Multi-modal test data uplink method, system and equipment

Abstract

The invention relates to the technical field of data uplink, in particular to a multi-modal test data uplink method, system and equipment. According to the method, a plurality of change representative positions are obtained according to the change trend of data in an ascending order representative sequence of index representativeness of all dimensions; obtaining a plurality of sub representative sequences of the ascending representative sequence according to the distribution of the change representative positions; constructing corresponding hash buckets by using the performance test data of all dimensions in each sub representative sequence, and obtaining the difference sensitivity of each hash bucket according to the correlation of the performance test data of each same moment position between the corresponding dimensions in each hash bucket; adjusting the hash bucket to obtain a new hash bucket; and according to the index representative distribution of each dimension in the new different hash buckets, obtaining the uplink sequence of each new hash bucket, and performing uplink on the performance test data. According to the method, the Hash buckets are adjusted by analyzing the accurate difference condition between the dimension data, and the test data uplink effect is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data uploading to a chain, and in particular to a method, system and device for uploading multimodal test data to a chain. Background Art

[0002] Multimodality refers to different forms of data or information. By uploading multimodal test data to the blockchain, the data is encrypted and stored in the blockchain network through blockchain technology. When the data in any block is changed, the hash value of the data block on the entire chain will point to an error in the previous block, protecting the data from illegal access and tampering, which can significantly improve the security and credibility of the data. Therefore, it is necessary to upload multimodal test data to the blockchain.

[0003] Taking into account the high repetitiveness and the distribution of many identical types of test data, in the prior art, in the process of uploading test data to the chain, the performance test data is divided into hash buckets through the Locality Sensitive Hashing (LSH) algorithm to build a data index; however, due to the high repetitiveness of the performance test data during processing and the distribution of many similar types, the traditional LSH algorithm has inaccurate recognition of the differential features of the data, and cannot accurately retrieve the data with significant distinguishing features in the local part of the test data, resulting in errors in the hash bucket division and poor chain-up effect. Summary of the invention

[0004] In order to solve the technical problem that the difference feature recognition of data is inaccurate, resulting in errors in hash bucket division and poor chain-up effect, the purpose of the present invention is to provide a method, system and device for chain-up multimodal test data. The technical solution adopted is as follows:

[0005] The present invention proposes a method for uploading multimodal test data to a chain, the method comprising:

[0006] Obtain multi-dimensional performance test data of the test facility at each moment to form a column vector in the sample space;

[0007] According to the distribution of performance test data corresponding to different time positions of each dimension in the sample space, the index representativeness of each dimension is obtained; the ascending representative sequence of the index representativeness of all dimensions is obtained, and according to the change trend of the data in the ascending representative sequence, multiple change representative positions are obtained; according to the distribution of the change representative positions, multiple sub-representative sequences of the ascending representative sequence are obtained;

[0008] The corresponding hash bucket is constructed with the performance test data of all dimensions in each sub-representative sequence, and the difference sensitivity of each hash bucket is obtained according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket; the hash bucket is adjusted according to the difference sensitivity of different hash buckets to obtain a new hash bucket;

[0009] According to the representative distribution of indexes in each dimension in different new hash buckets, the chaining order of each new hash bucket is obtained, and the performance test data is chained.

[0010] Furthermore, the method for obtaining the representativeness of the index includes:

[0011] Obtain the mean of the performance test data corresponding to all time positions in each dimension in the sample space as the data mean;

[0012] Calculate the deviation of the performance test data at each time position in each dimension of the sample space relative to the data mean, and standardize it as the standard deviation value;

[0013] According to the even power of the standard deviation value corresponding to different time positions of each dimension in the sample space, the index representativeness of each dimension is obtained, and the even power of the standard deviation value is positively correlated with the index representativeness.

[0014] Furthermore, the method for obtaining the change representative position includes:

[0015] Obtain the difference sequence of the ascending representative sequence; if the data on the difference sequence is greater than the adjacent data before and after, the middle position of the two adjacent data in the ascending representative sequence of the corresponding data is used as the representative position of the change.

[0016] Furthermore, the method for obtaining the sub-representative sequence includes:

[0017] The ascending representative sequence is divided based on the changing representative position to obtain a plurality of sub-representative sequences.

[0018] Furthermore, the method for obtaining the difference sensitivity includes:

[0019] For any hash bucket, obtain the encrypted hash value of the performance test data of each dimension at each time position;

[0020] For each time position, if the same elements exist in the same order in the encrypted hash values ​​between all dimensions, output the two-dimensional value 1, otherwise, output the two-dimensional value 0;

[0021] According to the number of two-dimensional values ​​1, the number of two-dimensional values ​​0, and the number of dimensions in each hash bucket at all time positions, the difference sensitivity of each hash bucket is obtained. The number of two-dimensional values ​​1 and the number of dimensions are positively correlated with the difference sensitivity, and the number of two-dimensional values ​​0 is negatively correlated with the difference sensitivity.

[0022] Furthermore, the method for obtaining the new hash bucket includes:

[0023] Select all hash buckets whose difference sensitivity is less than the preset sensitivity threshold, merge the corresponding hash buckets to form a new hash bucket;

[0024] For hash buckets whose difference sensitivity is greater than or equal to the preset sensitivity threshold, the corresponding hash bucket is used as a new hash bucket.

[0025] Furthermore, the method for obtaining the chaining order includes:

[0026] For any new hash bucket, obtain the mean of the corresponding index representations of all dimensions as the average representation;

[0027] The dimension whose traction representativeness is closest to the average representativeness among all dimensions is selected, and the corresponding dimension is taken as the typical dimension; all hash buckets are arranged in ascending order according to the index representativeness of the typical dimension, and the corresponding order is taken as the chaining order of each new hash bucket.

[0028] Furthermore, the method for obtaining the differential sequence includes:

[0029] Perform first-order difference processing on the ascending representative sequence to obtain the difference sequence.

[0030] The present invention proposes a multi-modal test data uplink system, the system comprising:

[0031] Test data acquisition module: obtains multi-dimensional performance test data of the test facility at each moment, forming a column vector in the sample space;

[0032] Representativeness analysis module: according to the distribution of performance test data corresponding to different time positions of each dimension in the sample space, obtain the index representativeness of each dimension; obtain the ascending representative sequence of the index representativeness of all dimensions, and according to the change trend of the data in the ascending representative sequence, obtain multiple representative positions of change; according to the distribution of the representative positions of change, obtain multiple sub-representative sequences of the ascending representative sequence;

[0033] Hash bucket optimization module: Construct corresponding hash buckets with the performance test data of all dimensions in each sub-representative sequence, obtain the difference sensitivity of each hash bucket according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket; adjust the hash bucket according to the difference sensitivity of the new different hash buckets to obtain a new hash bucket;

[0034] Data on-chain processing: Based on the representative distribution of indexes in each dimension in different new hash buckets, the on-chain order of each new hash bucket is obtained, and the performance test data is on-chain.

[0035] The present invention also proposes a system for uploading multimodal test data, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any one of the steps of the method for uploading multimodal test data.

[0036] The present invention has the following beneficial effects:

[0037] The present invention takes into account that performance test data of different dimensions at different time positions may have different distribution characteristics. According to the distribution of performance test data of each dimension at different time positions in the sample space, the index representativeness of each dimension is obtained, and the influence and discrimination of each dimension in the overall data are quantified; according to the change trend of data in the ascending representative sequence of the index representativeness of all dimensions, multiple change representative positions are obtained, which reflects the dynamic characteristics and potential laws of the data and identifies the key change points in the data; according to the distribution of the change representative positions, multiple sub-representative sequences of the ascending representative sequence are obtained, so that the data in each sub-representative sequence has higher similarity and consistency, which is convenient for more efficient data storage and retrieval in the hash bucket; the corresponding hash bucket is constructed with the performance test data of all dimensions in each sub-representative sequence, and the difference sensitivity of each hash bucket is obtained according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket, which reflects the similarity and difference degree of the data in the hash bucket; the hash bucket is adjusted to obtain a new hash bucket; according to the distribution of the index representativeness of each dimension in the new different hash buckets, the chaining order of each new hash bucket is obtained, and the performance test data is chained, so as to reduce the dependency conflict and verification complexity between the data and improve the chaining efficiency. The present invention adjusts the hash bucket by analyzing the accurate differences between dimensional data, thereby improving the effect of uploading experimental data to the chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A flowchart of a method for uploading multimodal test data to a chain provided by an embodiment of the present invention;

[0040] Figure 2 A flow chart of a method for obtaining a representative index provided by an embodiment of the present invention;

[0041] Figure 3 A flow chart of a method for obtaining difference sensitivity provided by an embodiment of the present invention;

[0042] Figure 4 A structural block diagram of a multi-modal test data uplink system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the method, system and device for uploading multimodal test data proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0044] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0045] The following is a detailed description of a method, system and device for uploading multimodal test data to a chain provided by the present invention in conjunction with the accompanying drawings.

[0046] See also Figure 1 , which shows a flow chart of a method for uploading multimodal test data to a chain provided by an embodiment of the present invention, and the specific method includes:

[0047] Step S1: Obtain multi-dimensional performance test data of the test facility at each moment to form a column vector in the sample space.

[0048] In the embodiment of the present invention, in the process of uploading multiple types of test data, in order to avoid the high similarity of facilities and the requirement of relatively unified standards, the distribution of high repetitiveness and many identical attributes affects the analysis of data difference characteristics. Therefore, it is necessary to perform characteristic differentiation processing on the performance test data between multiple dimensions; first, for each location, a multi-dimensional sensor with time alignment is deployed, and the multi-dimensional performance test data of the test facility at each moment is obtained to form a column vector in the sample space. Among them, each dimension corresponds to a type of performance test data.

[0049] It should be noted that, in one embodiment of the present invention, the test facility is a highway infrastructure, which can obtain multi-dimensional performance test data such as test pressure data in the mechanical properties test link, settlement depth data in the roadbed settlement test link, temperature state and humidity state inside the highway structure layer in the durability performance test link, physical properties data and traffic flow data.

[0050] It should be noted that in the embodiments of the present invention, due to the influence of factors such as erroneous measurements and equipment failures that may exist during the test process, resulting in poor accuracy of the results, it is necessary to preprocess the data, including: the implementers can obtain the range limits of the performance parameters of the test facilities based on human experience and relevant professional knowledge, and remove the corresponding unqualified data outside the performance parameter range limits; use the nearest neighbor interpolation method to complete the data according to the observation time series to obtain complete performance test data corresponding to each moment position; the specific nearest neighbor interpolation method is a technical means well known to those skilled in the art and will not be elaborated here.

[0051] It should be noted that in order to facilitate subsequent data processing and avoid differences in data units and numerical magnitudes, the multi-dimensional performance test data are standardized to eliminate the influence of dimensions in data calculations.

[0052] Step S2: According to the distribution of performance test data corresponding to different time positions of each dimension in the sample space, obtain the index representativeness of each dimension; obtain an ascending representative sequence of the index representativeness of all dimensions, and according to the change trend of the data in the ascending representative sequence, obtain multiple change representative positions; according to the distribution of the change representative positions, obtain multiple sub-representative sequences of the ascending representative sequence.

[0053] Since the materials and textures of the test facilities themselves are different, the performance test data of the same dimension at different times may show different distribution characteristics, reflecting the dynamic changes within the test facilities. The more obvious the differences between the performance test data in each dimension, the more obvious the performance of the test facilities in the corresponding dimension, and the higher the index representativeness. Therefore, according to the distribution of the performance test data of each dimension at different times in the sample space, the index representativeness of each dimension is obtained.

[0054] Preferably, in one embodiment of the present invention, the method for obtaining the representative index is as follows: Figure 2 , which shows a flow chart of a method for obtaining index representativeness, including:

[0055] Step S201: Obtain the mean of the performance test data corresponding to all time positions in each dimension in the sample space as the data mean.

[0056] In order to comprehensively reflect the overall situation of the performance test data of each dimension in the sample space, the performance test data corresponding to all time positions of each dimension are quantified by averaging to obtain the overall level of the performance test data corresponding to each dimension, which is convenient for subsequent comparison.

[0057] Step S202: Calculate the deviation of the performance test data at each time position in each dimension in the sample space relative to the data mean, and standardize it as the standard deviation value.

[0058] It should be noted that, in one embodiment of the present invention, standardization of the deviation of the performance test data at each time position relative to the data mean can be achieved by calculating the deviation and the standard deviation of the performance test data corresponding to all time positions in each dimension. The closer the standard deviation value is to 0, the closer the performance test data is to the data mean, and the more concentrated the data distribution is.

[0059] Step S203: Obtain the index representativeness of each dimension according to the even power of the standard deviation value of each dimension in the sample space corresponding to different time positions. The even power of the standard deviation value is positively correlated with the index representativeness.

[0060] Among them, the positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases, that is, the larger the even power of the standard deviation value, the greater the deviation between the performance test data at each moment and the data mean, the more uneven the data distribution, the more likely abnormal situations will occur, and the more representative the index is.

[0061] In one embodiment of the present invention, the index representative formula is expressed as:

[0062]

[0063] Among them, Y d represents the index representativeness of the dth dimension in the sample space; p d,i Represents the performance test data of the dth dimension corresponding to the i-th moment position in the sample space; It represents the mean of the performance test data at all time positions corresponding to the dth dimension in the sample space, that is, the data mean; Represents the standard deviation of the performance test data at all time positions corresponding to the dth dimension in the sample space; n d Represents the number of time positions corresponding to the dth dimension in the sample space.

[0064] In the representative formula of the index, It means calculating the ratio between the deviation of the performance test data at the i-th moment position in the d-th dimension in the sample space relative to the data mean and the standard deviation, that is, standardizing the deviation to obtain the standard deviation value; It means to accumulate the fourth power of the standard deviation value of all time positions corresponding to the dth dimension in the sample space. The larger the accumulated value, the greater the difference between the performance test data and the data mean, the more uneven the data distribution, the greater the discrimination of the data, and the more representative the index.

[0065] By sorting the index representativeness of all dimensions in ascending order from small to large, we can fully understand the impact of each dimension on the facility performance and reflect the changing rules and characteristics of the facility performance in different dimensions; obtain the ascending representative sequence of the index representativeness of all dimensions, and obtain multiple representative positions of change according to the changing trend of the data in the ascending representative sequence.

[0066] Preferably, in one embodiment of the present invention, the method for obtaining the change representative position includes:

[0067] Obtain the difference sequence of the ascending representative sequence; if the data on the difference sequence is greater than the adjacent data before and after, the middle position of the two adjacent data in the ascending representative sequence of the corresponding data is used as the representative position of the change.

[0068] The difference sequence can remove the unsteady trend in the ascending representative sequence, facilitate the observation and analysis of the changing rules and characteristics in the sequence, and highlight the changing part of the sequence, that is, the part that is different from the sequence trend; it should be noted that, in the embodiment of the present invention, the ascending representative sequence is subjected to first-order difference processing to obtain the difference sequence; wherein, the difference processing is a technical means well known to those skilled in the art and will not be elaborated here.

[0069] Since the representative position of change is usually the inflection point or important turning point of the data change trend, it reflects the significant change in the data performance characteristics between dimensions. Dividing the ascending representative sequence into sub-representative sequences through the distribution of the representative position of change can highlight the similar changes between different dimensions. Therefore, according to the distribution of the representative position of change, multiple sub-representative sequences of the ascending representative sequence are obtained.

[0070] Preferably, in one embodiment of the present invention, the method for obtaining the sub-representative sequence includes:

[0071] The ascending representative sequence is divided based on the change representative position to obtain multiple sub-representative sequences. The sub-representative sequences illustrate the concentration of performance test data of different dimensions in the test facility on the performance of the facility.

[0072] As an example, if there is an ascending representative sequence of (1, 2, 5, 6, 8, 9) and a differential sequence of (1, 2, 1, 2, 1), where the second and fourth data in the differential sequence are greater than the adjacent data before and after, the corresponding data in the ascending representative sequence are 2 and 5, 6 and 8, and the middle position of the two groups of data is the changing representative position, and the sub-representative sequences obtained by division are (1, 2), (5, 6), and (8, 9).

[0073] Step S3: Construct corresponding hash buckets with the performance test data of all dimensions in each sub-representative sequence, and obtain the difference sensitivity of each hash bucket according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket; adjust the hash bucket according to the difference sensitivity of different hash buckets to obtain a new hash bucket.

[0074] Each sub-representative sequence contains dimensions with smaller changing trends. Constructing corresponding hash buckets with the performance test data of all dimensions in each sub-representative sequence is helpful for centralized management and analysis of performance test data. Correlation can be used to evaluate whether there is a correlation between performance test data of different dimensions at the same time and the degree of correlation. The higher the degree of correlation, when the performance test data of one dimension changes, the performance data of other dimensions is likely to change accordingly, indicating that the difference sensitivity between dimensions is high. Therefore, the difference sensitivity of each hash bucket is obtained based on the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket.

[0075] Preferably, in one embodiment of the present invention, the method for obtaining the difference sensitivity can be found in Figure 3 , which shows a flow chart of a method for obtaining difference sensitivity, including:

[0076] Step S301: For any hash bucket, obtain the encrypted hash value of the performance test data of each dimension at each time position.

[0077] It should be noted that, in one embodiment of the present invention, the performance test data is hashed using the SHA-512 algorithm to obtain an encrypted hash value corresponding to the data; wherein the SHA-512 algorithm is a secure hash algorithm that generates a 512-bit hash value, that is, the encrypted hash value is a binary number with a length of 512 bits; the specific SHA-512 algorithm is a technical means well known to those skilled in the art and will not be elaborated here.

[0078] Step S302: For each time position, if the same elements exist in each same order in the encrypted hash values ​​between all dimensions, a two-dimensional value of 1 is output; otherwise, a two-dimensional value of 0 is output.

[0079] By comparing the numerical values ​​of encrypted hash values ​​between dimensions, we can analyze the similarities between the data, indicating that there is some kind of correlation or common characteristics between the data in these dimensions.

[0080] As an example, if the dimension is 3, for each time position, the encrypted hash values ​​of data of different dimensions are: 3acdcd0a3894f0fcb837, 3ardcd0a3724f0d8b837, 3aedcd0a3894f06cq837, and the corresponding output result is 1101111100111000111.

[0081] Step S303: According to the number of two-dimensional values ​​1, the number of two-dimensional values ​​0, and the number of dimensions in each hash bucket at all time positions, the difference sensitivity of each hash bucket is obtained. The number of two-dimensional values ​​1 and the number of dimensions are positively correlated with the difference sensitivity, and the number of two-dimensional values ​​0 is negatively correlated with the difference sensitivity.

[0082] Among them, the positive correlation indicates that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases, that is, the larger the number of two-dimensional values ​​1, the larger the number of dimensions, indicating that the data similarity between multiple dimensions is greater and the difference is smaller, which makes it more sensitive to data with differences and the greater the difference sensitivity; the negative correlation indicates that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases, that is, the larger the number of two-dimensional values ​​0, the greater the difference in data between dimensions, which makes it less sensitive to data with differences and the smaller the difference sensitivity.

[0083] In one embodiment of the present invention, the formula of difference sensitivity is expressed as:

[0084]

[0085] Among them, M T represents the difference sensitivity of the Tth hash bucket; n T,1 Indicates the number of two-dimensional values ​​1 in the Tth hash bucket; n T,0 Indicates the number of two-dimensional values ​​0 in the Tth hash bucket; n T,d Indicates the number of corresponding dimensions in the Tth hash bucket; max(n T′,d ) represents the maximum number of dimensions in hash buckets other than the Tth hash bucket; sigmoid() represents the logistic function.

[0086] In the formula for differential sensitivity, n T,0 +0.01 for n T,0 Add 0.01 to avoid the formula being meaningless when the value is 0; The larger the number of two-dimensional values ​​1 in the Tth hash bucket, the smaller the number of two-dimensional values ​​0, the smaller the ratio, the more similar the data between dimensions, the more sensitive to differences, and the greater the difference sensitivity; It means calculating the ratio between the number of corresponding dimensions in the Tth hash bucket and the maximum number of dimensions in other hash buckets except the Tth hash bucket. The larger the ratio, the larger the number of corresponding dimensions in the Tth hash bucket, which means that there are more data associations between dimensions. If there are differences, more dimensions are affected and the greater the sensitivity of the differences.

[0087] By adjusting the hash buckets based on the difference sensitivity, the amount of calculation is reduced, and the data dimension that can fully represent the actual situation during the experiment can occupy the main part of the on-chain contract. The hash buckets are adjusted according to the difference sensitivity of different hash buckets to obtain new hash buckets.

[0088] Preferably, in one embodiment of the present invention, the method for obtaining a new hash bucket includes:

[0089] Select all hash buckets whose difference sensitivity is less than the preset sensitivity threshold, merge the corresponding hash buckets to form a new hash bucket;

[0090] For hash buckets whose difference sensitivity is greater than or equal to the preset sensitivity threshold, the corresponding hash bucket is used as a new hash bucket.

[0091] It should be noted that, in one embodiment of the present invention, the preset sensitivity threshold is set to 0.1; in other embodiments of the present invention, the preset sensitivity threshold may be set according to specific circumstances, which is not limited or elaborated herein.

[0092] Step S4: According to the representative distribution of the index of each dimension in the new different hash buckets, the chaining order of each new hash bucket is obtained, and the performance test data is chained.

[0093] During the data upload process, there may be dependencies between data items. Determining the upload order according to the representative distribution of the index can ensure that the dependent items are uploaded before the dependent items, avoiding data inconsistency and facilitating subsequent data retrieval. Therefore, according to the representative distribution of the index of each dimension in the new different hash buckets, the upload order of each new hash bucket is obtained, and the performance test data is uploaded to the chain.

[0094] Preferably, in one embodiment of the present invention, the method for obtaining the chaining order includes:

[0095] For any new hash bucket, obtain the mean of the corresponding index representativeness of all dimensions as the average representativeness; select the traction representativeness of all dimensions that is closest to the average representativeness, and use the corresponding dimension as the typical dimension; arrange all hash buckets in ascending order according to the index representativeness of the typical dimension, and use the corresponding order as the chaining order of each new hash bucket.

[0096] It should be noted that, in another embodiment of the present invention, after obtaining the chain-up order of each new hash bucket, the data in each new hash bucket is stored in a linked list in turn; a smart contract that defines the data processing logic and rules is constructed, and the result after the linked list storage is passed to the contract by calling the preset deployed smart contract. If a transaction containing detailed information of the contract call is generated after the contract is executed, it needs to be verified and confirmed by the nodes in the blockchain network; if it complies with the consensus mechanism and rules of the blockchain, the block containing the transaction data will be added to the blockchain to realize the chain-up of the data; through the introduction of smart contracts, the transparency and traceability of the data are increased, making the data processing process more fair and open.

[0097] In summary, the present invention obtains the index representativeness of each dimension according to the distribution of the performance test data corresponding to different time positions of each dimension in the sample space; obtains multiple change representative positions according to the change trend of the data in the ascending representative sequence of the index representativeness of all dimensions; obtains multiple sub-representative sequences of the ascending representative sequence according to the distribution of the change representative positions; constructs corresponding hash buckets with the performance test data of all dimensions in each sub-representative sequence, obtains the difference sensitivity of each hash bucket according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket; adjusts the hash bucket to obtain a new hash bucket; obtains the chaining order of each new hash bucket according to the distribution of the index representativeness of each dimension in the new different hash buckets, and chains the performance test data. The present invention improves the effect of chaining the test data by analyzing the accurate differences between the dimensional data.

[0098] Based on the same application concept as the method for uploading multimodal test data provided in the embodiment of the present invention, this embodiment also proposes a system for uploading multimodal test data, such as Figure 4 As shown, the system includes: a test data acquisition module 401, a representative analysis module 402, a hash bucket optimization module 403 and a data chain processing 404:

[0099] Test data acquisition module 401: acquires multi-dimensional performance test data of the test facility at each moment, forming a column vector in the sample space;

[0100] Representativeness analysis module 402: according to the distribution of performance test data corresponding to different time positions of each dimension in the sample space, obtain the index representativeness of each dimension; obtain an ascending representative sequence of the index representativeness of all dimensions, and according to the change trend of the data in the ascending representative sequence, obtain multiple change representative positions; according to the distribution of the change representative positions, obtain multiple sub-representative sequences of the ascending representative sequence;

[0101] Hash bucket optimization module 403: construct corresponding hash buckets with the performance test data of all dimensions in each sub-representative sequence, obtain the difference sensitivity of each hash bucket according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket; adjust the hash bucket according to the difference sensitivity of different hash buckets to obtain a new hash bucket;

[0102] Data chain processing 404: According to the representative distribution of the index of each dimension in the new different hash buckets, the chain order of each new hash bucket is obtained, and the performance test data is chained.

[0103] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above.

[0104] It should be understood that the system provided in this embodiment is used to execute the above-mentioned method of uploading multimodal test data to the chain, and therefore has the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0105] The present invention also proposes a multimodal test data uplink device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any one of the steps of a multimodal test data uplink method.

[0106] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0107] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for uploading multimodal test data to a chain, characterized in that: The method comprises: Obtain multi-dimensional performance test data of the test facility at each moment to form a column vector in the sample space; According to the distribution of performance test data corresponding to different time positions of each dimension in the sample space, the index representativeness of each dimension is obtained; the ascending representative sequence of the index representativeness of all dimensions is obtained, and according to the change trend of the data in the ascending representative sequence, multiple change representative positions are obtained; according to the distribution of the change representative positions, multiple sub-representative sequences of the ascending representative sequence are obtained; The corresponding hash bucket is constructed with the performance test data of all dimensions in each sub-representative sequence, and the difference sensitivity of each hash bucket is obtained according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket; the hash bucket is adjusted according to the difference sensitivity of different hash buckets to obtain a new hash bucket; According to the representative distribution of indexes in each dimension in different new hash buckets, the chaining order of each new hash bucket is obtained, and the performance test data is chained.

2. A method for uploading multimodal test data to a chain according to claim 1, characterized in that: The method for obtaining the representativeness of the index includes: Obtain the mean of the performance test data corresponding to all time positions in each dimension in the sample space as the data mean; Calculate the deviation of the performance test data at each time position in each dimension of the sample space relative to the data mean, and standardize it as the standard deviation value; According to the even power of the standard deviation value corresponding to different time positions of each dimension in the sample space, the index representativeness of each dimension is obtained, and the even power of the standard deviation value is positively correlated with the index representativeness.

3. The method for uploading multimodal test data to a chain according to claim 1, characterized in that: The method for obtaining the change representative position includes: Obtain the difference sequence of the ascending representative sequence; if the data on the difference sequence is greater than the adjacent data before and after, the middle position of the two adjacent data in the ascending representative sequence of the corresponding data is used as the representative position of the change.

4. The method for uploading multimodal test data to a chain according to claim 1, characterized in that: The method for obtaining the sub-representative sequence includes: The ascending representative sequence is divided based on the changing representative position to obtain a plurality of sub-representative sequences.

5. The method for uploading multimodal test data to a chain according to claim 1, characterized in that: The method for obtaining the difference sensitivity includes: For any hash bucket, obtain the encrypted hash value of the performance test data of each dimension at each time position; For each time position, if the same elements exist in the same order in the encrypted hash values ​​between all dimensions, output the two-dimensional value 1, otherwise, output the two-dimensional value 0; According to the number of two-dimensional values ​​1, the number of two-dimensional values ​​0, and the number of dimensions in each hash bucket at all time positions, the difference sensitivity of each hash bucket is obtained. The number of two-dimensional values ​​1 and the number of dimensions are positively correlated with the difference sensitivity, and the number of two-dimensional values ​​0 is negatively correlated with the difference sensitivity.

6. The method for uploading multimodal test data to a chain according to claim 1, characterized in that: The method for obtaining the new hash bucket includes: Select all hash buckets whose difference sensitivity is less than the preset sensitivity threshold, merge the corresponding hash buckets to form a new hash bucket; For hash buckets whose difference sensitivity is greater than or equal to the preset sensitivity threshold, the corresponding hash bucket is used as a new hash bucket.

7. The method for uploading multimodal test data to a chain according to claim 1, characterized in that: The method for obtaining the chaining order includes: For any new hash bucket, obtain the mean of the corresponding index representations of all dimensions as the average representation; The dimension whose traction representativeness is closest to the average representativeness among all dimensions is selected, and the corresponding dimension is taken as the typical dimension; all hash buckets are arranged in ascending order according to the index representativeness of the typical dimension, and the corresponding order is taken as the chaining order of each new hash bucket.

8. The method for uploading multimodal test data to a chain according to claim 3, characterized in that: The method for obtaining the differential sequence includes: Perform first-order difference processing on the ascending representative sequence to obtain the difference sequence.

9. A system for uploading multimodal test data, characterized in that: The system comprises: Test data acquisition module: obtains multi-dimensional performance test data of the test facility at each moment, forming a column vector in the sample space; Representativeness analysis module: according to the distribution of performance test data corresponding to different time positions of each dimension in the sample space, obtain the index representativeness of each dimension; obtain the ascending representative sequence of the index representativeness of all dimensions, and according to the change trend of the data in the ascending representative sequence, obtain multiple representative positions of change; according to the distribution of the representative positions of change, obtain multiple sub-representative sequences of the ascending representative sequence; Hash bucket optimization module: Construct corresponding hash buckets with the performance test data of all dimensions in each sub-representative sequence, obtain the difference sensitivity of each hash bucket according to the correlation of the performance test data of each same time position between the corresponding dimensions in each hash bucket; adjust the hash bucket according to the difference sensitivity of different hash buckets to obtain a new hash bucket; Data on-chain processing: Based on the representative distribution of indexes in each dimension in different new hash buckets, the on-chain order of each new hash bucket is obtained, and the performance test data is on-chain.

10. A device for uploading multimodal test data, the device 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 steps of the method for uploading multimodal test data to a link as described in any one of claims 1 to 8 are implemented.

Images