Critical analysis of smart environment sensor data behavior pattern based on sequential data mining techniques
| DOI | https://doi.org/10.1108/IMDS-12-2014-0386 |
| Published date | 13 July 2015 |
| Date | 13 July 2015 |
| Pages | 1151-1178 |
| Author | Gebeyehu Belay Gebremeskel,Chai Yi,Chengliang Wang,Zhongshi He |
| Subject Matter | Information & knowledge management,Information systems,Data management systems |
Critical analysis of smart
environment sensor data behavior
pattern based on sequential data
mining techniques
Gebeyehu Belay Gebremeskel and Chai Yi
College of Automation, Chongqing University, Chongqing, China, and
Chengliang Wang and Zhongshi He
College of Computer Science, Chongqing University, Chongqing, China
Abstract
Purpose –Behavioral pattern mining for intelligent system such as SmEs sensor data are vitally
important in many applications and performance optimizations. Sensor pattern mining (SPM) is also
dynamic and a hot research issue to pervasive and ubiquitous of smart technologies toward improving
human life. However, in large-scale sensor data, exploring and mining pattern, which leads to detect the
abnormal behavior is challenging. The paper aims to discuss these issues.
Design/methodology/approach –Sensor data are complex and multivariate, for example, which
data captured by the sensors, how it is precise, what properties are recorded or measured, are
important research issues. Therefore, the method, the authors proposed Sequential Data Mining (SDM)
approachto explore patternbehaviors towarddetecting abnormalpatterns for smartspace fault diagnosis
and performance optimization in the intelligent world. Sensor data types, modeling ,d escriptions and SPM
techniques are discussed in depth using real sensor data sets.
Findings –The outcome of the paper is measured as introducing a novel idea how SDM technique’s scale-
up to sensor data pattern mining. In the paper, the approach and technicality of the sensor data pattern
analyzed, and finally the pattern behaviors detected or segmented as normal and abnormal patterns.
Originality/value –The paper is focussed on sensor data behavioral patterns for fault diagnosis and
performance optimizations. It is other ways of knowledge extraction from the anomaly of sensor data
(observation records), which is pertinent to adopt in many intelligent systems applications, including
safety and security, efficiency, and other advantages as the consideration of the real-world problems.
Keywords Activity description, Behavioral pattern, Data mining, Intelligent systems, Smart space,
Trajectories
Paper type Research paper
I. Introduction
Critical analysisof Smart Environment (SE) sensor data is a technological explorationof
intelligent system’s functionality and performance (Brian and C.J.H., 2007). Sensor
technologies are related to human cognitive capture and visualize behavioral patterns,
which adopted in almost every modern intelligent system. Personal (smart home), safety
(traffic management, military security), healthcare (cognitive behavior), business (sales
track) are few domains. Furthermore, industries (architectural control), environmental
monitoring, and location-aware services are an additional potential area of smart
technology applications. In these applications, sensors captured various properties of
Industrial Management & Data
Systems
Vol. 115 No. 6, 2015
pp. 1151-1178
©Emerald Group Publis hing Limited
0263-5577
DOI 10.1108/IMDS-12-2014-0386
Received 29 December 2014
Revised 25 March 2015
6 May 2015
18 May 2015
Accepted 18 May 2015
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
The authors are very thankful to the anonymous reviewers for their useful comments. The works
is supported by the National Natural Science Foundation of China under Grant No. 61004112 and
the Third Stage Building of “211 Project“under Grant No. S-10218 and the project of Innovative
Team Building under the Grant No. 2009091011.
1151
Smart
environment
sensor data
physical phenomena become a huge sensor data, which is challenging to explore and
mine behavioral patterns in a traditional approach. The problem is more noticeable to
analysis the hidden data relationship and searches useful information (Tao et al.,2009;
Joëlle et al.,2005).
Behavioral Pattern Mining (BPM) is the process of segmenting the signal or sensors’
paths, which are an essential to reveal, object trajectories and characteristics for fault
detections. The approach is an investigation of smart space applications in relation
to sensor technologies’performances. Sensor data captured from diverse sensor devices
and annotated as the objects’situations or moments. Therefore, data schemas, precision
or accuracy, units of measurement factors are a seductive issue of BPM ( Jae-Gil et al.,
2007). However,the behavioral pattern in large-scale sensordata is challenging. Since the
data do not have common features, and the process of data synthesizing is complex as a
set of instructions and event patterns (Elad, 2004). The issues can be summarized as: as
sensor technologies, and applicationsare increasing, sensor data handling and analyzing
become a challenge; how sensor data behavioral patterns are developed by considering
the domain contexts and actuators’behaviors towards abnormal behavior pattern
detections and synthesizing?; sensor data are not completely stored, which involves
real-time activities. How to combine and analyzes trend and active sensor data?; and
what analytic tool is more capable to explore sensor data as its types (volume, velocity,
variety, etc.), complexity, sensitive, and massiveness and other behavioral factors?
In this paper, we proposed Sequential Data Mining (SDM) technique for a critical
analysis of sensor data behavioral pattern towards Abnormal Behavior Pattern (ABP)
detection. The approach is an arrangement of instances or events in a sequentialmanner
or standard to define the objects’trajectories behaviors. It is the process of exploring the
data model, which designin a given sequence of sensordata to define objects’properties
how and why deviate from its normal conditions. Moreover, an intelligent system is the
agglomerationof advanced IT and smart technologies, whichis pertinent to the dynamic
applications of SDM.The technique of SDM for sensor design, pervasivecomputing and
pattern mining (Parisa and Diane, 2011) is a systematic approach demystifythe process.
It is essential to address each sequence of the behavioral pattern, which support to
describe the nature of sensor data (Yan et al., 2008; Thomas et al., 2006).
The ultimate goal of this paper is critically analyzing sensor data toward behavioral
pattern mining using SDM techniques. Therefore, its contributions can be summarized
as: introducing a novel idea and techniques to demystifying sensor data behavioral
patterns as the context of the domain; optimizing DM applications as to explore
sequential sensor events. SDM adapts to analysis sequence event patterns based on
clustering and distance mining techniques for sizing clustered elements, then extend it
into normal and abnormal sequence pattern characterizations. Proposed and present a
generic distance-based clustering algorithms to define sensor data pattern behavior,
which is significant to segment the boundary of any trajectories as normal and
abnormal properties. The approach is also scalable to overcome challenges that
happened in the process of SE performance optimizations and fault diagnosis.
The rest of the paper is organized into five sections. Section II summarizes the
related and synonymous research works in the field of pattern mining and sensor
technologies. In Section III, we discussed the sensor data features, models and model
descriptions, pattern clustering and defining rules and algorithms. Section IV is the
empirical analysis of the methods and the detail discussed. The last, Section V is
the conclusion of this work, in which followed by the acknowledgments and reference
for cited articles in this work.
1152
IMDS
115,6
To continue reading
Request your trial