Research On Pattern Mining Of Financial Time Series Volume Indicators

Data analysis has now become a hot topic in the world,the data generated on a daily basis has been calculated and analyzed by researchers to mine frequent sequence segments.The main content of pattern mining is to explore the significance of these sequences and the impact on the follow-up data.In the stock model mining under the financial background,there is a single one-sided problem of analysis index.Therefore,based on the traditional time series data mining,this paper takes the stock market in the financial environment as the background,integrates the trading volume index into the research of stock price Motif.This paper studies the three aspects of multiple time series model extraction,income model establishment and optimization,decision tree and integrated learning algorithm prediction.The main contents and results of this paper are as follows:(1)By constructing the multivariate time series of trading volume and stock price,the sliding window linear fitting algorithm is used to preprocess the multivariate time series to highlight the development trend of the data and improve the robustness to noise.By using the adaptive step segmentation algorithm,the unique volume price combination of the trading volume and the stock price development trend in the same period of time is obtained.Through the extraction of frequent patterns,the experiment shows that the trend of volume and price is relatively high,which can reflect the idea of volume before price,and help investors better grasp the development of stock in the perspective of trading volume.(2)By combining the theory of the relationship between volume and price,this paper designs the volume price Motif model based on the income as the measurement standard.By defining the unique pattern of two frequent volume price trends as buy and sell point signals to carry out the income experiment.For some stocks with unsatisfactory experimental return results,an improvement of the bottom-up algorithm has been developed.The experimental data show that the combination of volume price Motif and transaction can achieve desirable positive returns with high probability,which can enrich investors' trading theory.(3)Through cart decision tree algorithm to explore the application value of the singularity of trading volume,the input data of decision tree is set as opening price,closing price,highest price,lowest price,trading volume,turnover rate,ten day average price,ten day maximum trading volume and the corresponding closing price of the point,and the prediction error is reduced through Ada Boost algorithm.Combined with the prediction results,this paper puts forward the definition Motif "wave by wave ",designs the singularity mode mining experiment of trading volume,and carries out the income verification.The results show that this pattern can be combined with the buy point to locate the intervention time of the stock in time to grasp the late market.