Scribd
Upload
47 views10 pages

Distributed Data Mining in Wireless Sensor Networks Using Naive Bayes Classification

This document proposes using a naive Bayes classifier for distributed data mining in wireless sensor networks to predict faulty sensor nodes. It describes limitations of existing centralized and clustering techniques, and introduces the concepts of distributed data mining and naive Bayes classification. The proposed technique involves preprocessing sensor network data, classifying networks, determining node statuses using finite state elements, predicting statuses with naive Bayes, and calculating error metrics. It is evaluated through simulation for its ability to consider heterogeneous data dependencies at low complexity.

Uploaded by

ashish88bhardwaj_314
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
47 views10 pages

Distributed Data Mining in Wireless Sensor Networks Using Naive Bayes Classification

This document proposes using a naive Bayes classifier for distributed data mining in wireless sensor networks to predict faulty sensor nodes. It describes limitations of existing centralized and clustering techniques, and introduces the concepts of distributed data mining and naive Bayes classification. The proposed technique involves preprocessing sensor network data, classifying networks, determining node statuses using finite state elements, predicting statuses with naive Bayes, and calculating error metrics. It is evaluated through simulation for its ability to consider heterogeneous data dependencies at low complexity.

Uploaded by

ashish88bhardwaj_314
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 10

Distributed Data mining In Wireless Sensor

Networks Using Naive Bayes Classification

AIM
WE INTEND TO DEVELOP A TOOL WHICH SHOWS
THE PREDICTION OF FAULTY SENSOR NODES IN
WIRELESS SENSOR NETWORKS USING SENSOR
NODE PARAMETERS.

Limitations of Existing Data Mining Techniques


for WSNs
Most of the techniques do not take into account the heterogeneous data

and assume that the sensor data is homogenous. They ignore the fact
that different attributes together can improve the mining accuracy.
The majority of techniques uses centralized approach in which all data
is transmitted to the sink node for identifying certain patterns. These
techniques cause much communication overhead and delay the
response time.
Clustering techniques suffer from the choice of an appropriate
parameter of cluster width, and computing the distance between data
instances in heterogeneous data is computationally expensive, whereas
classification-based techniques require some prior knowledge to
classify the incoming data stream.

Application Area

WSNs performance
WSNs are usually unattended and deployed in harsh
environment. Sensor nodes are resource constrained
especially in terms of power. We are using data mining
techniques that will help to identify the faulty or dead
nodes.

Concepts Used
Distributed data mining
Naive Bayes classifier for prediction
Finite state of element

Proposed Technique

1.

2.
3.
4.
5.
6.
7.

Loading the data set: The data set of a wireless sensor network includes the
following attributes( wsn attributes):
Sensor ID
Network size
no. of nodes
Radius of event
Boundary
Preprocessing of the data
Classifying of the networks in area
Determining the frequency on the basis of status(using finite state of element).
Identifying the status type: perfect, imperfect, minimal, failure.
Finding frequency on the basis of current status using nave Bayer's classification.
We calculate error parameters on the basis of input parameters :Absolute error,
Percentage error, Estimation error.

Process Flowchart
Given a data set of a WSN and load it into the system

Measurement of node parameters

Preprocessing of data and classification of networks

Individual frequency of each network using finite state


of element

Find out the frequency on the basis of current status


using Nave Bayes classification.

Calculate errors

Evaluation Method

Simulation:
Simulation framework used is .NET.
It is the most popular and effective approach to
design and test any proposed scheme in terms of cost
and time.
It also provides higher level of details as compared
with real implementation.
Simplicity for the dataset and its calculation.

Our Contribution

The technique combines offline machine learning

mechanisms with distributed data processing.


The technique has low computation complexity and
easy to implement.
The technique considers heterogeneous data and
dependencies among spatial and attribute
correlations which may exist between adjacent
nodes.

References

Distributed Data Mining in Sensor Networks,

Springer,2010
Review Article Data Mining Techniques for Wireless
Sensor Networks: A Survey Hindawi Publishing
Corporation International Journal of Distributed Sensor
Networks,2013
Study on Distributed Data Mining Model in Wireless
Sensor Networks
Distributed data mining in peer to peer networks, IEEE
explorer ,2006

You might also like