The Application Of Data Mining In The Analysis Of K-Line

There are many technical analysis methods for stock price forecasting at present,however,due to the complexity of stock market volatility,many traditional technical analysis methods can not get good results in short-term stock price forecast.In recent years,the use of data mining technology to analyze the stock market has become a popular research direction.In the stock's K-line analysis,there is a K-line which has a long upper wick,when it happens,often means an escape of main investor or liquidation of stocks,the data consists of technical specifications of the previous day,the day(having a long shadow)and the following day.These technical specifications are mainly: volume,DDE large net,change,the 3-day technical indicator data is connected to form a data sample point,and the data set is mined using clustering and association analysis.These data are pelleted and normalized prior to cluster analysis,the best number of clusters is obtained by using the minimum descriptive length(MDL)criterion in the Gaussian mixture model(GMM),using k-means and GMM clustering algorithm to do clustering analysis of data sets under the same cluster number,getting the best clustering results by comparing the error squares of the results of these two clustering algorithms,then the fp-growth algorithm is used to do correlation analysis for each cluster obtained by cluster analysis.The experiment obtains some meaningful operation modes of long upper wick,in this paper,the operation mode of long upper wick refers to the relevant technical indicators of the previous day,the day(having a long shadow)and the following day meet a certain constraint,the stocks which meet the operation mode have a larger probability of rising on the second day after appearing of long upper wick.For stocks which meet the operation mode of long upper wick in this paper,experimental results show that these stocks have not only a greater probability of rising but also a higher expected gains on the second day after appearing of long upper wick,in addition,these operation modes derived from this paper are extracted from a large amount of data and they have a high degree of confidence.