The device and bandwidth of network. these

The objective of mobile data mining is to provideefficient techniques for the analysis and monitoring of critical data frommobile devices. The client programs are present on mobile devicesthat can invoke the data mining tasks of remote execution and show the miningresults is an additional value for organizations and users who need to perform dataanalysis stored where users are working, which is very far away from the site thatallows them to produce study regardless of their present location. Themobile data mining provides efficient techniques for the monitoring of criticaldata from mobile devices. Distributeddata mining environment is a critical issue which mobile data mining has toface, also some technological constraints such as low-bandwidth of networks, muchslower processors, lesser storage space, small screens to visualize the resultsas well as limited battery power.

MobiMineis an example of mobile data mining environment that is created for intelligentsupervisingfrom mobile devices of stock market. It is based on a client-serverarchitecture. The clients, which are running such as PDAs, monitora streamof all the data which are coming through a server. The stockmarketdata is collected from number of Web sources in a database and evaluate it on aregular basis using several mining techniques. The databaseis queried by clients for the latest information as for quotes and other essentialinformation. For communication which includes clients and the database a proxy is used.

Whena user have some query regarding the database, to the proxy the query is sendthat connects to the database, take out the results and pass them to the client. MobiMineuses a Fourier-dependent approach for the representation of the decision trees, for efficient communication of data mining models over wireless links that haslimitedBandwidth, it saves memory on mobile device and bandwidth of network. These are the major components of any datamining 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 basiccomponents: 1) Data SourcesData warehouse, World Wide Web, Database, some textfiles and other documents which are the actual sources of data present. Werequire large volumes of historic data to be successful for data mining.

Organizations mostly store data in data warehouses. It may contain one or moredatabases, files, spreadsheets or other kinds of information database. Mostly, data may present even in plain text files and spreadsheets. The another bigsource of data is World Wide Web or the Internet. Different ProcessesThe data should be cleaned, integrated and pickedup before passing it to the  datawarehouse 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 thedata mining process. So, first data needs to be cleaned and integrated for datamining process.

Similarly, more than required data will be taken from various datasources and only the interested data needs to be selected and passed to theserver. These processes are not as much simple. As part of cleaning, integration and selection as a number of techniques may be performed on thedata. 2)Mobile clientsThe 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 deviceprofile(MIDP). Ø The DMS(data mining service) is a Web Service stubthat allows the MIDlet to invoke the operations of a remotedata mining server. Working1.      TheMIDlet invokes the submitTask operation of the remote DMS. Whenever the taskhas been submitted, the DMS returns a unique id for that task. 2.

TheMIDlet monitors its status by querying the DMS. It periodically invokes thegetStatus operation, that receives the id of the task and returns its currentstatus. 3.      Wheneverthe getStatus operation return done, the MIDlet invokes the getResult operationto receive the result of the data mining analysis. 3) Mining Server Server nodes are essential forstoring the data generated by some data providers and for executing the miningtasks submitted by mobile clients. Each mining server uncovers itsfunctionalities 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.  DataMining 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 mobileclients 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 allthe algorithms.