Application of Stream Data Time-Series Pattern Reliance Mining in Stock Market Analy

[ajitsuna]

The Application of Stream Data Time-Series Pattern Reliance Mining in Stock Market Analysis

With the rapid development of computer information technology, a new data model-the stream data emerged. The stream data is coming continuously, quickly, changing with time, and may be unpredictable and unlimited in the way. The emergence of stream data has brought new challenges from the nature ofthe data applied to static database technology and data mining technology. At present, the stream data has been widely used in telecommunications, financial securities, retail trade and other fields.

Stream data mining is a process of finding and extracting potential information and knowledge hidden in the stream data, which is useful but people do not know in advance. Due to the stream data is on the continuous arrival of large or even unlimited data, which can not be all stored, so many traditional data mining algorithms are not suitable for the stream data mining. The paper proposes a mining model and algorithm of stream data time-series pattern reliance in a dynamic stock market through the research of stream data time-series pattern mining algorithms and applications, and then does short-term forecasts on the stocks with pattern reliance to provide rational guidance for stock investors.

The stock price data coming from securities transactions are real-time, continuous, ever-changing with time, and the long-term accumulation of data on stock transactions that can be regarded as massive and unlimited. Therefore, the stock price data is stream data, at the same time, the stock price data shares sequential nature, which can be used to analyze stream data time-series pattern mining methods. In addition, there are many affecting factors of stock prices, making the price data show non-linear features, which bring new challenges to the traditional data mining algorithms.

The analysis of the stock market based on time series, including the analysis of historical market and forecasts of market trends. This theory is based on the cyclical principle of stocks, observing a series of data by chronological sequence intiming order. It studies the trend of changes implied in market data with the discovery process of time-series model, including the analysis of sequence trend, sequence mode mining, cycle mode mining and the query of the time sequence similarity, and so on. More reliable results, greater achievement value, relatively simple and easy to accomplish model are its characteristics.

The real time data stream can be obtained from yahoo finance and the analysis can be done using JAVA