelligent urban management system, the size of the sensor network has been an unprecedented expansion. ll events that can be used to urban decision are detected by the sensor network. Sensor are used to sense many tings. during the opera- ti-on of the intelligent urban management system,the sensor network continuously sub- mit a large number of sensor data to the system. Among the data, there are all kinds of inevitable errors or mistakes. If the high-level application just uses the data directly, it can cause a lot of problems. Therefore the data submitted by the sensor must be clean- ed.This paper has analyzed the requirements for the sensor data submitted in the intelli- gent city applications. Based on the result this paper introduces some useful methods to clean sensor data. In these methods we clean the sensor by time window technique and Linear regression technique. We build a two level data cleaning system model. To veri- fy the efficiency of the method we also design and conduct experiments. The results sh- ow that the cleaning method proposed in this paper,can achieve good data cleansing effect.
Keywords data cleaning,data fitting,linear regression,time window,         forecast
目录
1 绪论    5
1.1课题来源以及介绍    6
1.2相关技术    7
2 数据清洗介绍    9
2.1数据清洗概述    9
2.2.1 国外研究现状    11
2.2.2 国内研究现状    12
2.3 数据清洗定义    12
2.4 数据清洗研究对象    13 智慧城市管理系统中的传感器数据采集与清洗方法研究:http://www.751com.cn/jisuanji/lunwen_6748.html