Study On Stock Data Analysis Based On Image Processing

At present, people’s financial awareness is increasing, causing more and moreinvestors to invest in stocks. Therefore, it is necessary to analyze and predict the stockmarket. Researchers have also been committed to predict the trend of a stock, a stockindex or stock of different sections with various methods. Although the volatility of thestock may lead to uncertain gains for investors, there are many factors to change theprice of stocks. Therefore, it is very difficult to grasp the exact law of ups and downsfully for investors. However, if the ups and downs of the stock are estimated basically,to some extent, it can also give investors some recommendations. After investorsunderstand the ups and downs of the whole plate and whether there is stock which isdifferent from other stocks in the plate, the study of all stock data of a plate will behelpful to provide some reference information for analyzing the later development of aplate of stock. Therefore, the estimation of the trend of rise or fall of a plate stock isvaluable.Investors hope that they can master the vagaries of the stock market’s law of upsand downs. There are a lot of technical analysis and stock forecasting methods, but to acertain extent, technical analysis is subjective. Different researchers will have differentconclusions. Stock forecasting methods mainly contain mathematical models andmodel-free. Traditional research on the analysis and forecast of the stock data mainlyfocused on a single stock or stocks not on the same plate, especially empirical analyzingthe data using the specified model. When creating a mathematical model, theappropriate variables is selected correctly, in order to ensure reliability and accuracy ofthe empirical analysis and forecasting, but we need test various variables significantlyand appropriately, therefore, computation is relatively large and it is not intuitive.The content of this study is to analyze stock data combined image processing.Firstly, normalize stock data of a plate collected and then make them to gray image. Thevertical direction of the image denotes the different stock, while the horizontal directiondenotes the closing price stock data of different dates. The correlation strength betweenthe various stocks will affect the consistency and smoothness of the image in thevertical direction. Therefore we need rearrange stocks of the plate. Investors alwaysfocused on the general trend of the stocks, ignoring the small fluctuations in the stock,which regarded as noise. When analyzing gray image formed by stock data of the plate, noise will affect the clarity of the image. This essay makes use of the mean smoothing,median smoothing and adaptive Wiener filtering method to de-noise the image from thevertical direction, combing the characteristics of the gray image formed by stock data.Comparing the experimental results analyzing the change of stock plate combined withthe de-noised image. Then it can conclude that whether the plate has the obvious plateeffect. Finally, extract edge of the de-noised image in the horizontal and verticaldirection. The edge of the vertical direction reflect the maximum or minimum value ofthe stock plate in the magnitude of the ups and downs of a moment; however the edgeof the horizontal direction mainly reflects the singularity rally of a stock in the plate.When extracting Edge, mainly use wavelet transform, find the extreme values in thedata and then find the edge in the corresponding direction. Estimate the cycle of thechange of the stocks and give some suggestions and ideas, combined with the locationof the edge. Compared with the previous method, this article analyzes the stock datafrom a two-dimensional perspective and it is more intuitive to reflect the rise or fall ofthe stock plate.