Outlier Detection of Data in Wireless Sensor Networks Using Kernel Density Estimation

This article addresses the problem of outlier detection for wireless sensor networks. As increasing amounts of observational data are tending to be high-dimensional and large scale, it is becoming increasingly difficult for existing techniques to perform outlier detection accurately and efficiently. Although dimensionality reduction tools (such as deep belief network) have been utilized to compress the high-dimensional data to support outlier detection, these methods may not achieve the desired performance due to the special distribution of the compressed data. Furthermore, because most existed classification methods must solve a quadratic optimization problem in their training stage, they cannot perform well in large-scale datasets. In this article, we developed a new form of classification model called “deep belief network online quarter-sphere support vector machine,” which combines deep belief network with online quarter-sphere one-class support vector machine. Based on this model, we first propose a model training method that learns the radius of the quarter sphere by a sorting method. Then, an online testing method is proposed to perform online outlier detection without supervision. Finally, we compare the proposed method with the state of the arts using extensive experiments. The experimental results show that our method not only reduces the computational cost by three orders of magnitude but also improves the detection accuracy by 3%–5%.

Download Full-text

Outlier Detection Using Improved Support Vector Data Description in Wireless Sensor Networks

Sensors ◽

10.3390/s19214712 ◽

2019 ◽

Vol 19 (21) ◽

pp. 4712

Author(s):

Pei Shi ◽

Guanghui Li ◽

Yongming Yuan ◽

Liang Kuang

Keyword(s):

Wireless Sensor Networks ◽

Sensor Networks ◽

Relative Density ◽

Outlier Detection ◽

Sensor Data ◽

Wireless Sensor ◽

Support Vector ◽

Support Vector Data Description ◽

Vector Data ◽

Data Description

Wireless sensor networks (WSNs) are susceptible to faults in sensor data. Outlier detection is crucial for ensuring the quality of data analysis in WSNs. This paper proposes a novel improved support vector data description method (ID-SVDD) to effectively detect outliers of sensor data. ID-SVDD utilizes the density distribution of data to compensate SVDD. The Parzen-window algorithm is applied to calculate the relative density for each data point in a data set. Meanwhile, we use Mahalanobis distance (MD) to improve the Gaussian function in Parzen-window density estimation. Through combining new relative density weight with SVDD, this approach can efficiently map the data points from sparse space to high-density space. In order to assess the outlier detection performance, the ID-SVDD algorithm was implemented on several datasets. The experimental results demonstrated that ID-SVDD achieved high performance, and could be applied in real water quality monitoring.

Download Full-text

Outlier Detection in Wireless Sensor Networks using Machine Learning Techniques: A Survey

2020 International Conference on Electrical and Electronics Engineering (ICE3) ◽

10.1109/ice348803.2020.9122865 ◽

2020 ◽

Cited By ~ 1

Author(s):

Rajendra Kumar Dwivedi ◽

Arun Kumar Rai ◽

Rakesh Kumar

Keyword(s):

Machine Learning ◽

Wireless Sensor Networks ◽

Sensor Networks ◽

Outlier Detection ◽

Machine Learning Techniques ◽

Wireless Sensor ◽

Learning Techniques

Download Full-text

Outlier Detection in Wireless Sensor Networks Using Model Selection-Based Support Vector Data Descriptions

Sensors ◽

10.3390/s18124328 ◽

2018 ◽

Vol 18 (12) ◽

pp. 4328 ◽

Cited By ~ 5

Author(s):

Zhan Huan ◽

Chang Wei ◽

Guang-Hui Li

Keyword(s):

Wireless Sensor Networks ◽

Sensor Networks ◽

Model Selection ◽

Outlier Detection ◽

Sensor Data ◽

Wireless Sensor ◽

Quality Data ◽

Support Vector ◽

Support Vector Data Description ◽

Vector Data

Wireless sensor networks (WSNs) are often deployed in harsh and unattended environments, which may cause the generation of abnormal or low quality data. The inaccurate and unreliable sensor data may increase generation of false alarms and erroneous decisions, so it’s very important to detect outliers in sensor data efficiently and accurately to ensure sound scientific decision-making. In this paper, an outlier detection algorithm (TSVDD) using model selection-based support vector data description (SVDD) is proposed. Firstly, the Toeplitz matrix random feature mapping is used to reduce the time and space complexity of outlier detection. Secondly, a novel model selection strategy is realized to keep the algorithm stable under the low feature dimensions, this strategy can select a relatively optimal decision model and avoid both under-fitting and overfitting phenomena. The simulation results on SensorScope and IBRL datasets demonstrate that, TSVDD achieves higher accuracy and lower time complexity for outlier detection in WSNs compared with existing methods.

Download Full-text