The device and bandwidth of network. These

The objective of mobile data mining is to provide
efficient techniques for the analysis and monitoring of critical data from
mobile devices. The client programs are present on mobile devices
that can invoke the data mining tasks of remote execution and show the mining
results is an additional value for organizations and users who need to perform data
analysis stored where users are working, which is very far away from the site that
allows them to produce study regardless of their present location.
The
mobile data mining provides efficient techniques for the monitoring of critical
data from mobile devices.

Distributed
data mining environment is a critical issue which mobile data mining has to
face, also some technological constraints such as low-bandwidth of networks, much
slower processors, lesser storage space, small screens to visualize the results
as well as limited battery power.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

 

MobiMine
is an example of mobile data mining environment that is created for intelligent

supervising
from mobile devices of stock market. It is based on a client-

server
architecture. The clients, which are running such as PDAs, monitor

a stream
of all the data which are coming through a server. The stock

market
data is collected from number of Web sources in a database and evaluate it on a
regular basis using several mining techniques.

The database
is queried by clients for the latest information as for quotes and other essential

information.
For communication which includes clients and the database a proxy is used. When
a user have some query regarding the database, to the proxy the query is send
that connects to the database, take out the results and pass them to the client.

MobiMine
uses a Fourier-dependent approach for the representation of the decision trees,
for efficient communication of data mining models over wireless links that has
limited

Bandwidth,
it saves memory on mobile device and bandwidth of network.

 

These are the major components of any data
mining system: data sources, data warehouse server, engine for data mining,
graphical user interface (GUI), pattern evaluation module and knowledge base.
System architecture of mobile data mining is based on three types of basic
components:

1) Data Sources

Data warehouse, World Wide Web, Database, some text
files and other documents which are the actual sources of data present. We
require large volumes of historic data to be successful for data mining.
Organizations mostly store data in data warehouses. It may contain one or more
databases, files, spreadsheets or other kinds of information database. Mostly,
data may present even in plain text files and spreadsheets. The another big
source of data is World Wide Web or the Internet.

Different Processes

The data should be cleaned, integrated and picked
up before passing it to the  data
warehouse server. The data is from number of sources and in different formats,
so it cannot be used directly the data may not be complete and reliable for the
data mining process. So, first data needs to be cleaned and integrated for data
mining process. Similarly, more than required data will be taken from various data
sources and only the interested data needs to be selected and passed to the
server. These processes are not as much simple. As part of cleaning,
integration and selection as a number of techniques may be performed on the
data.

2)
Mobile clients

The mobile client is composed of three components:
the MIDlet , the DMS, and the Record Management System (RMS)

Ø 
The MIDlet is a J2ME(java 2 micro edition)
application.

Ø 
MIDlet uses mobile information device
profile(MIDP).

Ø 
The DMS(data mining service) is a Web Service stub
that allows the MIDlet to invoke the operations of a remote
data mining server.

 Working

1.      The
MIDlet invokes the submitTask operation of the remote DMS. Whenever the task
has been submitted, the DMS returns a unique id for that task.

2.      The
MIDlet monitors its status by querying the DMS.It periodically invokes the
getStatus operation, that receives the id of the task and returns its current
status.

3.      Whenever
the getStatus operation return done, the MIDlet invokes the getResult operation
to receive the result of the data mining analysis.

 

3) Mining Server

Server nodes are essential for
storing the data generated by some data providers and for executing the mining
tasks submitted by mobile clients. Each mining server uncovers its
functionalities by the help of two web services:

Data Collection Service (DCS)

The
DCS is invoked by data providers to store data on
the server.For
uploading new data.For
deleting an existing data.For
updating data set.

  Data
Mining Service (DMS)

Invoked
by mobile clients to perform data mining tasks. Obtaining the list of the
available data sets and algorithms.Submitting
a data mining task.Getting
the current status of a computation, along with the result of a given
task. The
DCS is invoked by data providers to store data on the server. Data
uploaded through the DCS is stored as plain data sets in the local file
system.

The DMS invokes the mobile
clients to perform data mining tasks. Its interface defines a set of operations
(DMS ops). The data analysis is performed by the DMS using a subset of all
the algorithms.

x

Hi!
I'm Ethel!

Would you like to get a custom essay? How about receiving a customized one?

Check it out