Data processing and data analysis system and method based on block chain

Abstract

The invention discloses a data processing and data analysis system and method based on a block chain, and the method comprises the following steps: S1, obtaining time series data in real time, and carrying out the data preprocessing of the time series data, and obtaining a training sample; s2, inputting the training sample into a neural network for training, and outputting a convolution kernel carrying time sequence data information; s3, constructing a pruning model, pruning the convolution kernel according to a pruning rule, and outputting a result; s4, an output result is compared with the time sequence data, if the output result is correct, trimming succeeds, and if the output result is wrong, the steps S2-S4 are repeated; according to the method, the pruning model is adopted to compress the data in the neural network, and the pruning model can simultaneously reduce the volume of the data and the training amount of the data and can be combined with other data compression methods more easily, so that the compression of the time series data is realized, and the storage space of the time series data is reduced.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a blockchain-based data processing and data analysis system and method. Background Art

[0002] Blockchain (or block chain) is a shared, distributed database technology that uses a distributed database to identify, disseminate and record information in an intelligent peer-to-peer network. First of all, the main function of blockchain is to store information. Any information that needs to be saved can be written into the blockchain and can also be read from it, so it is a database.

[0003] A time series database, short for time series database, refers to a specialized database system that stores, queries, and analyzes time-tagged data (data that changes in chronological order, i.e., time series). It can also be used as a type of blockchain database. Currently, two types of data, represented by the Internet of Things and financial analysis, continuously generate large amounts of sensor values ​​or event data over time. Time series data is a continuous sequence of values ​​formed by the time (timestamp) when the data (event) occurs. Whether it is machine-generated sensor data or social event data generated by human activities, they share some common characteristics:

[0004] 1. High acquisition frequency: dozens, hundreds, hundreds of thousands or even millions of times per second;

[0005] 2. High acquisition accuracy: at least support millisecond-level acquisition, some need to support microsecond-level and nanosecond-level acquisition;

[0006] 3. Large collection span: 7*24 hours continuous collection of data for several years or even decades;

[0007] 4. Long storage period: It is necessary to support the persistent storage of time series data, and some data may even need to be permanently stored for hundreds of years (such as earthquake data);

[0008] Therefore, high requirements are placed on the storage of time series databases.

[0009] Chinese patent document publication number CN109684328A discloses a method for compressing and storing high-dimensional time series data. Based on the dimension of the time series data and the number of data points after compression of each dimension, the input time series data is linearly scanned. During the scanning process, each component data point is merged using an interval merging method. The start and end time tags of all merged component intervals are extracted, globally sorted, and time series data is regenerated based on the sorted time tags. Finally, the generated time series data is stored in a relational database. This method can be applied to the compressed storage of time series data in various industrial control systems. It is suitable for situations with large data flow and a variety of sampling types. It can better preserve the characteristics of the original data, has good real-time performance, a high compression ratio, and strong practicality. The compressed results can be stored in a relational database. Existing technologies generally directly compress and store time series data, but this ignores the fact that time series data generally contains a lot of identical data information. Therefore, further processing the time series data before compression to reduce the storage volume of its original data information can further reduce the compressed volume of the time series data.

[0010] To this end, a data processing and data analysis system and method based on blockchain is proposed. Summary of the Invention

[0011] In view of this, the embodiments of the present invention hope to provide a blockchain-based data processing and data analysis system and method to solve or alleviate the technical problems existing in the existing technology and at least provide a beneficial option;

[0012] The technical solution of the embodiment of the present invention is implemented as follows: First, a data processing and data analysis method based on blockchain is proposed, comprising the following steps:

[0013] S1: Acquire time series data in real time, perform data preprocessing on the time series data, and obtain training samples;

[0014] S2: Input the training sample into the neural network for training, and output a convolution kernel carrying time series data information;

[0015] S3: Build a pruning model, prune the convolution kernel according to the pruning rules, and output the results;

[0016] S4: Compare the output result with the time series data. If correct, the pruning is successful. If wrong, repeat steps S2-S4.

[0017] After pruning the convolution kernel, the following is satisfied:

[0018] f(M',D,M)=min(val(M,D)-val(M',D));

[0019]

[0020] Where val(M,D) is the accuracy of the convolution kernel M on the time series data D, val(M',D) is the accuracy of the pruned convolution kernel M' on the time series data D, size(M) is the volume of the convolution kernel M, and size(M') is the volume of the pruned convolution kernel M'.

[0021] Furthermore, in the blockchain-based data processing and data analysis method, the step of inputting the training samples into a neural network for training and outputting a convolution kernel carrying time series data information includes:

[0022] The training sample is compared with the previous training sample. If there is the same training sample information, the same pruning rule is used to prune the same convolution kernel.

[0023] Furthermore, in the blockchain-based data processing and data analysis method, in S3, the step of constructing the pruning model includes:

[0024] Use time series data for forward propagation and backpropagation to obtain the initial model weights of the time series data;

[0025] Evaluate the importance of each convolution kernel in the time series data to the model based on gradient matrix analysis;

[0026] The convolution kernels are soft-pruned according to the importance evaluation results. The soft pruning operation includes: after pruning some convolution kernels, continuing training and adjusting the weights of the remaining convolution kernels, and continuing pruning after the model converges until the final set pruning rate is achieved.

[0027] Furthermore, in the blockchain-based data processing and data analysis method, the step of using time series data for forward propagation and backward propagation to obtain the initial model weight of the time series data includes:

[0028] The convolution kernel is subjected to a gradient-free separation operation, which is separated into the direction convolution kernel and the convolution kernel radius;

[0029] Perform a gradient convolution kernel recovery operation, forward propagate the recovered convolution kernel, and then obtain the directional convolution and the gradient of the convolution kernel radius parameter through backpropagation;

[0030] Update the weights according to the gradient.

[0031] Furthermore, in the blockchain-based data processing and data analysis method, since the weight values ​​of the directional convolution kernel and the convolution kernel radius parameters change after the convolution kernel performs gradient update through backpropagation, the directional convolution kernel and the convolution kernel radius parameters lose their original definitions and need to be restored to their original definitions before the next forward propagation:

[0032]

[0033] Recover the true weight TW of the convolution kernel i l , separate the weights into convolution kernel radius parameters and directional convolution kernel W i l .

[0034] Furthermore, in the blockchain-based data processing and data analysis method, the volume of the convolution kernel M after pruning is:

[0035] The convolution kernel M is pruned after the convolution layer of L layers.

[0036] size(M)=∑ L＝1 (1-p L )size(L);

[0037] Where size(L) is the volume of the L-th convolution kernel, p L is the pruning rate of the Lth convolutional layer.

[0038] In another aspect, a blockchain-based data processing and data analysis system is proposed, which is applied to the steps of any of the blockchain-based data processing and data analysis methods described above, and the data processing and analysis system includes:

[0039] Acquisition unit: acquires time series data in real time, performs data preprocessing on the time series data, and obtains training samples;

[0040] Training unit: inputs the training samples into the neural network for training, and outputs a convolution kernel carrying time series data information;

[0041] Construction unit: builds a pruning model, prunes the convolution kernel according to the pruning rules, and outputs the results;

[0042] Comparison unit: compares the output result with the time series data. If they are correct, the pruning is successful. If they are wrong, repeat steps S2-S4.

[0043] Furthermore, the data processing and data analysis system based on blockchain further includes a temporary storage unit;

[0044] The temporary storage unit is used to store the previous training sample.

[0045] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any one of the methods for data processing and data analysis based on blockchain.

[0046] A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any one of the methods for data processing and data analysis based on blockchain are implemented.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] 1. The present invention proposes a blockchain-based data processing and data analysis method including a pruning model. This method processes traditional time series data before compressing it. The pruning model compresses the data in the neural network. The time series data input into the neural network can be pre-processed to eliminate noise and interference during data transmission. The pruning model can also reduce the volume of data and the amount of data training. The pruning model is also easier to combine with other data compression methods to achieve pre-processing of time series data, thereby achieving time series data compression and ensuring that it occupies minimal storage space.

[0049] 2. The present invention proposes a data processing and data analysis method based on blockchain. According to the characteristics of time series data, time series data is data recorded in chronological order, and the data points are arranged in chronological order. Therefore, according to the correlation of their data, the previous time series and the next time series are identical or similar in some aspects. Here, the same pruning rules can be used to prune them, reducing the computational complexity of the model and speeding up the processing speed of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 Schematic diagram of a data processing and data analysis method based on blockchain of the present invention;

[0052] Figure 2 Schematic diagram of a pruning model in a blockchain-based data processing and data analysis method of the present invention;

[0053] Figure 3 A schematic diagram of a computer device for implementing a blockchain-based data processing and data analysis method according to the present invention;

[0054] Figure 4 Schematic diagram of a blockchain-based data processing and data analysis system of the present invention. DETAILED DESCRIPTION

[0055] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. The following description sets forth many specific details to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art can make similar improvements without violating the scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0056] It should be noted that the terms "first," "second," "symmetrical," "array," etc. are used only to distinguish descriptions from positional descriptions and should not be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, limitations on features such as "first" and "symmetrical" may explicitly or implicitly include one or more of these features; similarly, when the number of certain features is not limited in the form of words such as "two" or "three," it should be noted that these features also explicitly or implicitly include one or more of the number of features.

[0057] In the present invention, unless otherwise expressly specified or limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediate medium. Furthermore, a first feature being "above," "above," or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is at a higher level than the second feature. A first feature being "below," "below," or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is at a lower level than the second feature. At the same time, all axial descriptions, such as the X-axis, Y-axis, Z-axis, one end of the X-axis, the other end of the Y-axis, or the other end of the Z-axis, are based on a Cartesian coordinate system.

[0058] In the present invention, unless otherwise expressly specified or limited, terms such as "installation," "connection," and "fixation" should be understood broadly; for example, they may refer to fixed connection, detachable connection, or integral molding; they may refer to mechanical connection, direct connection, welding, or indirect connection through an intermediate medium; they may refer to internal communication between two components or interaction between two components. Those skilled in the art will be able to understand the specific meanings of the above terms in the present invention based on the specification and drawings in conjunction with specific circumstances.

[0059] In the existing technology, time series data is usually compressed and stored directly, but it ignores the fact that time series data generally contains a lot of identical data information. Therefore, further processing the time series data before compression to reduce the storage volume of its original data information can further reduce the compressed volume of time series data. To this end, please refer to Figures 1-4 , the present invention provides a technical solution to solve the above technical problems: a specific method for using a data processing and data analysis system based on blockchain;

[0060] In some specific embodiments of this application, please refer to Figure 1 :

[0061] A data processing and data analysis method based on blockchain is proposed, which includes the following steps:

[0062] S1: Acquire time series data in real time, perform data preprocessing on the time series data, and obtain training samples;

[0063] S2: Input the training sample into the neural network for training, and output a convolution kernel carrying time series data information;

[0064] S3: Build a pruning model, prune the convolution kernel according to the pruning rules, and output the results;

[0065] S4: Compare the output result with the time series data. If correct, the pruning is successful. If wrong, repeat steps S2-S4.

[0066] After pruning the convolution kernel, the following is satisfied:

[0067] f(M',D,M)=min(val(M,D)-val(M',D));

[0068]

[0069] Where val(M,D) is the accuracy of the convolution kernel M on the time series data D, val(M',D) is the accuracy of the pruned convolution kernel M' on the time series data D, size(M) is the volume of the convolution kernel M, and size(M') is the volume of the pruned convolution kernel M'.

[0070] In this embodiment, in response to the defects of the prior art, this embodiment further reduces the storage volume occupied by time series data by processing the time series data before compressing it; first, real-time time series data is obtained. Here, the real-time time series data does not include the first time series data provided by a data source. This solution is implemented with the second time series data and the time series data thereafter. After the time series data is obtained, the time series data is preprocessed and training samples are generated and input into the neural network. In this process, there is an error between the time series data D and the convolution kernel M. The error is due to the noise and uncertainty factors in the data transmission process, which are eliminated by the neural network during the training process. In the process of pruning, the error is satisfied. Only then can it be determined that the convolution kernel has been pruned. When the accuracy of the pruned convolution kernel M' on the time series data D is at a preset level with the accuracy of the convolution kernel M before pruning on the time series data D, the correct storage information can be output, then the pruning is successful, otherwise, the pruning is repeated; the data in the neural network is compressed by the pruning model, and the time series data input into the neural network can eliminate noise and interference during data transmission through preprocessing, and the pruning model can simultaneously reduce the volume of the data and the amount of data training. The pruning model is also easier to combine with other data compression methods to realize the advance processing of the time series data, so as to realize the compression of the time series data and ensure that it occupies the minimum storage space.

[0071] In a specific embodiment, it is assumed that the temperature sensor data transmits the following information: time 12:01, location XX, temperature 38.5°C, self-status: good, over-temperature alarm: not reported; in this time series, if the system wants to store data related to temperature data, then the data information of self-status: good and over-temperature alarm: not reported does not need to be stored, and it is pruned in the pruning model.

[0072] In one embodiment, the step of inputting the training samples into a neural network for training and outputting a convolution kernel carrying time series data information includes:

[0073] The training sample is compared with the previous training sample. If there is the same training sample information, the same pruning rule is used to prune the same convolution kernel.

[0074] In a specific embodiment, the same data as in the previous embodiment is assumed to be transmitted by the temperature sensor data as follows: time 12:01, location XX, temperature 38.5°C, self-status: good, over-temperature alarm: not reported; and the data information of the previous time series data is also known to be: time 12:00, location XX, temperature 38.3°C, self-status: good, over-temperature alarm: not reported; the previous time series data has been pruned, and the same pruning rules can be used to prune the current time series data. Secondly, Since the locations that appeared in the previous time series are the same, this data feature can also be pruned. According to the time interval, the time information does not need to be saved. Finally, it is pruned into a convolution kernel containing only temperature. This embodiment is based on the characteristics of time series data. Time series data is recorded in chronological order, and the data points are arranged in chronological order. Therefore, according to the correlation of the data, the previous time series and the next time series are identical or similar in some aspects. Here, the same pruning rules can be used to prune them, reducing the computational complexity of the model and speeding up the processing speed of the model.

[0075] Furthermore, in S3, the step of constructing the pruning model includes:

[0076] S31: Use time series data for forward propagation and backward propagation to obtain the initial model weights of the time series data;

[0077] S32: Evaluate the importance of each convolution kernel in the time series data to the model according to gradient matrix analysis;

[0078] S33: Perform a soft pruning operation on the convolution kernel according to the importance evaluation result. The soft pruning operation includes: after pruning some convolution kernels, continuing training and adjusting the weights of the remaining convolution kernels, and continuing pruning after the model converges until the final set pruning rate is achieved.

[0079] Furthermore, the step of using the time series data for forward propagation and backward propagation to obtain the initial model weights of the time series data includes:

[0080] The convolution kernel is subjected to a gradient-free separation operation, which is separated into the direction convolution kernel and the convolution kernel radius;

[0081] Perform a gradient convolution kernel recovery operation, forward propagate the recovered convolution kernel, and then obtain the directional convolution and the gradient of the convolution kernel radius parameter through backpropagation;

[0082] Update the weights according to the gradient.

[0083] Furthermore, since the weight values ​​of the directional convolution kernel and the convolution kernel radius parameters change after the convolution kernel is updated through back propagation, the directional convolution kernel and the convolution kernel radius parameters lose their original definitions and need to be restored before the next forward propagation:

[0084]

[0085] Recover the true weight TW of the convolution kernel i l , separate the weights into convolution kernel radius parameters and directional convolution kernel W i l .

[0086] Furthermore, the volume of the convolution kernel M after pruning is:

[0087] The convolution kernel M is pruned after the convolution layer of L layers.

[0088] size(M)=∑ L＝1 (1-p L )size(L);

[0089] Where size(L) is the volume of the L-th convolution kernel, p L is the pruning rate of the Lth convolutional layer.

[0090] In another aspect, a blockchain-based data processing and data analysis system is proposed, which is applied to the steps of any of the blockchain-based data processing and data analysis methods described above, and the data processing and analysis system includes:

[0091] The acquisition unit is used to acquire time series data in real time, perform data preprocessing on the time series data, and obtain training samples;

[0092] The training unit is used to input the training samples into the neural network for training, and output a convolution kernel carrying time series data information;

[0093] The construction unit is used to build a pruning model, prune the convolution kernel according to the pruning rules, and output the results;

[0094] The comparison unit is used to compare the output result with the time series data. If they are correct, the pruning is successful. If they are wrong, steps S2-S4 are repeated.

[0095] Furthermore, the data processing and analysis system further includes a temporary storage unit;

[0096] The temporary storage unit is used to store the previous training sample.

[0097] The stored previous training sample is used to compare with the current training sample, and the same data information is compared and then input into the pruning model, so that the pruning model uses the same pruning rules for pruning the previous convolution kernel and the current convolution kernel information.

[0098] Reference Figure 3 In the embodiment of the present application, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, memory, a network interface, and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used for data such as training samples. The network interface of the computer device is used to communicate with an external terminal via a network connection. When executed by the processor, the computer program implements a control method for blockchain-based data processing and data analysis methods.

[0099] Those skilled in the art will understand that Figure 3 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0100] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of a blockchain-based data processing and data analysis method are implemented, specifically:

[0101] S1: Acquire time series data in real time, perform data preprocessing on the time series data, and obtain training samples;

[0102] S2: Input the training sample into the neural network for training, and output a convolution kernel carrying time series data information;

[0103] S3: Build a pruning model, prune the convolution kernel according to the pruning rules, and output the results;

[0104] S4: Compare the output result with the time series data. If correct, the pruning is successful. If wrong, repeat steps S2-S4.

[0105] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0106] It should be noted that, in this application, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0107] The above description is only a preferred embodiment of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

[0108] Although the embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

[0109] Of course, the present invention may have many other implementations. Based on this implementation, other implementations obtained by ordinary technicians in this field without any creative work are all within the scope of protection of the present invention.

Claims

1. A data processing and data analysis method based on blockchain, characterized in that: The steps include: S1: acquiring time series data in real time, performing data preprocessing on the time series data, and obtaining training samples; S2: input the training sample into the neural network for training, and output a convolution kernel carrying time series data information; S3: Build a pruning model, prune the convolution kernel according to the pruning rules, and output the results; S4: Compare the output result with the time series data. If correct, the pruning is successful. If wrong, repeat steps S2-S4. After pruning the convolution kernel, the following is satisfied: f(M',D,M)=min(val(M,D)-val(M',D)); Where val(M,D) is the accuracy of the convolution kernel M on the time series data D, val(M',D) is the accuracy of the pruned convolution kernel M' on the time series data D, size(M) is the volume of the convolution kernel M, and size(M') is the volume of the pruned convolution kernel M'.

2. According to claim 1, a data processing and data analysis method based on blockchain is characterized in that: The step of inputting the training samples into the neural network for training and outputting the convolution kernel carrying the time series data information comprises: The training sample is compared with the previous training sample, and if there is the same training sample information, the same convolution kernel is pruned using the same pruning rule.

3. According to a blockchain-based data processing and data analysis method according to claim 1, it is characterized in that: In S3, the step of constructing the pruning model includes: Use time series data for forward propagation and back propagation to obtain the initial model weights of the time series data; Evaluating the importance of each convolution kernel in the time series data to the model based on gradient matrix analysis; The convolution kernel is softly pruned according to the importance evaluation result. The soft pruning operation includes: after pruning some convolution kernels, continuing training and adjusting the weights of the remaining convolution kernels, and continuing pruning after the model converges until the final set pruning rate is completed.

4. A blockchain-based data processing and data analysis method according to claim 3, characterized in that: The step of using the time series data for forward propagation and backward propagation to obtain the initial model weight of the time series data includes: The convolution kernel is subjected to gradient-free separation, which is separated into the directional convolution kernel and the convolution kernel radius. Perform a gradient convolution kernel recovery operation, forward propagate the restored convolution kernel, and then obtain the directional convolution and the gradient of the convolution kernel radius parameter through back propagation; Update the weights according to the gradient.

5. A blockchain-based data processing and data analysis method according to claim 4, characterized in that: Since the weight values ​​of the directional convolution kernel and the convolution kernel radius parameters have changed after the convolution kernel is updated through back propagation, the directional convolution kernel and the convolution kernel radius parameters have lost their original definitions and need to be restored before the next forward propagation: Recover the true weight TW of the convolution kernel i l , separate the weights into convolution kernel radius parameters and directional convolution kernel W i l .

6. The method for data processing and data analysis based on blockchain according to claim 1, characterized in that: The volume of the convolution kernel M after pruning is: The convolution kernel M is pruned after the convolution layer of L layers. size(M)=∑ L＝1 (1-p L )size(L); Where size(L) is the volume of the L-th convolution kernel, p L is the pruning rate of the Lth convolutional layer.

7. A data processing and data analysis system based on blockchain, characterized in that: The steps of a blockchain-based data processing and data analysis method applied to any one of claims 1 to 6, wherein the data processing and analysis system comprises: Acquisition unit: acquires time series data in real time, performs data preprocessing on the time series data, and obtains training samples; Training unit: inputting the training samples into the neural network for training, and outputting a convolution kernel carrying time series data information; Construction unit: construct a pruning model, prune the convolution kernel according to the pruning rules, and output the results; Comparison unit: compares the output result with the time series data. If correct, the pruning is successful. If wrong, repeat steps S2-S4.

8. A blockchain-based data processing and data analysis system according to claim 7, characterized in that: The data processing and analysis system also includes a temporary storage unit; The temporary storage unit is used to store the last training sample.

Images