Dynamic Time Warping Algorithm Research Articles

Dynamic Time Warping Algorithm in Modeling Systemic Risk in the European Insurance Sector.

We are looking for tools to identify, model, and measure systemic risk in the insurance sector. To this aim, we investigated the possibilities of using the Dynamic Time Warping (DTW) algorithm in two ways. The first way of using DTW is to assess the suitability of the Minimum Spanning Trees’ (MST) topological indicators, which were constructed based on the tail dependence coefficients determined by the copula-DCC-GARCH model in order to establish the links between insurance companies in the context of potential shock contagion. The second way consists of using the DTW algorithm to group institutions by the similarity of their contribution to systemic risk, as expressed by DeltaCoVaR, in the periods distinguished. For the crises and the normal states identified during the period 2005–2019 in Europe, we analyzed the similarity of the time series of the topological indicators of MST, constructed for 38 European insurance institutions. The results obtained confirm the effectiveness of MST topological indicators for systemic risk identification and the evaluation of indirect links between insurance institutions.

A CNN model embedded with local feature knowledge and its application to time-varying signal classification

A novel convolutional neural network is proposed for local prior feature embedding and imbalanced dataset modeling for multi-channel time-varying signal classification. This model consists of a single-channel signal feature parallel extraction unit, a multi-channel signal feature integration unit, a local feature embedding and feature similarity measurement unit, a full connection layer, and a Softmax classifier. An algorithm combining dynamic clustering and sliding window was used to select segments signals with typical local features in each pattern class, forming a typical local feature set. The one-dimensional CNNs were used to extract features from the single-channel signal in parallel, a comprehensive feature matrix of the multi-channel signal and the local feature matrix templates were produced. Using the method of external embedding, based on the sliding window and dynamic time warping (DTW) algorithm, the local feature similarities between the local feature template of each pattern class and the comprehensive feature sub-matrix of the input signal were measured, and the maximum values were selected to construct a local feature similarity vector in order. The information fusion was realized through a full connection layer. The proposed methodology can extract and represent both global and local signals features, strengthen the role of prior local feature in classification and improve the modeling properties of imbalanced datasets. A comprehensive learning algorithm is presented in this paper. The classification diagnosis of cardiovascular disease based on 12-lead ECG signals was used as a verification experiment. Results showed that the accuracy and generalization for the proposed technique were significantly improved.

Method for Measuring the Similarity of Multiple Metrological Sequences in the Key Phenological Phase of Rice-based on Dynamic Time

The automatic classification of historical data of myriad diverse meteorological sequences in the annual period can help to find the climate differences through key phenology of rice. In this paper, a hybrid gradients-shape dynamic time warping (HGSDTW) algorithm is proposed to measure the similarity of meteorological data during the diverse rice growth period at various locations. The weighting calculation of Euclidean distance uses the form factor in the rice jointing and heading stage. The distance matrix constructs first & second-level gradient single-factor transformation sequences during the period. The dynamic programming method obtains the similarity distances of single and multiple meteorological factors. The results show that the classification accuracy rate from HGSDTW of the heading & jointing stage is higher than that of other similar algorithms. Furthermore, it can observe that the clustering number increases the classification accuracy, and the HGSDTW algorithm maintains the accuracy of 14% for varieties of rice at diverse locations to multiple years of jointing. Besides, the automatic classification experiment of sequence period shows that the classification accuracy of this method is higher than that of another similarity measure. The classification accuracy rate of the heading stage sequence is 10%~14% higher than that of a similar previous standard measurement algorithm, and the jointing period is 1%~9% higher. In this case, the cluster number increasing the classification accuracy, and the HGSDTW maintain the overall accuracy of 14%. Thus, this method can be effectively combined with the classification algorithm to improve the efficiency of the automatic classification of multi-weather sequence data in key phenological periods of rice.

STA-VPR: Spatio-Temporal Alignment for Visual Place Recognition

Recently, the methods based on Convolutional Neural Networks (CNNs) have gained popularity in the field of visual place recognition (VPR). In particular, the features from the middle layers of CNNs are more robust to drastic appearance changes than handcrafted features and high-layer features. Unfortunately, the holistic mid-layer features lack robustness to large viewpoint changes. Here we split the holistic mid-layer features into local features, and propose an adaptive dynamic time warping (DTW) algorithm to align local features from the spatial domain while measuring the distance between two images. This realizes viewpoint-invariant and condition-invariant place recognition. Meanwhile, a local matching DTW (LM-DTW) algorithm is applied to perform image sequence matching based on temporal alignment, which achieves further improvements and ensures linear time complexity. We perform extensive experiments on five representative VPR datasets. The results show that the proposed method significantly improves the CNN-based methods. Moreover, our method outperforms several state-of-the-art methods while maintaining good run-time performance. This work provides a novel way to boost the performance of CNN methods without any re-training for VPR. The code is available at https://github.com/Lu-Feng/STA-VPR.

P-P and Dynamic Time Warped P-SV Wave AVA Joint-Inversion With ℓ1–2 Regularization

S-wave velocity and mass density, which are two essential parameters in distinguishing lithology and hydrocarbon detection, are very difficult to retrieve even when long-offset gather data are used. P-P and P-SV waves joint inversion has been verified as an effective tool to accurately extract such fluid related parameters. However, registering the travel-time of P-P and P-SV waves on the identical coordinate is a significant but knotty problem for the joint inversion. Therefore, an improved strategy of prestack seismic joint inversion is proposed for accurately transforming the recorded data to elastic-parameter-based interpretive information. First, to overcome the weaknesses of artificial strenching and the local cross correlation-based method, a nonstrenching and globally optimal registration algorithm, i.e., dynamic time warping (DTW), is exploited to match the P-P and P-SV waves. Then, a new method has been developed for the joint inversion method by combining the logarithmic absolute-criterion-based misfit function with <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">l</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1-2</sub> norm-based penalty, which can significantly improve the vertical resolution and stability of inversion results. The numerical examples demonstrate that the DTW algorithm can obtain better registered results than the conventional method. Moreover, the proposed joint inversion method performs better than the conventional prestack inversion in both resolution and accuracy, especially for the S-wave velocity.

Understanding Stock Price Movement on the COVID-19 Period Using Dynamic Time Warping Algorithm

Unlike the impact of MERS or SARS, the COVID-19 has had a huge impact on both social and economic environments. This study investigated the COVID-19 economic impact using the S&P 500 index by applying the DTW (Dynamic Time Warping) algorithm, an artificial intelligence technique. The DTW algorithm is designed for the identification of events the capital market impact resembles in the past. It is shown that the impact of the capital market on COVID-19 showed a completely different pattern from the patterns seen in MERS and SARS. Also, the results showed patterns similar to those of the 2015 China’s Stock Market Crash, the 2019 U.S.-China Trade Dispute, the 2008 Global Financial Crisis, which are specifically referred to as financial crises. However, it was unexpectedly similar to the shock of the U.S.-North Korea Conflict and the 2001 9·11 Terror. They were treated as political and diplomatic crises. These events are closely related to APEC countries, meaning that APEC countries’ economic and political activities are essential in the global economic crisis. The results imply that the policies implemented by APEC countries to overcome the economic crises are important to resolve the global economic crisis from COVID-19.

A Human–Computer Interaction framework for emotion recognition through time-series thermal video sequences

Infrared-Thermal Imaging is a non-contact mechanism for psychophysiological research and application in Human–Computer Interaction (HCI). Real-time detection of the face and tracking the Regions of Interest (ROI) in the thermal video during HCI is challenging due to head motion artifacts. This paper proposes a three-stage HCI framework for computing the multivariate time-series thermal video sequences to recognize human emotion and provides distraction suggestions. The first stage comprises of face, eye, and nose detection using a Faster R-CNN (region-based convolutional neural network) architecture and used Multiple Instance Learning (MIL) algorithm for tracking the face ROIs across the thermal video. The mean intensity of ROIs is calculated which forms a multivariate time series (MTS) data. In the second stage, the smoothed MTS data are passed to the Dynamic Time Warping (DTW) algorithm to classify emotional states elicited by video stimulus. During HCI, our proposed framework provides relevant suggestions from a psychological and physical distraction perspective in the third stage. Our proposed approach signifies better accuracy in comparison with other classification methods and thermal data-sets.

Predicting total electron content in ionosphere using vector autoregression model during geomagnetic storm

Abstract The ionospheric total electron content (TEC) severely impacts the positional accuracy of a single frequency Global Positioning System (GPS) receiver at the equatorial latitudes. The ionosphere causes a frequency-dependent group delay in the GPS-ranging signals, which reduces the receiver’s accuracy. Further, the variations in TEC due to various space weather phenomena make the ionosphere’s behaviour nonhomogeneous and complex. Hence, developing an accurate forecast model that can track the dynamic behaviour of the ionosphere remains a challenge. However, advances in emerging data-driven algorithms have been found helpful in tracking non-stationary behavior in TEC. These models help forecast the delays in advance. The multivariate Vector Autoregression model (VAR) predicts the Ionospheric TEC in the proposed model. The prediction model uses input data compiled in real-time from the lag values of incoming TEC data and features extracted from TEC. The TEC is predicted in real-time and tested for different prediction intervals. The metrics – Mean Percentage Error (MAPE), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) are used for testing and validating the accuracy of the model statistically. Testing the predicted output accuracy is also done with the dynamic time warping (DTW) algorithm by comparing it with the actual value obtained from the dual-frequency receiver. The model is tested for storm days of the year 2015 for Bangalore and Hyderabad stations and found to be reliable and accurate. A prediction interval of twenty-minute shows the highest accuracy with an error within 10 TECU for all the storm days.

Asymmetric Gait Analysis Using a DTW Algorithm with Combined Gyroscope and Pressure Sensor.

Walking is one of the most basic human activities. Various diseases may be caused by abnormal walking, and abnormal walking is mostly caused by disease. There are various characteristics of abnormal walking, but in general, it can be judged as asymmetric walking. Generally, spatiotemporal parameters can be used to determine asymmetric walking. The spatiotemporal parameter has the disadvantage that it does not consider the influence of the diversity of patterns and the walking speed. Therefore, in this paper, we propose a method to analyze asymmetric walking using Dynamic Time Warping (DTW) distance, a time series analysis method. The DTW distance was obtained by combining gyroscope data and pressure data. The experiment was carried out by performing symmetrical walking and asymmetrical walking, and asymmetric walking was performed as a simulation of hemiplegic walking by fixing one ankle using an auxiliary device. The proposed method was compared with the existing asymmetric gait analysis method. As a result of the experiment, a p-value lower than 0.05 was obtained, which proved that there was a statistically significant difference.

Similarities among glacials and interglacials in the LR04 benthic oxygen isotope stack over the last 1.014 million years revealed by cluster analysis and a DTW algorithm

Quantitative algorithmic examination of the LR04 oxygen-isotope record using cluster analysis and dynamic time warping (DTW) reveals significant new understandings of changes in glaciation and sea level during the Quaternary. For example, four relatively recent and major glacial stages, MIS 2 (or 2–4), 6, 12 and 16, emerge as a distinct group and suggest an Early Brunhes Shift to more extreme glacial stages about 660 thousand years ago. Secondly, two nominally interglacial stages, MIS 3 and 23, group with glacial stages. Of these two, recognition of MIS 3 as part of the last glacial phase is not a novel finding. However, recognition of MIS 23 as a glacial stage is significant because viewing MIS 22, 23, and 24 as one glacial phase formalizes a ~100,000-year glacial cycle from 960 to 859 ka. That result pushes the Mid-Pleistocene Transition, when the periodicity of climate cycles changed from 41 thousand to 100 thousand years, back to no later than 970 ka, counter to previous views that the Mid-Pleistocene Transition could have been as recent as 700 thousand years ago.

A People-Counting and Speed-Estimation System Using Wi-Fi Signals.

Counting the number of people and estimating their walking speeds are essential in crowd control and flow. In this work, we propose a system that uses prevalent Wi-Fi signals to identify the number of people entering and leaving a room through a door. It selects the best subcarrier of Wi-Fi signals and applies the Hampel filter to remove outlier information first. Then, it employs a double threshold method to determine the start and end times of entering or leaving. Afterward, it compares the detected signals with the precollected database using the dynamic time-warping algorithm and determines the number of people. It uses a variance threshold method to identify the states of entering or leaving. It also employs a nonlinear fitting approach to calculate the walking speeds. The experiments show that, in a large empty laboratory, the accuracy rates in determining the number of people are 100% for one person, 81% for two persons, and 95% for three persons. In a small office, the accuracy rates for detecting the number of people are 98% for one or two persons, 82% for three persons, 93% for four, and 75% for five persons. For the walking speed estimation, the accuracy rate for a speed error of less than 0.2410 m/s is 75% for a single person.

Determination of Key Phenological Phases of Winter Wheat Based on the Time-Weighted Dynamic Time Warping Algorithm and MODIS Time-Series Data

Accurate determination of phenological information of crops is essential for field management and decision-making. Remote sensing time-series data are widely used for extracting phenological phases. Existing methods mainly extract phenological phases directly from individual remote sensing time-series, which are easily affected by clouds, noise, and mixed pixels. This paper proposes a novel method of phenological phase extraction based on the time-weighted dynamic time warping (TWDTW) algorithm using MODIS Normalized Difference Vegetation Index (NDVI) 5-day time-series data with a spatial resolution of 500 m. Firstly, based on the phenological differences between winter wheat and other land cover types, winter wheat distribution is extracted using the TWDTW classification method, and the results show that the overall classification accuracy and Kappa coefficient reach 94.74% and 0.90, respectively. Then, we extract the pure winter-wheat pixels using a method based on the coefficient of variation, and use these pixels to generate the average phenological curve. Next, the difference between each winter-wheat phenological curve and the average winter-wheat phenological curve is quantitatively calculated using the TWDTW algorithm. Finally, the key phenological phases of winter wheat in the study area, namely, the green-up date (GUD), heading date (HD), and maturity date (MD), are determined. The results show that the phenological phase extraction using the TWDTW algorithm has high accuracy. By verification using phenological station data from the Meteorological Data Sharing Service System of China, the root mean square errors (RMSEs) of the GUD, HD, and MD are found to be 9.76, 5.72, and 6.98 days, respectively. Additionally, the method proposed in this article is shown to have a better extraction performance compared with several other methods. Furthermore, it is shown that, in Hebei Province, the GUD, HD, and MD are mainly affected by latitude and accumulated temperature. As the latitude increases from south to north, the GUD, HD, and MD are delayed, and for each 1° increment in latitude, the GUD, HD, and MD are delayed by 4.84, 5.79, and 6.61 days, respectively. The higher the accumulated temperature, the earlier the phenological phases occur. However, latitude and accumulated temperature have little effect on the length of the phenological phases. Additionally, the lengths of time between GUD and HD, HD and MD, and GUD and MD are stable at 46, 41, and 87 days, respectively. Overall, the proposed TWDTW method can accurately determine the key phenological phases of winter wheat at a regional scale using remote sensing time-series data.

Non stretching NMO (Normal-Moveout) correction with DTW (Dynamic Time Warping) algorithm, study case: shallow target seismic

Normal Move Out (NMO) Correction using Common Mid-Point (CMP) data based on shallow target seismic. It has several problems for NMO correction. This constraint is caused by changes in extreme velocity gradients, resulting in less optimum value shallow target Common Mid-Point (CMP). The increased velocity causes a stretch in the Normal moveout (NMO) process which can reduce the S/N ratio. To minimize the stretch effect on NMO processing, we use a dynamic time warping algorithm analysis seismic signal with dynamic time warping algorithm based calculation time shift at time series sequence. The results of the NMO-DTW show that there is no change phase and frequency in the trace seismic with the far offset. Because algorithm calculation is based on the initial trace on CMP.

A Sensor-Based Data Driven Framework to Investigate PM2.5 in the Greater Detroit Area

PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> are inhalable particles with aerodynamic diameters of 2.5 micrometers or smaller. PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> concentrations require close monitoring as they impose negative effects on both human health and air quality. Monitoring PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> concentrations in the metropolitan Detroit Area is increasingly important as its residents are being disproportionately exposed to harmful air pollution due to health inequities through economic divestment, limited educational and employment opportunities. The relations between PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> and meteorological factors can be critical in understanding how particulate matter affects humans and the environment. This study utilizes PurpleAir sensors to measure PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> along with some methodological factors such as humidity and applies temporal analysis of the impact of meteorological factors on PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> concentrations and spatiotemporal analysis of PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> distributions at different locations over the Greater Detroit Area via Long Short Term Memory (LSTM) neural networks and Dynamic Time Warping (DTW) algorithms, respectively. Our findings show that although LSTMs with exogenous variables (i.e., the current values of PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> concentration, meteorological features, and weather conditions) can accurately (i.e., average RMSE of 3.2 μg/m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> ) predict levels of PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> , but there is no significant relation between the mentioned meteorological factors and PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> concentrations over the Greater Detroit Area. Furthermore, DTW analysis portraits the similarity of PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> behavioral patterns over the Greater Detroit Area.

Assessment of flourishing levels of individuals by using resting-state fNIRS with different functional connectivity measures

Flourishing is an important criterion for assessing well-being, however, controversy remains, especially while assessing it with self-report measures. Therefore, to understand the underlying neural mechanisms of well-being, researchers often use neuroimaging techniques. However, previous neuroimaging studies using conventional statistical approaches provided answers in an average sense rather than individual answers. In this study, we used machine learning algorithms to classify highly flourishing from normally flourishing individuals using a publicly available resting-state functional near-infrared spectroscopy (rs-fNIRS) dataset collected from 43 participants to obtain an answer for individual level. We utilized both the Pearson’s correlation (CC) and the Dynamic Time Warping (DTW) algorithm to estimate functional connectivity matrices from the rs-fNIRS data on the temporo-parieto-occipital region and used them as input for machine learning algorithms. Our results showed that we were able to classify flourishing individuals with 90 % accuracy with AUC 0.90 and 0.93 using Nearest Neighbor and Radial Basis Kernel Support Vector Machine using oxyhemoglobin concentration change with Pearson’s correlation (CC – ΔHbO) and deoxy hemoglobin concentration change with dynamic time warping (DTW – ΔHb). This finding suggests that temporo-parieto-occipital region-based resting-state functional connectivity might be a potential biomarker to identify the levels of flourishing and using both connectivity measures might allow us to find different potential biomarkers.

Denial of service detection using dynamic time warping

SummaryWith the rapid growth of security threats in computer networks, the need for developing efficient security‐warning systems is substantially increasing. Distributed denial‐of‐service (DDoS) and DoS attacks are still among the most effective and dreadful attacks that require robust detection. In this work, we propose a new method to detect TCP DoS/DDoS attacks. Since analyzing network traffic is a promising approach, our proposed method utilizes network traffic by decomposing the TCP traffic into control and data planes and exploiting the dynamic time warping (DTW) algorithm for aligning these two planes with respect to the minimum Euclidean distance. By demonstrating that the distance between the control and data planes is considerably small for benign traffic, we exploit this characteristic for detecting attacks as outliers. An adaptive thresholding scheme is implemented by adjusting the value of the threshold in accordance with the local statistics of the median absolute deviation (MAD) of the distances between the two planes. We demonstrate the efficacy of the proposed method for detecting DoS/DDoS attacks by analyzing traffic data obtained from publicly available datasets.

Robust Visual Place Recognition in Changing Environments Using Improved DTW

Recently, the methods based on Convolutional Neural Networks (CNNs) have achieved state-of-the-art performance in visual place recognition. CNN is a class of multilayer perceptrons, but unlike common multilayer perceptrons that it is not usually fully connected networks. It can acquire more general image features and make the image processing computationally manageable through filtering the connections by proximity. In this paper, we utilize the deep features generated by CNNs and the dynamic time warping (DTW) algorithm for image sequence place recognition. We propose a novel image similarity measurement, which is derived from cosine distance and can better distinguish match and mismatch. Meanwhile, we improve the DTW algorithm to design a local matching method that can reduce time complexity from O(n3) to O(n). To test the proposed method, four datasets (Nordland, Gardens Point, St. Lucia, and UoA datasets) are used as benchmarks; using two traverses in each dataset with one for reference and the other for testing. The results show high precision-recall characteristics of our method in the cases of severe appearance changes. Besides, our method achieves substantial improvements over the methods using the deep feature representations of a single image for recognition, which reflects that the spatiotemporal information contained in the image sequence is significant for the task of visual place recognition. Moreover, the proposed method also shows to outperform the classical sequence-based method SeqSLAM.

Multivariate time-series clustering based on component relationship networks

Clustering is a powerful technique for providing class labels of data objects for learning guidance. Because traditional clustering methods rarely consider the component correlations of a multivariate time series(MTS), an MTS clustering method based on a component relationship network (CRN) is proposed in the present study. An MTS dataset is mapped to a multi-relationship network (MRN), which consists of a set of CRNs. Every CRN reflects the relationship of the MTS data under each component. The proposed method applies two steps. First, the distance function and K-nearest neighbors are combined to transform an MTS dataset into an MRN. Second, a non-negative matrix factorization is designed to identify time-series clusters. Because asynchronous time-series data exist in the form of an approximate case, we recommend an improved penalty-coefficient based dynamic time-warping algorithm to measure the similarity between two-time sequences. The similarity can reflect the correlation between the asynchronous MTS data. Through an experiment, we compared the proposed method to other clustering methods and discussed of the effects of the parameters in detail. The numerical results of the experiment demonstrate that the proposed method can improve the accuracy and quality of MTS data clustering.

High-Throughput Dynamic Time Warping Accelerator for Time-Series Classification With Pipelined Mixed-Signal Time-Domain Computing

Time-series classification (TSC) is a challenging problem in machine learning and significant efforts have been made to improve its speed and computation efficiency. Among various approaches, dynamic time warping (DTW) algorithm is one of the most prevalent methods for TSC due to its succinctness and generality. To improve the throughput of the operation, this work presents a mixed-signal DTW accelerator utilizing mixed-signal time-domain (TD) computing where signals are encoded and processed using time pulses. A pipelined operation is enabled by a specially designed time flip-flop (TFF) circuit leading to dramatic improvements in performance and scalability of the operation. A 65-nm CMOS test chip was implemented and measured. The results show more than 9× improvements in throughput compared with prior work on TSC. As most existing TD designs suffer from the lack of TD storage elements, this work utilizes sequential circuit elements in TD computing extending the capability of time-based circuits.

A Study on Stock Graph Recognition Based on Wavelet Denoising and DTW Algorithm

The “classical pattern” of stock price formation has long been widely used in the determination of future price trends of stocks, and the identification and analysis of classical price patterns have an important guiding role in investors’ decision-making and trading. The wavelet transform is a useful tool to remove some of the noise of time series because it has the characteristic of multiresolution. In this study, we propose a method for stock price pattern recognition based on the wavelet transform and dynamic time warp (DTW). A pattern recognition method with similar quantified results is developed to obtain accurate pattern recognition results. That is, using the wavelet transform to smooth the original price graph, and then using the DTW algorithm improved in this study to find the graph with the smallest distance from the target graph under the sliding window method, the identification and analysis of the target graph can be realized. In order to improve the recognition rate of the target graph, we preprocessed the raw price sequence using the moving average convergence and divergence (MACD) algorithm based on the control experiments set up in this study. The pattern recognition method used in this study will identify the price patterns of a certain time window as a whole, thus avoiding the problem of how to objectively select the important points that constitute a price pattern and the mathematical definition of different price patterns in the previous traditional methods.