[Singh, 2(10): October, 2013] ISSN: 2277-9655 Impact Factor: 1.852 http: // www.ijesrt.com (C) International Journal of Engineering Sciences & Research Technology [2767-2771] IJESRT INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH TECHNOLOGY Survey on Pattern Optimization for Novel Class in MCM for Stream Data Classification Vinay Singh *1 , Divakar Singh 2 *1,2 Department of Computer Science & Engineering, BUIT, Bhopal, M.P, India vinay.cse5@gmail.com Abstract The classification of stream data is somehow difficult. Existing data stream classification techniques assume that total number of classes in the stream is fixed. Therefore, instances belonging to a novel class are misclassified by the existing techniques. Because data streams have endless length, conventional multi pass learning algorithms are not appropriate as they would require infinite storage and training time. Concept-drift occurs in the stream when the underlying concept of the data changes over time. Thus, the classification model must be updated continuously so that it reflects the most recent concept. In this paper we are presenting some efficient research approaches suggested by numerous scholars. Keywords: Data stream, multi-class miner Introduction Stream data classification faced a problem of new class generation during process of pattern evaluation. The evaluation process of pattern raised new pattern of data for classification. The evolving new pattern mismatches the assigned class for stream data classification, now the generated pattern creates new class for classification process. For this process of handling multi- class miner process are used. But the multi-class miner failed a process of new pattern evaluation mechanism. For the generation of pattern and optimized pattern process for improving of multi-class miner used pattern optimization technique using genetic algorithm. The optimized pattern settled the new class and improved the efficiency multi-class miner. Pattern optimization plays an important role in stream data classification. The feature evaluation process of stream data induced a problem for classification such as infinite length. So it is not possible to store the data and use it for training. Infinite length, concept-evolution and concept- drift are major challenges in data streaming. The data stream is infinite amount of data; data continuous arrived and can only be read for one or a few times. So the faster method of data stream mining need to be updated. Data-stream mining is a technique which can find valuable information or knowledge from a great deal of primitive data. Unlike mining static databases, mining data streams poses many new challenges [1]. Data stream has different characteristics of data collection to the traditional database model. Such as the date of data stream continuous generation with time progresses and the data stream is dynamic and the arrival of the data stream cannot be controlled by the order. The data of data stream can be read and process based on the order of arrival. The order of data cannot be changed to improve the results of treatment. Therefore, the processing of the data stream requires first, each data element should be examined almost one time, because it is unrealistic to keep the entire stream in the main memory. Second, each data element in data streams should be processed as fast as possible. Third, the memory usage for mining data streams should be bounded even though new data elements are continuously generated. Finally, the results generated by the online algorithms should be instantly available when user requested [1]. Data stream compared with traditional data collection, the data stream is a real-time, continuous, orderly, time-varying, infinite tulle data stream has the following distinctive features such as orderly, Cannot Reproduce, High-Speed, Infinite, High Dimensional and Dynamic. Learning, like intelligence covers such a broad range of processes that it is difficult to define precisely. A dictionary definition includes phrases such as “to gain knowledge, or understanding of, or skill in, by study, instruction, or experience”, and “modification of a behavioural tendency by experience”. Certainly, many techniques in Machine Learning derive from the efforts of psychologists to make more precise their theories of animal and human learning through computational models. It seems likely also that the concepts and