Time series data query method and system, readable storage medium and computer
By constructing a conditional generation model and performing data dimensionality upgrade processing, the accuracy problem of the time series data query method in the prior art when processing non-uniformly sampled data is solved, and more efficient and accurate data query results are achieved.
Patent Information
- Application Number
- CN202510254114.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-05
- Publication Date
- 2025-06-27
AI Technical Summary
When the existing time series data query method processes non-uniformly sampled data, it is easy to cause some valuable time series to be eliminated, reducing the accuracy of query results.
By obtaining the historical time series data set, extracting data labels and generating similar sequence data, building a condition generation model, performing data dimensionality upgrade processing and mark point recognition, constructing fuzzy query conditions, calculating the matching degree and outputting the query results.
It improves the accuracy and efficiency of time series data query, enhances the exploration and query capabilities of the data set, can better represent the complex structure in the data and maintain the slope characteristics of the data.
Smart Images
Figure CN120216563A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and particularly relates to a method, a system, a readable storage medium and a computer for querying time series data. Background Art
[0002] With the rapid development of technology and the improvement of people's living standards, time series data has played an increasingly important role in various fields.
[0003] Data query is a common type of interaction in big data analysis. Users can directly filter out unnecessary data information in the dataset, so as to further obtain the data subset that the user is interested in. For example, researchers in a certain university can use the time series query method to find out the attendance situation of students and their participation in extracurricular activities during a certain period. However, the current data query methods usually adopt visual query algorithms, which require all the primitive elements within the required time period to fully meet the given range constraints. Even if there is one primitive element outside the query range, its corresponding time series will be excluded, resulting in the exclusion of time series that can provide valuable information, leading to fewer matching information in the query results, thereby reducing the accuracy of the data query results. Summary of the Invention Based on this, the purpose of the present invention is to provide a method, a system, a readable storage medium and a computer for querying time series data to solve the above deficiencies in the technology.
[0004] The present invention proposes a method for querying time series data, including: Obtain a historical time series dataset, where each time series data in the historical time series dataset consists of an ordered list of two-dimensional points sampled densely and non-uniformly; Extract the data labels of each time series data in the historical time series dataset, and perform data processing according to the data labels and each time series data to obtain the similar sequence data of each time series data; Input each similar sequence data, each time series data and its corresponding data label into a preset model learner to construct a corresponding conditional generation model; Perform data dimensionality increase processing on the obtained current time series data to obtain an ordered point list of each current time series data, and given a data threshold, compare the ordered point list of each current time series data with the data threshold to obtain the marked points of each current time series data; Construct a fuzzy query condition, and input the fuzzy query condition and the marked points of each current time series data into the conditional generation model to calculate the matching degree of the current time series data, and output a query result according to the matching degree.
[0005] Further, the steps of extracting data tags of each time series data in the historical time series dataset and performing data processing based on the data tags and each time series data to obtain similar series data of each time series data include: Extract data tags of each time series data in the historical time series dataset, and input each time series data and its corresponding data expression into a preset model generator and model discriminator respectively; Given a sample data from the output of the model generator, and using the model discriminator to perform data discrimination on the sample data. If the sample data belongs to true sample data, then use the sample data to train the model generator to obtain similar series data of each time series data.
[0006] Further, the steps of inputting each similar series data, each time series data and its corresponding data tag into a preset model learner to construct a corresponding conditional generation model include: Construct a generative adversarial network model, and add a rectified linear function to the first two fully connected layers of the generative adversarial network model to obtain a generative adversarial network optimization model, where the generative adversarial network optimization model uses three fully connected layers, and the output of the previous connected layer is used as the input of the next connected layer; Construct a loss function of the generative adversarial network optimization model, and input each similar series data, each time series data and its corresponding data tag into the generative adversarial network optimization model to construct a corresponding conditional generation model.
[0007] Further, the steps of giving a data threshold and comparing the ordered point list of each current time series data with the data threshold to obtain the marked points of each current time series data include: Connect two adjacent ordered points in the ordered point list of each current time series data to obtain a corresponding sequence curve; Find the ordered point in the ordered point list of each current time series data that is farthest from the sequence curve. If the distance is greater than the data threshold, then use this point as the marked point.
[0008] Further, the steps of constructing a fuzzy query condition, inputting the fuzzy query condition and the marked points of each current time series data into the conditional generation model to calculate the matching degree of the current time series data, and outputting a query result according to the matching degree include: Define a fuzzy query condition , and screen out all the marker points within the query range of the fuzzy query condition among the marker points of each current time series data, and denote them as the marker point set; All the marker points are recorded as the marker point set; Input the marker point set and the fuzzy query condition into the condition generation module to calculate the matching degree of the current time series data, and extract the current time series data whose matching degree is greater than or equal to the matching rate of the fuzzy query condition to generate the corresponding query result.
[0009] The present invention also proposes a time series data query system, including: A data acquisition module for acquiring a historical time series data set, wherein each time series data in the historical time series data set consists of an ordered list of two-dimensional points sampled densely and non-uniformly; A data processing module for extracting the data labels of each time series data in the historical time series data set, and performing data processing based on the data labels and each time series data to obtain the similar sequence data of each time series data; A model construction module for inputting each similar sequence data, each time series data and its corresponding data label into a preset model learner to construct a corresponding condition generation model; A data comparison module for performing data dimensionality increase processing on the acquired current time series data to obtain an ordered point list of each current time series data, and given a data threshold, comparing the ordered point list of each current time series data with the data threshold to obtain the marker points of each current time series data; A data query module for constructing a fuzzy query condition, and inputting the fuzzy query condition and the marker points of each current time series data into the condition generation model to calculate the matching degree of the current time series data, and outputting a query result according to the matching degree.
[0010] Further, the data processing module includes: A label extraction unit for extracting the data labels of each time series data in the historical time series data set, and respectively inputting each time series data and its corresponding data expression into a preset model generator and a model discriminator; A data training unit for giving a sample data from the output of the model generator, and using the model discriminator to perform data discrimination on the sample data. If the sample data belongs to true sample data, then use the sample data to train the model generator to obtain the similar sequence data of each time series data.
[0011] Further, the model construction module includes: A model optimization unit, configured to construct a generative adversarial network model, and add a rectified linear function to the first two fully connected layers of the generative adversarial network model to obtain a generative adversarial network optimized model, wherein the generative adversarial network optimized model adopts three fully connected layers, and the output of the previous connection layer is used as the input of the subsequent connection layer; A model construction unit, configured to construct a loss function of the generative adversarial network optimized model, and input each of the similar sequence data, each of the time series data, and their corresponding data labels into the generative adversarial network optimized model to construct a corresponding conditional generation model.
[0012] Further, the data comparison module includes: A data processing unit, configured to connect two adjacent ordered points in the ordered point list of each of the current time series data to obtain a corresponding sequence curve; A data comparison unit, configured to find the ordered point with the farthest distance from the sequence curve in the ordered point list of each of the current time series data, and if the distance is greater than the data threshold, use this point as a marked point.
[0013] Further, the data query module includes: A condition definition unit, configured to define a fuzzy query condition and filter out all marked points passing through the query range of the fuzzy query condition among the marked points of each of the current time series data to form a marked point set; A data query unit, configured to input the marked point set and the fuzzy query condition into the conditional generation module to calculate the matching degree of the current time series data, and extract the current time series data with the matching degree greater than or equal to the matching rate of the fuzzy query condition to generate a corresponding query result.
[0014] The present invention also provides a readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned time series data query method is implemented.
[0015] The present invention also provides a computer, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the above-mentioned time series data query method is implemented.
[0016] The time series data query method, system, readable storage medium and computer in the present invention extract data tags of each time series data in the historical time series dataset, and use the data tags and each time series data for data processing to obtain similar sequence data of the time series data. The similar sequence data, each time series data and its corresponding data tags are input into a preset model learner to construct a conditional generation model, and the model is used to learn the distribution law of the data in the dataset to generate sufficient samples for each data in the dataset, thereby improving the estimation accuracy and precision of the dataset; performing data dimensionality increase processing on the obtained current time series data can better represent the complex structure in the data, using a preset data threshold to obtain the marking points of each current time series data, thereby better maintaining the slope characteristics of the data, constructing a fuzzy query condition, calculating the matching degree using the fuzzy query condition, marking points and the conditional generation model, and outputting the query result according to the matching degree. Based on the fuzzy query method, the exploration and query capabilities of the time series dataset are enhanced, and thus the data query efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 is a flowchart of the time series data query method in the first embodiment of the present invention; Figure 2 is Figure 1 a detailed flowchart of step S102 in Figure 3 is Figure 1 a detailed flowchart of step S103 in Figure 4 is Figure 1 a detailed flowchart of step S104 in Figure 5 is a structural block diagram of the time series data query system in the second embodiment of the present invention; Figure 6 is a structural block diagram of the computer in the third embodiment of the present invention.
[0018] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. SPECIFIC EMBODIMENTS
[0019] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.
[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs. The terms used in the specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.
[0021] Embodiment 1 Please refer to Figure 1 , which shows the time series data query method in the first embodiment of the present invention. The method specifically includes steps S101 to S105: S101. Obtain a historical time series data set, where each time series data in the historical time series data set consists of an ordered list of two-dimensional points sampled densely and non-uniformly; In specific implementation, obtain a historical time series data set , where the historical time series data set each time series data in consists of an ordered list of two-dimensional points sampled densely and non-uniformly composed of, where represents the th time series data in the historical time series data set , represents the matching rate, which is used to represent what proportion of the data needs to meet the constraint conditions, represents the historical time series data set the th ordered list of time series data, represents the th two-dimensional point in.
[0022] S102. Extract the data labels of each time series data in the historical time series data set, and perform data processing based on the data labels and each time series data to obtain the similar sequence data of each time series data; Further, please refer to Figure 2 , the step S102 specifically includes steps S1021 to S1022: S1021. Extract the data labels of each time series data in the historical time series data set, and respectively input each time series data and its corresponding data expression into a preset model generator and model discriminator; S1022, Given a sample data from the output of the model generator, and using the model discriminator to perform data discrimination on the sample data. If the sample data belongs to true sample data, then use the sample data to train the model generator to obtain similar sequence data for each of the time series data.
[0023] In specific implementation, extract the data labels of each time series data in each historical event sequence dataset, update the parameters of the preset model discriminator, fix the parameters of the preset model generator, input the random noise of each time series data into the model generator, use the output of the model generator as negative samples, and input each time series data as positive samples into the model discriminator; Furthermore, update the parameters of the model generator, fix the parameters of the model discriminator, train the model generator to learn the data distribution of each time series data. When inputting random noise, the model generator will generate generated data similar to each time series data. After the model discriminator calculates the error, perform loop operations until the model discriminator and the model generator meet the preset requirements (in this embodiment, the preset requirement is that the model generator and the model discriminator reach Nash equilibrium). At this time, the model generator will output similar sequence data for each time series data.
[0024] S103, Input each of the similar sequence data, each of the time series data, and their corresponding data labels into a preset model learner to construct a corresponding conditional generation model; Furthermore, please refer to Figure 3 , The step S103 specifically includes steps S1031~S1032: S1031, Construct a generative adversarial network model, and add a rectified linear function to the first two fully connected layers of the generative adversarial network model to obtain a generative adversarial network optimization model, where the generative adversarial network optimization model uses three fully connected layers, and the output of the previous connected layer is used as the input of the next connected layer; S1032, Construct the loss function of the generative adversarial network optimization model, and input each of the similar sequence data, each of the time series data, and their corresponding data labels into the generative adversarial network optimization model to construct a corresponding conditional generation model.
[0025] In specific implementation, construct a generative adversarial network model, and add a rectified linear function to the first two fully connected layers of the generative adversarial network model to obtain a generative adversarial network optimization model, where the generative adversarial network optimization model uses three fully connected layers, and the output of the previous connected layer is used as the input of the next connected layer. In this embodiment, the rectified linear function is selected as the LeakyReLU function; Further, construct the loss function of the above-mentioned generative adversarial network optimization model, input the obtained similar sequence data, each time series data and its corresponding data label into the generative adversarial network optimization model, and perform sequential processing through the fully connected layer of the generative adversarial network optimization model to construct a corresponding conditional generation model.
[0026] S104. Perform data dimensionality increase processing on the obtained current time series data to obtain an ordered point list of each current time series data, and given a data threshold, compare the ordered point list of each current time series data with the data threshold to obtain the marked points of each current time series data. Further, please refer to Figure 4 , and the step S104 specifically includes steps S1041 to S1042: S1041. Connect two adjacent ordered points in the ordered point list of each current time series data to obtain a corresponding sequence curve. S1042. Find the ordered point in the ordered point list of each current time series data that is farthest from the sequence curve. If the distance is greater than the data threshold, use this point as the marked point.
[0027] In specific implementation, given a data threshold , use the digital image processing algorithm to compare the ordered point list of the obtained current time series data with the data threshold, and add and to the result set to connect two adjacent ordered points in the ordered point list of each current time series data, and construct a sequence curve from point to point . Specifically, find the point in the ordered point list of each current time series data that is farthest from the sequence curve, and determine whether the distance is greater than the above-mentioned data threshold. If the distance is greater than the data threshold, use this point as the marked point. This embodiment simplifies the data using the data threshold to facilitate indexing through the simplified data.
[0028] S105. Construct a fuzzy query condition, and input the fuzzy query condition and the marked points of each current time series data into the conditional generation model to calculate the matching degree of the current time series data, and output a query result according to the matching degree.
[0029] Further, the step S105 specifically includes steps S1051 to S1052: S1051. Define the fuzzy query condition , and screen out all the marked points in each of the current time series data that pass through the query range of the fuzzy query condition as the marked point set; S1052, input the marked point set and the fuzzy query condition into the condition generation module to calculate the matching degree of the current time series data, and extract the current time series data with the matching degree greater than or equal to the matching rate of the fuzzy query condition to generate the corresponding query result.
[0030] In specific implementation, define a fuzzy query condition, which is a binary tuple , where the query range is a six-tuple in the topological vector space, including the start time and end time of the query range , the start value and end value of the fuzzy query, and the start angular slope and end angular slope of the fuzzy query; perform range traversal through the preset spatial indexing algorithm, so as to screen the marked points of the above current time series data with this fuzzy query condition to obtain all the marked points passing through the query range as the marked point set; Specifically, traverse the query range to find the marked curve formed by all the marked points intersecting with the topological vector space defined by the query range , and at the same time record the set of all primitive elements (the primitive elements include points or line segments) intersecting with the query range on this marked curve.
[0031] Furthermore, input the above marked point set and the fuzzy query condition into the condition generation module, and perform binary search on the point segment list in the marked point set through the condition generation module, so as to obtain the minimum index value and maximum index value of the primitive elements of the corresponding current time series data in the query range in its point segment list, and calculate the matching degree of the current time series from the number of primitive elements, this minimum index value and this maximum index value obtained above: ; In the formula, represents the number of primitive elements calculated from the th current time series data and the fuzzy query condition, , respectively represent the maximum index value and minimum index value of the primitive elements of the th current time series data in the query range in its point segment list, represents the compensation coefficient, which is set by the user.
[0032] Further, after obtaining the matching degree of the above current time series data, the current time series data with a matching rate greater than or equal to the fuzzy query condition is extracted, and the extracted current time series data is the query result of the fuzzy query condition.
[0033] In summary, in the time series data query method in the above embodiments of the present invention, by extracting the data tags of each time series data in the historical time series data set, and using the data tags and each time series data for data processing to obtain the similar series data of the time series data, inputting the similar series data, each time series data and its corresponding data tags into a preset model learner to construct a conditional generation model, and using the model to learn the distribution law of the data in the data set to generate sufficient samples for each data in the data set, thereby improving the estimation accuracy and accuracy of the data set; performing data dimensionality increase processing on the obtained current time series data can better represent the complex structure in the data, using a preset data threshold to obtain the marked points of each current time series data, thereby better maintaining the slope characteristics of the data, constructing a fuzzy query condition, using the fuzzy query condition, marked points and conditional generation model to calculate the matching degree, and outputting the query result according to the matching degree, enhancing the exploration and query capabilities of the time series data set based on the fuzzy query method, and further improving the data query efficiency.
[0034] Embodiment 2 On the other hand, the present invention also proposes a time series data query system. Please refer to Figure 5 which shows the time series data query system in the second embodiment of the present invention. The system includes: A data acquisition module 11, configured to acquire a historical time series data set, where each time series data in the historical time series data set consists of an ordered list of two-dimensional points with dense and non-uniform sampling; A data processing module 12, configured to extract the data tags of each time series data in the historical time series data set, and perform data processing according to the data tags and each time series data to obtain the similar series data of each time series data; Further, the data processing module 12 includes: A label extraction unit, configured to extract the data tags of each time series data in the historical time series data set, and respectively input each time series data and its corresponding data expression into a preset model generator and a model discriminator; A data training unit, configured to obtain a sample data from the output of the model generator, and use the model discriminator to perform data discrimination on the sample data. If the sample data belongs to true sample data, the sample data is used to train the model generator to obtain similar sequence data for each of the time series data.
[0035] A model construction module 13, configured to input each of the similar sequence data, each of the time series data, and their corresponding data labels into a preset model learner to construct a corresponding conditional generation model. Further, the model construction module 13 includes: A model optimization unit, configured to construct a generative adversarial network model, and add a rectified linear function to the first two fully connected layers of the generative adversarial network model to obtain a generative adversarial network optimization model. The generative adversarial network optimization model adopts three fully connected layers, and the output of the previous connection layer is used as the input of the next connection layer. A model construction unit, configured to construct a loss function of the generative adversarial network optimization model, and input each of the similar sequence data, each of the time series data, and their corresponding data labels into the generative adversarial network optimization model to construct a corresponding conditional generation model.
[0036] A data comparison module 14, configured to perform data dimensionality increase processing on the obtained current time series data to obtain an ordered point list for each of the current time series data, and set a data threshold, and compare the ordered point list of each of the current time series data with the data threshold to obtain a marked point for each of the current time series data. Further, the data comparison module 14 includes: A data processing unit, configured to connect two adjacent ordered points in the ordered point list of each of the current time series data to obtain a corresponding sequence curve. A data comparison unit, configured to find the ordered point with the farthest distance from the sequence curve in the ordered point list of each of the current time series data. If the distance is greater than the data threshold, the point is used as the marked point.
[0037] A data query module 15, configured to construct a fuzzy query condition, and input the fuzzy query condition and the marked points of each of the current time series data into the conditional generation model to calculate the matching degree of the current time series data, and output a query result according to the matching degree.
[0038] Further, the data query module 15 includes: A condition definition unit, configured to define a fuzzy query condition , and screen out all the marker points within the query range of the fuzzy query condition among the marker points of each current time series data, and denote them as the marker point set; All the marker points are denoted as the marker point set; A data query unit, configured to input the marker point set and the fuzzy query condition into the condition generation module to calculate the matching degree of the current time series data, and extract the current time series data with the matching degree greater than or equal to the matching rate of the fuzzy query condition to generate a corresponding query result.
[0039] The functions or operation steps implemented when the above modules and units are executed are substantially the same as those in the above method embodiments, and will not be elaborated here.
[0040] The time series data query system provided by the embodiments of the present invention has the same implementation principle and the same technical effects as those in the foregoing method embodiments. For a brief description, for the parts not mentioned in the system embodiments, reference can be made to the corresponding content in the foregoing method embodiments.
[0041] Embodiment 3 The present invention also proposes a computer. Please refer to Figure 6 . As shown in the figure, the computer in the third embodiment of the present invention includes a memory 10, a processor 20, and a computer program 30 stored on the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, the above time series data query method is implemented.
[0042] Among them, the memory 10 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc. The memory 10 may be an internal storage unit of the computer in some embodiments, such as the hard disk of the computer. The memory 10 may also be an external storage device in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. Further, the memory 10 may also include both an internal storage unit of the computer and an external storage device. The memory 10 can be used not only to store application software and various types of data installed on the computer, but also to temporarily store data that has been output or will be output.
[0043] Among them, in some embodiments, the processor 20 may be an Electronic Control Unit (ECU, also known as the vehicle computer), a Central Processing Unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips, and is used to run the program code stored in the memory 10 or process data, such as executing an access restriction program, etc.
[0044] It should be noted that Figure 6 The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown in the figure, or combine certain components, or have different component arrangements.
[0045] An embodiment of the present invention also provides a readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the time series data query method as described above is implemented.
[0046] Those skilled in the art can understand that the logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
[0047] More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection part (electronic device) having one or more wirings, a portable computer disk cartridge (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or processing it in other suitable ways as necessary, and then storing it in a computer memory.
[0048] It should be understood that each part of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.
[0049] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.
[0050] The above-described embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.
Claims
1. A time series data query method, characterized in that: include: Acquire a historical time series data set, wherein each time series data in the historical time series data set consists of an ordered list of densely and non-uniformly sampled two-dimensional points; Extracting data labels of each time series data in the historical time series data set, and performing data processing according to the data labels and each time series data to obtain similar sequence data of each time series data; Inputting each of the similar sequence data, each of the time series data and their corresponding data labels into a preset model learner to construct a corresponding conditional generation model; Performing data dimension upscaling processing on the acquired current time series data to obtain an ordered point list of each current time series data, and giving a data threshold, comparing the ordered point list of each current time series data with the data threshold to obtain a marking point of each current time series data; Construct a fuzzy query condition, and input the fuzzy query condition and each mark point of the current time series data into the condition generation model to calculate the matching degree of the current time series data, and output the query result according to the matching degree.
2. The time series data query method according to claim 1, characterized in that: The steps of extracting data labels of each time series data in the historical time series data set, and performing data processing according to the data labels and each time series data to obtain similar sequence data of each time series data include: Extracting data labels of each time series data in the historical time series data set, and inputting each time series data and its corresponding data expression into a preset model generator and model discriminator respectively; A sample data is given from the output of the model generator, and the model discriminator is used to perform data discrimination on the sample data. If the sample data belongs to true sample data, the model generator is trained using the sample data to obtain similar sequence data of each time series data.
3. The time series data query method according to claim 1, characterized in that: The steps of inputting each of the similar sequence data, each of the time series data and their corresponding data labels into a preset model learner to construct a corresponding conditional generation model include: Constructing a generative adversarial network model, and adding a modified linear function to the first two fully connected layers of the generative adversarial network model to obtain a generative adversarial network optimization model, wherein the generative adversarial network optimization model adopts three fully connected layers, and the output of the previous connection layer is used as the input of the next connection layer; A loss function of the generative adversarial network optimization model is constructed, and each of the similar sequence data, each of the time series data and their corresponding data labels are input into the generative adversarial network optimization model to construct a corresponding conditional generation model.
4. The time series data query method according to claim 1, characterized in that: Given a data threshold, the step of comparing the ordered point list of each current time series data with the data threshold to obtain the marked point of each current time series data includes: Connecting two adjacent ordered points in the ordered point list of each current time series data to obtain a corresponding sequence curve; In the ordered point list of each current time series data, find the ordered point farthest from the sequence curve, and if the distance is greater than the data threshold, use the point as a marking point.
5. The time series data query method according to claim 1, characterized in that: The steps of constructing a fuzzy query condition, inputting the fuzzy query condition and the marking points of each current time series data into the condition generation model to calculate the matching degree of the current time series data, and outputting the query result according to the matching degree include: Define fuzzy query conditions , and filter out the query range of the mark points of each current time series data that pass through the fuzzy query condition All the marked points of are recorded as the marked point set; The set of marked points and the fuzzy query conditions are input into the condition generation module to calculate the matching degree of the current time series data, and the matching rate of the fuzzy query conditions is greater than or equal to the matching degree. The current time series data is extracted to generate the corresponding query results.
6. A time series data query system, characterized in that: include: A data acquisition module, used to acquire a historical time series data set, wherein each time series data in the historical time series data set consists of an ordered list of densely and non-uniformly sampled two-dimensional points; A data processing module, used for extracting data labels of each time series data in the historical time series data set, and performing data processing according to the data labels and each time series data to obtain similar sequence data of each time series data; A model building module, used for inputting each of the similar sequence data, each of the time series data and their corresponding data labels into a preset model learner to build a corresponding conditional generation model; A data comparison module is used to perform data dimension upscaling processing on the acquired current time series data to obtain an ordered point list of each current time series data, and to give a data threshold, compare the ordered point list of each current time series data with the data threshold to obtain a marking point of each current time series data; The data query module is used to construct fuzzy query conditions, and input the fuzzy query conditions and the marking points of each current time series data into the condition generation model to calculate the matching degree of the current time series data, and output the query result according to the matching degree.
7. The time series data query system according to claim 6, characterized in that: The data processing module comprises: A label extraction unit, used to extract data labels of each time series data in the historical time series data set, and input each time series data and its corresponding data expression into a preset model generator and model discriminator respectively; A data training unit is used to give a sample data from the output of the model generator, and use the model discriminator to perform data discrimination on the sample data. If the sample data belongs to true sample data, the model generator is trained using the sample data to obtain similar sequence data of each time series data.
8. The time series data query system according to claim 6, characterized in that: The model building module includes: A model optimization unit, used to construct a generative adversarial network model, and add a modified linear function to the first two fully connected layers of the generative adversarial network model to obtain a generative adversarial network optimization model, wherein the generative adversarial network optimization model adopts three fully connected layers, and the output of the previous connection layer is used as the input of the next connection layer; A model building unit is used to build a loss function of the generative adversarial network optimization model, and input each of the similar sequence data, each of the time series data and their corresponding data labels into the generative adversarial network optimization model to build a corresponding conditional generation model.
9. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the time series data query method as described in any one of claims 1 to 5 is implemented.
10. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the time series data query method as described in any one of claims 1 to 5 is implemented.