53 results about "Time series data analysis" patented technology

The present invention relates to a method and an apparatus for determining which one or more time series parameters of a plurality of time series parameters relating to operation of a system are correlated with a first operation state of the system. According to the invention, the method comprises providing time series data including data relating to a time series of each of the plurality of time series parameters; determining at least two first time periods, wherein the system is in the first operation state during the at least two first time periods; determining at least one second time period, wherein the system is in a second operation state during the at least one second time period; determining, for each respective time series parameter of the plurality of time series parameters, a first characteristic parameter relating to a first characteristic of the time series of the respective time series parameter for each of the at least two first time periods and the at least one second time period; and determining which one or more time series parameters of the plurality of time series parameters relating to the operation of the system are correlated with the first operation state of the system by determining, for each respective time series parameter of the plurality of time series parameters, whether or not the respective time series parameter is correlated with the first operation state of the system based on the first characteristic parameters of the respective time series parameter determined for each of the at least two first time periods and the at least one second time period.

The invention discloses a method for predicting reserve fund of bank outlets based on long-short-term memory cyclic neural network, which includes three stages of data preprocessing, model training and prediction. The data preprocessing stage takes the day as the unit, counts the total daily deposits and daily withdrawals of cash transactions at bank outlets, and the date attribute of the day, and constructs a feature vector; calculates the daily net amount based on the daily cash transaction records. In the model training phase, the LSTM model is trained based on historical feature vectors and daily net data. In the forecasting stage, the eigenvectors of several days before the forecast date of the bank branch are counted, and the range of the daily net amount predicted by the LSTM model is input, and the random value in the range is taken as the reserve fund requirement of the day. The invention makes full use of the historical data and the advantages of the long-short-term memory cyclic neural network in the time-series data analysis, and effectively improves the accuracy rate of reserve fund prediction at bank outlets.

A text data storage unit stores a plurality of text data having attribute data and time data. A dictionary storage unit stores a plurality of events each associated with text data. An analysis condition indication unit indicates an analysis target as attribute data and an analysis condition as an event sequence. A time series data generation unit assigns an event to each of the plurality of text data by referring to the dictionary storage unit, extracts a group of text data each having the same attribute data as the analysis target from the plurality of text data, and generates time series data each representing the event assigned to the text data of the group in order of the time data of the text data. A time series data analysis unit analyzes the time series data each having the same event sequence as the analysis condition.

The invention relates to an electric pump well fault real-time diagnostic system and method based on time series data analysis, and the electric pump well fault real-time diagnostic system and method are applied to oilfield downhole electric pump fault diagnosis and treatment. The system comprises a downhole submersible device, a cable, a downhole parameter measuring device, a control cabinet, a well mouth casing pressure, oil pressure and temperature measuring device, a ground data collection and analysis system, a fault analysis processor and a control cabinet current, voltage and frequency data collection system. The method comprises the steps that fault type characteristic parameters and a corresponding treatment measure database are set up; production characteristic parameters are obtained in real time; real-time flow is calculated, average values of the production characteristic parameters and the flow in four time periods are calculated correspondingly; variation amplitude is calculated (please see the formula in the specification ); an integrated assessment value is calculated (please see the formula in the specification); the electric pump well fault type is determined; and a corresponding electric pump well fault treatment measure is recommended. According to the electric pump well fault real-time diagnostic system and method based on time series data analysis, electric pump well fault real-time diagnosis is achieved, and fault misjudgement is effectively reduced.

An unnecessary load is prevented from being applied to the system even in cases where the time series data required for the time series data analysis processing has not been collected. A time series data processing apparatus assigns an arrival time, which is a time that the time series data arrived, to the time series data sent from the data source, determines whether the requested time series data has arrived, and predicts the arrival time of the time series data, which was determined by the data arrival determination unit as not yet arrived, based on the arrival time assigned to each of the time series data.

A time series data analyzer includes a segment condition input section, an analysis condition input section, and an optimum condition deriving section for analyzing all segments based on the segment conditions and analysis conditions inputted in the respective input sections, under all analysis conditions by a maximum entropy method and a nonlinear least squares method. The time series data analyzer derives the optimum segment length and the optimum lag value in correspondence to selected results, and an analysis execution section executes analysis by the maximum entropy method by setting the optimum analysis conditions derived as described above. The trending of the spectrum of electroencephalogram data is used as an indicator of the state of the subject based on the findings that the spectrum of electroencephalogram data is an exponential spectrum and the gradient changes depending on the state of the subject.

The invention provides an Eclat-based multivariate time series association rule mining method. the method comprises the steps of 1, generating a perpendicular dataset; 2, generating a MINHASH matrix,wherein the MINHASH matrix needs a designated parameter k; 3, utilizing the MINHASH matrix for estimating a candidate item set in an original data set; 4, according to the minimum support, pruning thecandidate item set to obtain frequent item sets 1; 5, combining two Hash frequent item sets 1 and generating a new frequent item set 2; 6, repeatedly executing the step 5 till combination cannot be performed, and ending an algorithm. The association rule mining speed is remarkably increased, the purpose of obtaining the time series data analysis result in time is achieved, even though the miningprecision is lowered, the mining efficiency can be greatly improved, and the machine memory can be saved.

A text data storage unit stores a plurality of text data having attribute data and time data. A dictionary storage unit stores a plurality of events each associated with text data. An analysis condition indication unit indicates an analysis target as attribute data and an analysis condition as an event sequence. A time series data generation unit assigns an event to each of the plurality of text data by referring to the dictionary storage unit, extracts a group of text data each having the same attribute data as the analysis target from the plurality of text data, and generates time series data each representing the event assigned to the text data of the group in order of the time data of the text data. A time series data analysis unit analyzes the time series data each having the same event sequence as the analysis condition.

The invention relates to the technical field of computer time series data analysis and prediction, in particular to a time series data filling and restoring method based on machine learning. The method includes: filling the missing value by using a domain-based median and mean value filling method; estimating a true value of an expected sampling moment through a linear rule; detecting wave crestsand wave troughs of the time sequence, and smoothing abnormal values; taking hundreds of thousands of collected real data as samples, designing and generating time sequence characteristics, taking real results as labels, and training a machine learning model based on an XGBoost (Extreme Gradient Boost) for predicting a large number of unknown data. According to the method, the problems of multiplemissing values, large volatility, error accumulation and the like of specific time sequence data are solved, and the accuracy of data filling and restoring is effectively improved; moreover, the complexity of a machine learning model is well controlled, the filling and restoration of hundreds of millions of data records can be completed within an hour level, and the method has a high practical value.

A time-series data analyzing apparatus which extracts a composite factor time-series pattern from time-series data. The apparatus includes a dividing device which divides the time-series data into pattern generation time-series data and pattern inspection time-series data which do not include pattern generation time-series data. A first generating device generates a transitional pattern including a support time data indicating a transition of support time and having a transition occurrence probability higher than a minimum occurrence probability in the pattern generation time-series data. A second generating device generates frequently appearing integrated transitional patterns. A second computing device computes cause-and-effect strength of each of the frequently appearing integrated transitional patterns using the pattern inspection time-series data. A display device displays the composite factor time-series pattern having the cause-and-effect strength higher than the minimum cause-and-effect strength given preliminarily.

The invention discloses a crop growth multi-angle remote sensing spectral detection device and a use method thereof. The device comprises an upright column; a control box is arranged on the upright column; an observation pipe is arranged at the top of the upright column; a crop canopy reflective spectral curve detection device is arranged at one end of the observation pipe; and a solar incidence spectral curve detection device is arranged at the other end of the observation pipe. The crop growth multi-angle remote sensing spectral detection device can integrate two spectrometers with differentwavelength ranges and different spectral resolution, realizes double-channel synchronous spectral detection, greatly reduces time difference among data and enhances comparability among data; the cropcanopy is subjected to different remote-sensing sensor spectral detection by utilizing double-channel design, the consistency of the detected target objects is guaranteed, and the reliability and thecredibility of long-time sequence data analysis result and scientific research are guaranteed; due to the automatic tracking solar and multi-angle detection crop canopy spectral curve, the data precision is improved and the credibility of a remote-sensing monitoring model is enhanced; the crop growth multi-angle remote sensing spectral detection device is simple in structure, convenient to use and convenient to carry and use in the field; and the cost is saved.

The present invention provides a time series data processing method and apparatus. The method comprises: according to a target enlargement position selected by a user on a time series data curve, determining a time interval on a timeline corresponding to the target enlargement position; and carrying out fisheye enlargement on a first time series data curve corresponding to the time interval, wherein the first series data curve comprises the series data curve of at least one object. The time series data processing method and apparatus provided by the present invention can improve efficiency of time series data analysis by the user according to the series data curve.

The invention discloses a human body lower limb explosive power evaluation method and device. The method comprises the steps of acquiring time sequence data of force applied to a flat plate by a subject in a longitudinal jump process on an evaluation hardware device, specifically, the longitudinal jump comprises static squatting jump, downward squatting jump and jump depth; according to the time sequence data, analyzing and confirming a mapping relationship between the pressure and the time of each stage, specifically, each stage comprises a squatting stage, a force exerting stage, a jumping stage, a flying stage and a landing stage; obtaining a plurality of evaluation parameters according to the mapping relationship between the pressure and the time of each stage; and based on the plurality of evaluation parameters, calculating muscle explosive power level parameters of the subject. By applying the method, the level parameters of the lower limb explosive power can be accurately evaluated, improvement can be guided and scientific training means and methods can be selected according to characteristics of different sports events on the bounce ability in combination with the evaluation result of the lower limb muscle explosive power level of athletes, the training of the lower limb bounce ability is enhanced, and the lower limb explosive power is further improved.

The invention discloses a lightweight unsupervised anomaly detection method based on multivariate time series data analysis. The method comprises two models: a detection model and an inference model; the detection model firstly extracts time dependence characteristics of the captured multivariate time series data through a random convolutional neural network, and then encodes and decodes the multivariate time series data after the characteristics are extracted by using a deep Bayesian network, and the detection model can determine a detection precision range; the inference model is composed of a score attention unit, a threshold value automatic selection unit and a point adjustment unit, the score attention unit adopts an attention mechanism to expand the feature difference between abnormal data and normal data and provides a theoretical basis for abnormal interpretation, and the threshold value automatic selection unit can automatically calculate a threshold value, the point adjustment unit can simulate the generation process of real anomalies, and the inference model can improve the accuracy, stability and interpretability of anomaly detection. According to the invention, the method can cope with rapidly increased data scale and complex and changeable exception types.

The invention discloses a similarity measurement computing method of linear dynamical systems. The method comprises the following steps that time series data to be processed are collected; linear dynamical system model parameters corresponding to the time series data are respectively computed; subspace included angles between the corresponding linear dynamical systems are obtained by computing the linear dynamical system model parameters; frame offset optimization is carried out on the subspace included angles to obtain the similarity measurement of the time series data. The method can be applied to time series data analysis, video content retrieval, intelligent video monitoring and other services.

The invention discloses a data center task scale prediction method based on time series data analysis, and the prediction method predicts the overall level of a data center task in a certain period oftime in the future through the collection and analysis of historical task input of a data center. The method can be applied to reasonable distribution of data center resources, data center jitter caused by frequent resource scheduling is avoided, and the overall service quality of the data center is reduced; and the method can also be applied to rapid detection and early warning of abnormal inputof the data center task scale when the deviation between the actual input quantity and the predicted value is too large.

To provide a time-series data analysis device which allows to compare time-series data easily.A time-series data analysis device analyzes the time-series data output by a machine tool, the time-series data analysis device including a time-series data acquisition unit configured to acquire the plurality of time-series data including operation conditions and operation results of the machine tool, a classification unit configured to classify the plurality of time-series data according to their respective operation conditions, a display control unit configured to perform control for displaying the plurality of time-series data according to the respective operation condition into which data was classified in the form of a list, and a calculation unit configured to calculate differences in the operation results for a plurality of time-series data selected from the plurality of time-series data included in one operation condition into which data was classified, wherein the display control unit performs control for displaying the calculated differences.

The invention provides a model training and data analysis method, device and equipmen for time series data and a storage medium. The method comprises the steps of constructing a training sample, the training sample comprising one or more pieces of time series data, different pieces of time series data corresponding to different feature dimensions, and each piece of time series data comprising a plurality of pieces of data arranged according to a time sequence; and training a time series data analysis model by using the one or more training samples, wherein the time series data analysis model is used for analyzing the category to which the sample comprising the one or more time series data belongs. Therefore, the automatic interpretation of the time sequence data can be realized by using the trained model.

The objective of the present invention is to efficiently and accurately obtain a time lag and a time window which are different for each explanatory variable in a multidimensional time series prediction problem. In determining a time lag and a time window, an explanatory variable time series is not, as-is, subject to normalization and optimization. Instead, once transformation to a cumulative time series, normalization, and optimization are performed, an optimal time lag and time window are determined thereby. By introducing a regularization term in the cumulative time series, the complexity of the obtained model is adjusted. Furthermore, by obtaining the weights of two outstanding cumulative values (which are of opposite signs) (simplified to that extent by normalization), it is possible to obtain the time lag and the time window therefrom.

The invention discloses a method and a device for analyzing time series data. The method is utilized to accurately and efficiently combine multiple lines of time series data into an event. The method comprises the following steps: searching for a start character in the time series data; matching the time series data after the start character according to the preset first characteristic information; combining the matched time series data into a same event, thereby acquiring an analysis result.

The invention discloses a wind turbine generator operation data interpolation method, which comprises the steps of determining a data type of missing data to be interpolated and a time window of the missing data to be interpolated, with the data type comprising environmental data and / or unit data; and determining an interpolation strategy corresponding to the data type according to the data type to complete interpolation of missing data. An operation data interpolation method formed on the basis of multivariate time series data analysis, working condition identification and deep learning is utilized, the method is practicable and reasonable in conclusion, the integrity of the operation data of the wind turbine generator is improved, and an accurate and reliable data basis is provided for subsequent data analysis and mining.

The invention discloses a method and a device for analyzing time series data. The method is utilized to accurately and directly display the time series data. The method comprises the following steps: extracting characteristic information of the time series data; searching for time series data with same characteristic information; analyzing if the time series data with same characteristic information are generated from a same source code; if yes, displaying the time series data with same characteristic information in a clustering form.

The invention discloses a continuous stirred tank reactor operation state monitoring method based on time series data analysis, and aims to solve the problem of monitoring the operation state of a continuous stirred tank reactor by monitoring the time series abnormal change of the real-time sampling data of the continuous stirred tank reactor. The method comprises the steps of firstly deducing a time series correlation characteristic analysis algorithm according to the maximum time series characteristic typical correlation coefficient; secondly, further using an auto-regression model to describe a time sequence dynamic relation between time sequence related characteristics, and finally achieving the purpose of monitoring the running state of the continuous stirred tank reactor by monitoring the errors of the auto-regression model. Compared with a traditional method, the method has the advantages that the potential characteristic components, which are typically related to a time sequence, of the sampling data can be effectively extracted, and the superiority and effectiveness of the method in monitoring the running state of the continuous stirred tank reactor are verified through specific embodiments.

The invention discloses a high-precision long-term time series prediction method based on multivariate time series data analysis, and the method comprises the steps: employing a discrete network to extract global features and local features of a multivariate time series in a layered and parallel manner; the calculation complexity is reduced while the multivariate time sequence prediction precision is improved, the model scale is reduced, and the prediction length of the model is increased. According to the method, a mechanism for extracting the global features and the local features of the multivariate time series in a layered and parallel manner is adopted, the prediction precision is improved, the memory usage amount of the model is reduced, the local features are utilized to improve the fitting capability of local fine fluctuations of the multivariate time series, and the prediction length of the model is increased; and the effect of the model on multivariate time sequence prediction is greatly improved.

The invention provides an autonomous learning business risk control rule engine system and a risk assessment method. The system comprises a service data fan out assembly, a time series data storage system, a time series data analysis system and a real-time risk control rule engine, and the service data fan out assembly performs mirror image copying on log data from a service system and distributesthe log data to the time series data storage system and the real-time risk control rule engine; the time series data storage system stores the log data as time series service data in a time series form; the time series data analysis system performs statistical analysis on the time series service data at fixed time intervals, finds out characteristic values of abnormal data and abstracts the characteristic values into dynamic prevention and control rules; and the real-time risk control rule engine executes predefined prevention and control rules and dynamic prevention and control rules on thelog data one by one, and returns a risk assessment result obtained by execution to the business system. The invention has the advantages that the real-time performance of risk control rule protectionis guaranteed, and the automation degree of the risk control rule engine is improved.

A machining environment measurement device, which enables automatic identification of a vibration source which may cause defects in a machined surface based on pieces of vibration data without performing an actual machining process, includes a reference machining data storage unit that stores reference machining data computed based on time-series data of vibration measured in advance in an ideal machining environment; a data acquisition unit that acquires time-series data of vibration of at least a holder attached to a spindle of the machine tool detected by a holder measurement sensor; an analysis unit that analyzes the time-series data and computes feature data indicating a feature of the time-series data; a comparative determination unit that compares the feature data with the reference machining data and determines a machining environment of the machine tool; and a display unit that displays the determination result of the comparative determination unit.

A time series data analysis apparatus: generates first internal data, based on first feature data groups, first internal parameter, and first learning parameter; transforms first feature data's position in a feature space, based on the first internal data and second learning parameter; reallocates the first feature data, based on a first transform result and first feature data groups; calculates a first predicted value, based on a reallocation result and third learning parameter; optimizes the first-third learning parameters by statistical gradient, based on a response variable and first predicted value; generates second internal data, based on second feature data groups, second internal parameter, and optimized first learning parameter; transforms the second feature data's position in a feature space, based on the second internal data and optimized second learning parameter; and calculates importance data for the second feature data, based on a second transform result and optimized third learning parameter.

The invention discloses a time series data motif identification method and device, and belongs to the field of time series data analysis. The time series data motif identification method comprises the steps that time series data needing to be analyzed are divided into at least two data subsequences, and each data subsequence is converted into a symbol subsequence; random projection is performed on the symbol subsequences, and the times that each projected symbol subsequence and other projected symbol subsequences have same signals are recorded; two data subsequences which are corresponding to the time exceeding a threshold value in the recorded times and between which the distance is smaller than a first preset distance serve as identified standard motifs; clustering is performed on the standard motifs in each preset range to obtain a center data subsequence, and the variance of each preset range is calculated according to the standard motifs in each preset range and the center data subsequence; the threshold value is decreased, the distances between two data subsequences which are corresponding to the time exceeding the decreased threshold value in the recorded times and the center data subsequence in the preset range where the two data subsequences are located, and the data subsequence of which the distance is smaller than the variance of the preset range where the data subsequence is located serves as the identified motif. According to the time series data motif identification method and device, under the condition that the motif identification speed is guaranteed, the motif identification accuracy can be improved.

A plurality of analytics in a cloud-based environment is accessed. Each of the plurality of analytics performs an operation on time series data. Within the cloud-based environment, a selected one or more of the plurality of analytics is chosen. A set of time series data is uploaded to the cloud-based environment and the selected one of the plurality of analytics is optimized on that set of time series data.