Data processing method, device and electronic device

Abstract

The present application discloses a data processing method, apparatus and electronic device. The method includes: determining a target impact parameter; determining a target correlation coefficient of the target impact parameter and a target weight corresponding to the target correlation coefficient based on the correlation relationship between the target impact parameter and a target parameter; and determining a delay impact value of the target impact parameter on the target parameter based on the target correlation coefficient and the target weight.

Description

Technical Field

[0001] This application relates to the technical field of time-delay control, and particularly to a data processing method, apparatus, and electronic device. Background Art

[0002] During process control, most controlled objects exhibit slow time-varying and pure-delay processes in terms of dynamic characteristics. For example, the delay phenomenon between the input signal or control signal of the system and the output signal of the system under its action; for example, if there is a time delay in measurement, it will cause the controller to be unable to detect the change of the controlled quantity in time; for example, if there is a time delay in control, the effect cannot be generated in time. And the greater the time delay, the more obvious the above phenomena are, and the control difficulty is also significantly increased.

[0003] Therefore, there is an urgent need for a technical solution that can accurately detect time delays. Summary of the Invention

[0004] In view of this, this application provides a data processing method, apparatus, and electronic device, as follows:

[0005] A data processing method includes:

[0006] Determine the target influencing parameter;

[0007] Based on the correlation relationship between the target influencing parameter and the target parameter, determine the target correlation coefficient of the target influencing parameter and the target weight corresponding to the target correlation coefficient;

[0008] Based on the target correlation coefficient and the target weight, determine the time-delay influence value of the target influencing parameter on the target parameter.

[0009] In the above method, preferably, the step of determining the time-delay influence value of the target influencing parameter on the target parameter based on the target correlation coefficient and the target weight includes:

[0010] Based on the target correlation coefficient and the target weight, obtain the influence effect value of the target influencing parameter on the target parameter;

[0011] Based on the influence effect value, determine the time-delay influence value of the target influencing parameter on the target parameter.

[0012] In the above method, preferably, the step of determining the time-delay influence value of the target influencing parameter on the target parameter based on the influence effect value includes:

[0013] Based on the influence effect value, determine the target influence effect value;

[0014] Based on the target influence effect value, determine the time-delay influence value of the target influencing parameter on the target parameter.

[0015] Preferably, for the above method, based on the target correlation coefficient and the target weight, obtaining the influence effect value of the target influence parameter on the target parameter includes:

[0016] In at least one first sub-period within the first period, based on the target correlation coefficient and the target weight, obtaining at least one influence effect value of each target influence parameter on the target parameter.

[0017] Preferably, for the above method, determining the time-delay influence value of the target influence parameter on the target parameter based on the target influence effect value includes:

[0018] Based on the interval between the first time corresponding to obtaining the target influence effect value and the start time of the first period, determining the time-delay influence value corresponding to the first period.

[0019] Preferably, for the above method, determining the target influence parameter includes:

[0020] Obtaining candidate influence parameters;

[0021] Performing a first process on the candidate influence parameters to obtain an influence parameter sequence;

[0022] Determining the target influence parameter based on the influence parameter sequence.

[0023] Preferably, for the above method, the target weight is obtained by the following method:

[0024] Based on a normalization algorithm, processing the target correlation coefficient to determine the target weight corresponding to the target correlation coefficient.

[0025] Preferably, for the above method, before determining the target influence parameter, the method further includes:

[0026] Monitoring whether an update condition is satisfied; the update condition is related to a second period; the second period is greater than the first period;

[0027] When the update condition is satisfied, performing the step of determining the target influence parameter.

[0028] A data processing device includes:

[0029] A parameter determination unit for determining a target influence parameter;

[0030] A target determination unit for determining the target correlation coefficient of the target influence parameter and the target weight corresponding to the target correlation coefficient based on the association relationship between the target influence parameter and the target parameter;

[0031] A time delay determination unit, configured to determine a time delay influence value of the target influence parameter on the target parameter based on the target correlation coefficient and the target weight.

[0032] An electronic device, comprising:

[0033] A memory, configured to store a computer program and data generated by running the computer program;

[0034] A processor, configured to execute the computer program to implement: determining a target influence parameter; determining a target correlation coefficient of the target influence parameter and a target weight corresponding to the target correlation coefficient based on an association relationship between the target influence parameter and a target parameter; determining a time delay influence value of the target influence parameter on the target parameter based on the target correlation coefficient and the target weight.

[0035] As can be seen from the above technical solutions, in a data processing method, apparatus, and electronic device disclosed in this application, after determining a target influence parameter, a target correlation coefficient of the target influence parameter and a target weight corresponding to the target correlation coefficient can be determined according to the association relationship between the target influence parameter and the target parameter, and then a time delay influence value of the target influence parameter on the target parameter can be determined according to the target correlation coefficient and the target weight. It can be seen that in this application, the target correlation coefficient of the target influence parameter and the corresponding weight are used to determine the time delay influence value, so as to facilitate process control. Description of the Drawings

[0036] In order to more clearly illustrate the technical solutions of the embodiments of this application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of this application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0037] Figure 1 A flowchart of a data processing method provided in Embodiment 1 of this application;

[0038] Figure 2 A partial flowchart of a data processing method provided in Embodiment 1 of this application;

[0039] Figure 3 Another partial flowchart of a data processing method provided in Embodiment 1 of this application;

[0040] Figure 4 Another partial flowchart of a data processing method provided in Embodiment 1 of this application;

[0041] Figure 5 Another flowchart of a data processing method provided in Embodiment 1 of this application;

[0042] Figure 6 It is a schematic structural diagram of a data processing device provided in the second embodiment of the present application;

[0043] Figure 7 It is a schematic structural diagram of an electronic device provided in the third embodiment of the present application;

[0044] Figure 8 It is an example flowchart of the present application applicable to obtaining the time-delay effect of the outlet moisture content. Specific embodiments

[0045] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0046] Refer to Figure 1 As shown, it is a flowchart of the implementation of a data processing method provided in the first embodiment of the present application. This method can be applied to an electronic device capable of performing a data processing method, such as a computer or a server. The technical solution in this embodiment is mainly used to accurately detect the time-delay influence value of the target influence parameter on the target parameter.

[0047] Specifically, the method in this embodiment may include the following steps:

[0048] Step 101: Determine the target influence parameter.

[0049] Among them, there may be one or more target influence parameters. The target influence parameter is an influence parameter that affects the target parameter. Taking the target parameter as the outlet moisture content of the cut tobacco conditioning machine as an example, the target influence parameters may include environmental temperature, ambient humidity, etc.

[0050] Step 102: Based on the correlation relationship between the target influence parameter and the target parameter, determine the target correlation coefficient of the target influence parameter and the target weight corresponding to the target correlation coefficient.

[0051] Among them, the target correlation coefficient of the target influence parameter may include Pearson correlation coefficient, Spearman rank correlation coefficient, Feature Imports importance, high-order partial correlation coefficient, etc.

[0052] Specifically, in this embodiment, the actual values of the target impact parameter and the target parameter in the process control can be collected to obtain the actual values of the target impact parameter and the target parameter within a certain time period, such as 90 days. The correlation between the target impact parameter and the target parameter is characterized by the corresponding relationship between the actual value of the target impact parameter and the actual value of the target parameter. Then, the actual value of the target impact parameter and the actual value of the target parameter are used for calculation to obtain the target correlation coefficient of the target impact parameter and the target weight corresponding to the target correlation coefficient.

[0053] Among them, the target correlation coefficient quantitatively characterizes the correlation between the target impact parameter and the target parameter in the process control, and the target weight corresponding to the target correlation coefficient quantitatively characterizes the importance of the time delay generated on the target parameter due to the change of the target impact parameter in the process control.

[0054] Step 103: Based on the target correlation coefficient and the target weight, determine the time delay impact value of the target impact parameter on the target parameter.

[0055] Among them, the time delay impact value of the target impact parameter on the target parameter indicates the time delay value generated on the target parameter due to the change of the target impact parameter in the process control. For example, in the process control, at the first moment, the target impact parameter changes, and at the second moment, the target parameter changes. The time difference between the second moment and the first moment is the time delay value generated on the target parameter due to the change of the target impact parameter.

[0056] Specifically, in this embodiment, the target correlation coefficient and the target weight can be calculated to obtain the time delay impact value of the target impact parameter on the target parameter.

[0057] It can be seen from the above technical solution that in a data processing method provided in Embodiment 1 of the present application, after determining the target impact parameter, the target correlation coefficient of the target impact parameter and the target weight corresponding to the target correlation coefficient can be determined according to the correlation between the target impact parameter and the target parameter, and then the time delay impact value of the target impact parameter on the target parameter can be determined according to the target correlation coefficient and the target weight. It can be seen that in this embodiment, the target correlation coefficient of the target impact parameter and the corresponding weight are used to determine the time delay impact value, so as to facilitate the realization of process control.

[0058] In one implementation, when determining the time delay impact value of the target impact parameter on the target parameter in step 103 based on the target correlation coefficient and the target weight, it can be implemented in the following manner, such as Figure 2 shown in

[0059] Step 201: Based on the target correlation coefficient and the target weight, obtain the impact effect value of the target impact parameter on the target parameter.

[0060] Among them, the influence effect value can be understood as the delay effect value of the target influence parameter on the target parameter, and this delay effect value can characterize the degree of time-delay influence of the target influence parameter on the target parameter.

[0061] Specifically, in this embodiment, the target weight can be used to perform weighted summation on the target correlation coefficient to obtain the influence effect value of the target influence parameter on the target parameter.

[0062] Step 202: Based on the influence effect value, determine the time-delay influence value of the target influence parameter on the target parameter.

[0063] Specifically, in this embodiment, data statistics can be performed on the influence effect values within a certain time period, and then the time-delay influence value of the target influence parameter on the target parameter can be obtained based on the statistical results.

[0064] In one implementation, step 202 can be implemented in the following manner, as Figure 3 shown in

[0065] Step 301: Based on the influence effect value, determine the target influence effect value.

[0066] For example, when there is one influence effect value, this influence effect value is determined as the target influence effect value; when there are multiple influence effect values, any one of these multiple influence effect values can be randomly selected as the target influence effect value, or the median value among these multiple influence effect values can be selected as the target influence effect value, or the maximum value among these multiple influence effect values can be selected as the target influence effect value, or the average value of these multiple influence effect values can be calculated and the obtained average value can be used as the target influence effect value.

[0067] Specifically, in this embodiment, within at least one first sub-period of the first period, based on the target correlation coefficient and the target weight, at least one influence effect value of each target influence parameter on the target parameter can be obtained. That is to say, in this embodiment, the influence effect values can be statistically analyzed according to the first sub-period within the first period, and thus the influence effect values corresponding to each first sub-period within the first period can be obtained.

[0068] For example, it is preset that the first period is 600 seconds and the first sub-period is 10 seconds. Within 600 seconds, for each 10 seconds, based on the target correlation coefficient and the target weight of each above-mentioned target influence parameter, a time-delay effect value of the ambient temperature on the moisture content of the cut tobacco at the outlet, that is, the influence effect value, is obtained, and then 60 influence effect values are obtained. Based on this, the maximum influence effect value among these 60 influence effect values is found as the target influence effect value.

[0069] Step 302: Determine the time-delay influence value of the target influence parameter on the target parameter based on the target influence effect value.

[0070] Among them, in this embodiment, the time-delay influence value of the target influence parameter on the target parameter can be determined based on the moment corresponding to the target influence effect value.

[0071] Specifically, the target influence effect value is the target influence effect value obtained by statistically analyzing the influence effect value according to the first sub-period within the first period. Based on this, in this embodiment, the time-delay influence value of the target influence parameter on the target parameter can be determined according to the moment corresponding to the first sub-period of the target influence effect value within the first period. Among them, the moment corresponding to the first sub-period of the target influence effect value within the first period is the first time when the target influence effect value is obtained. Based on this, in this embodiment, the time-delay influence value of the target influence parameter on the target parameter corresponding to the first period can be determined based on the interval between the first time and the start time of the first period.

[0072] For example, in this embodiment, the largest influence effect value is found from 60 influence effect values as the target influence effect value, and then the moment corresponding to obtaining this target influence effect value is found, such as the moment of the 8th 10 seconds. Based on this, 80 seconds is determined as the time-delay influence value (i.e., time lag) of the environmental temperature on the moisture content of the cut tobacco at the outlet within the 600-second period.

[0073] In one implementation manner, when determining the target influence parameter in step 101, it can be specifically implemented in the following manner, such as Figure 4 as shown in

[0074] Step 401: Obtain candidate influence parameters.

[0075] Among them, in this embodiment, hundreds or even thousands of process production parameters can be obtained. These process production parameters are the initial influence parameters affecting the target parameter. Then, for these process production parameters, at least one correlation coefficient between each process production parameter and the target parameter is calculated. For example, the weighted sum of the correlation coefficients between each process production parameter and the target parameter is calculated to obtain the correlation value between each process production parameter and the target parameter. For example, the weighted sum of the Pearson correlation coefficient, Spearman rank correlation coefficient, etc. between each process production parameter and the target parameter is calculated to obtain the correlation value. Then, the process production parameters with the correlation value greater than or equal to the correlation threshold or the top N process production parameters sorted from large to small in terms of the correlation value are used as candidate influence parameters. N is a positive integer greater than or equal to 1.

[0076] Step 402: Perform a first processing on the candidate influence parameters to obtain an influence parameter sequence.

[0077] Specifically, the first process is to screen parameters in a corresponding proportion using the Feature Importances mechanism.

[0078] In this embodiment, in the candidate impact parameters, a model based on the Feature Importances mechanism can be used to select parameters that meet the screening conditions. The screening condition can be that the feature importance values output by the model based on the Feature Importances mechanism corresponding to the candidate impact parameters are sorted from largest to smallest and are in the top M%. M is a positive integer greater than or equal to 1. The feature importance represents the contribution amount of the candidate impact parameter to the time delay impact value of the target parameter. Based on this, in this embodiment, an impact parameter sequence containing multiple candidate impact parameters can be obtained.

[0079] It should be noted that the first process also includes: performing autocorrelation tests on the candidate impact parameters to eliminate redundant parameters in the candidate impact parameters.

[0080] For example, in this embodiment, first, a weight fusion coefficient model is used to perform weighted summation on each process production parameter and the target parameter in terms of Pearson correlation coefficient, Spearman rank correlation coefficient, etc. to obtain a correlation value. Then, the process production parameters with a correlation value greater than or equal to the correlation threshold or the top N in the order of decreasing correlation value are used as candidate impact parameters; then, autocorrelation tests are performed on the candidate impact parameters to eliminate redundant parameters in the candidate impact parameters; finally, the remaining candidate impact parameters are screened using a model based on the Feature Importances mechanism to obtain an impact parameter sequence.

[0081] Step 403: Determine the target impact parameter based on the impact parameter sequence.

[0082] Among them, in this embodiment, the candidate impact parameters in the impact parameter sequence can be determined as the target impact parameters. There is one or more target impact parameters.

[0083] Based on the above implementation solution, in this embodiment, the target correlation coefficient can be processed based on the normalization algorithm to determine the target weight corresponding to the target correlation coefficient.

[0084] For example, after obtaining the target impact parameters: Pearson correlation coefficient, Spearman rank correlation coefficient, Feature Imports importance, high-order partial correlation coefficient, etc., these correlation coefficients are normalized, and the normalized values are the target weights of the corresponding correlation coefficients.

[0085] In one implementation, before step 101 in this embodiment, the following steps may also be included, as Figure 5 shown in

[0086] Step 100: Monitor whether the update condition is satisfied. If the update condition is satisfied, execute Step 101. If the update condition is not satisfied, then continue to execute Step 100, that is, continue to monitor whether the update condition is satisfied until the update condition is satisfied and then execute Step 101.

[0087] Among them, the update condition is related to the second period. Here, the second period is greater than the first period.

[0088] That is to say, in this embodiment, the update period of the time delay influence value of the target parameter by the update target influence parameter is set, that is, the second period. Every time the duration of the second period elapses, the time delay influence value is updated once.

[0089] For example, in this embodiment, the second period is set to one day. Based on this, the technical solution in this embodiment is executed once a day, that is, the time delay influence value of each target influence parameter on the target parameter is updated every day.

[0090] Reference Figure 6 , which is a schematic structural diagram of a data processing device provided in Embodiment 2 of the present application. This device can be configured in an electronic device capable of performing a data processing method, such as a computer or a server. The technical solution in this embodiment is mainly used to accurately detect the time delay influence value of the target influence parameter on the target parameter.

[0091] Specifically, the device in this embodiment may include the following units:

[0092] A parameter determination unit 601, configured to determine a target influence parameter;

[0093] A target determination unit 602, configured to determine a target correlation coefficient of the target influence parameter and a target weight corresponding to the target correlation coefficient based on the correlation relationship between the target influence parameter and the target parameter;

[0094] A time delay determination unit 603, configured to determine a time delay influence value of the target influence parameter on the target parameter based on the target correlation coefficient and the target weight.

[0095] It can be seen from the above technical solution that in a data processing device provided in Embodiment 2 of the present application, after determining the target influence parameter, the target correlation coefficient of the target influence parameter and the target weight corresponding to the target correlation coefficient can be determined according to the correlation relationship between the target influence parameter and the target parameter, and then the time delay influence value of the target influence parameter on the target parameter can be determined according to the target correlation coefficient and the target weight. It can be seen that in this embodiment, the target correlation coefficient of the target influence parameter and the corresponding weight are used to determine the time delay influence value, so as to facilitate process control.

[0096] In one implementation, the latency determination unit 603 is specifically configured to: obtain the influence effect value of the target influence parameter on the target parameter based on the target correlation coefficient and the target weight; and determine the latency influence value of the target influence parameter on the target parameter based on the influence effect value.

[0097] Further, when the latency determination unit 603 determines the latency influence value of the target influence parameter on the target parameter based on the influence effect value, it is specifically configured to: determine a target influence effect value based on the influence effect value; and determine the latency influence value of the target influence parameter on the target parameter based on the target influence effect value.

[0098] In one implementation, when the latency determination unit 603 obtains the influence effect value of the target influence parameter on the target parameter based on the target correlation coefficient and the target weight, it is specifically configured to: obtain at least one influence effect value of each target influence parameter on the target parameter based on the target correlation coefficient and the target weight in at least one first sub - period within the first period.

[0099] In one implementation, when the latency determination unit 603 determines the latency influence value of the target influence parameter on the target parameter based on the target influence effect value, it is specifically configured to: determine the latency influence value corresponding to the first period based on the interval between the first time corresponding to obtaining the target influence effect value and the start time of the first period.

[0100] In one implementation, the parameter determination unit 601 is specifically configured to: obtain candidate influence parameters; perform a first process on the candidate influence parameters to obtain an influence parameter sequence; and determine the target influence parameter based on the influence parameter sequence.

[0101] Wherein, the target weight is obtained through the following method: based on a normalization algorithm, process the target correlation coefficient to determine the target weight corresponding to the target correlation coefficient.

[0102] In one implementation, before determining the target influence parameter, the parameter determination unit 601 is further configured to: monitor whether an update condition is satisfied; the update condition is related to a second period; the second period is greater than the first period; and when the update condition is satisfied, perform the operation of determining the target influence parameter.

[0103] It should be noted that the specific implementation of each unit in this embodiment can refer to the corresponding content in the previous text and will not be elaborated here.

[0104] Reference Figure 7, which is a schematic structural diagram of an electronic device provided in Embodiment 3 of the present application. The electronic device is an electronic device capable of performing a data processing method, such as a computer or a server. The technical solution in this embodiment is mainly used to accurately detect the time delay influence value of a target influence parameter on a target parameter.

[0105] Specifically, the electronic device in this embodiment may include the following structure:

[0106] A memory 701, configured to store a computer program and data generated by the running of the computer program;

[0107] A processor 702, configured to execute the computer program to implement: determining a target influence parameter; determining a target correlation coefficient of the target influence parameter and a target weight corresponding to the target correlation coefficient based on the correlation relationship between the target influence parameter and the target parameter; and determining a time delay influence value of the target influence parameter on the target parameter based on the target correlation coefficient and the target weight.

[0108] It can be seen from the above technical solution that in an electronic device provided in Embodiment 3 of the present application, after determining the target influence parameter, the target correlation coefficient of the target influence parameter and the target weight corresponding to the target correlation coefficient can be determined according to the correlation relationship between the target influence parameter and the target parameter, and then the time delay influence value of the target influence parameter on the target parameter can be determined according to the target correlation coefficient and the target weight. It can be seen that in this embodiment, the target correlation coefficient of the target influence parameter and the corresponding weight are used to determine the time delay influence value, so as to facilitate process control.

[0109] Taking a tobacco shred making device as an example, since it involves different devices such as an electronic scale, a conditioning machine, a drying machine, and a flavoring machine, and the process results of each device affect each other, the time lag effects of various process production parameters on the moisture content of the tobacco shreds at the outlet are different.

[0110] In view of this, a large time delay processing solution based on artificial intelligence technology in real time and online is proposed in this embodiment.

[0111] Such as Figure 8As shown, in this solution, first, from hundreds or even thousands of process production parameters, a first screening is carried out according to the weighted fusion coefficient model of Pearson correlation coefficient and Spearman rank correlation coefficient. For example, in this solution, from hundreds or thousands of process production parameters, the Pearson correlation coefficient of each process production parameter (such as calculating process parameter A this time) with the outlet moisture content (denoted as A1, and the corresponding weight of A1 is denoted as W1), and the Spearman rank correlation coefficient with the outlet moisture content (denoted as A2, and the corresponding weight of A2 is denoted as W2) are calculated. Then what the weighted fusion coefficient model needs to do is to calculate the correlation value of process production parameter A as A1*W1 + A2*W2; calculate the correlation values of each process production parameter B, C, D, E... in this way, and take the process production parameters corresponding to the correlation value greater than a certain threshold or the top N as the process production parameters after this round of screening.

[0112] Where A1 is the Pearson correlation coefficient, A2 is the Spearman rank correlation coefficient, A1 and A2 can be calculated, and W1 and W2 can be set according to experience or using the normalization method. For example, the settings can be 0.5 and 0.5.

[0113] After that, this solution conducts autocorrelation tests on the screened process production parameters to filter out the process production parameters that have little impact on the outlet moisture content of cut tobacco. For example, each process production parameter (such as the actual pressure of the progressive thin plate steam in the cut tobacco dryer, the actual pressure of the steam entering the radiator of the cut tobacco dryer, the actual opening of the mixing air damper of the cut tobacco dryer, the steam pressure of the HT preheating entering the sandwich plate, etc.) passes the autocorrelation test, and redundant parameters can be eliminated.

[0114] Then, in this solution, a model based on the Feature Imports (Feature Importances) mechanism is established to select the process production parameters whose model output results are ranked in the top 30%. That is to say, the process production parameters first pass through the screening of the weighted fusion coefficient model and autocorrelation test, and the screened process production parameters then pass through the feauture imports mechanism for the second round of screening.

[0115] Thus, in this embodiment, the weighted fusion coefficient model is used for parameter screening, which is fast and can quickly eliminate irrelevant parameters, reducing the processing volume of the Feature Imports mechanism. Moreover, in this embodiment, the weighted fusion coefficient model and the Feature Imports mechanism calculate the relationship between process production parameters and target parameters from different angles. The weighted fusion coefficient deals with linearly correlated relationships, and the Feature Imports mechanism can deal with non-linearly correlated relationships. The focus of FeatureImports is on feature importance, that is, the contribution of parameters in the model.

[0116] Finally, based on the current artificial intelligence technology, this solution establishes a self-sliding detection correlation model and outputs the lag time of the process production parameters in the next production cycle (i.e., the delay impact value in the previous text).

[0117] Among them, the main steps of the self-sliding detection correlation model are as follows:

[0118] 1. Set the maximum delay time to 600 seconds (the first cycle, because for a drying and roasting machine, a production cycle from the first process to the last process, the time of a set of production lines is 10 minutes), set the time period length to 90 days, and set the update duration to 1 day (the second cycle);

[0119] 2. According to the set time period length, intercept the production data of the past 90 days as the data for analysis;

[0120] 3. According to the results of Feature Imports, sort the important process production parameters (i.e., target impact parameters) in the results from the most important to the least important;

[0121] 4. Select the first process production parameter, and at intervals of 10 seconds according to time, correlate the data of this process production parameter with the moisture content at the outlet, and calculate the delay effect (i.e., the impact effect value) of each correlation. When the maximum delay time (600 seconds) is reached, stop the calculation. The calculation method is as follows;

[0122] (1) Delay effect = w1 * Pearson correlation coefficient + w2 * Spearman rank correlation coefficient + w3 * Feature Imports importance + w4 * high-order partial correlation coefficient, where the Pearson correlation coefficient, Spearman rank correlation coefficient, Feature Imports importance, and high-order partial correlation coefficient are calculated based on the production data of the previous 90 days. For example, for the Pearson correlation coefficient, obtain the Pearson correlation coefficient values calculated P times in the previous 90 days, take the average of these N Pearson correlation coefficient values, and use this average as the value of the Pearson correlation coefficient in step (1). Similarly, the other three types of correlation coefficients are obtained in a similar manner; obtaining various correlation coefficients based on recent historical data can be closer to the production situation in the next production cycle.

[0123] (2) Calculation of weights w1 - w4: Based on the historical data of two years, calculate the Pearson correlation coefficient C1, Spearman rank correlation coefficient C2, Feature Imports importance C3, and high - order partial correlation coefficient C4 for each process production parameter. Specifically, similar to step (1), obtain C1 - C4 based on the average value of the historical data of two years, and then normalize C1, C2, C3, and C4. The normalized values w1, w2, w3, and w4 are used as weights. Obtaining weights based on historical data with a long time span can avoid atypical data that may be brought by a short time span.

[0124] 5. Compare the delay effect values every 10 seconds, and take the time interval corresponding to the maximum delay effect value as the lag time (delay impact value) of this process production parameter; for example, if the delay effect value calculated at the 8th 10 - second interval is the largest, then 80 seconds is the lag time of this process production parameter.

[0125] 6. Return to step 4 and sequentially loop to the next process production parameter;

[0126] 7. Automatically update according to the update duration, that is, execute steps 1 - 6 every update duration, and output the lag time of each process production parameter after the update.

[0127] It can be seen that in this solution, the deviation of human subjective estimation can be reduced. By analyzing the relationships between different process production parameters, the stability and accuracy of solving the time - delay problem are improved. Moreover, in this solution, according to real - time data, continuous iterative updates are carried out to improve applicability.

[0128] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and reference can be made to the description in the method part for related parts.

[0129] Professionals can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0130] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented directly in hardware, in software modules executed by a processor, or in a combination thereof. The software modules may be located in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium well known in the art.

[0131] The foregoing description of the disclosed embodiments enables those skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A data processing method, comprising: Determining a target impact parameter, where the target impact parameter is an impact parameter that affects a target parameter; Based on the correlation relationship between the target impact parameter and the target parameter, determining a target correlation coefficient of the target impact parameter and a target weight corresponding to the target correlation coefficient; Based on the target correlation coefficient and the target weight, determining a delay impact value of the target impact parameter on the target parameter.

2. The method according to claim 1, wherein the determining the delay impact value of the target impact parameter on the target parameter based on the target correlation coefficient and the target weight comprises: Based on the target correlation coefficient and the target weight, obtaining an impact effect value of the target impact parameter on the target parameter; Based on the impact effect value, determining the delay impact value of the target impact parameter on the target parameter.

3. The method according to claim 2, wherein the determining the delay impact value of the target impact parameter on the target parameter based on the impact effect value comprises: Determining a target impact effect value based on the impact effect value; Based on the target impact effect value, determining the delay impact value of the target impact parameter on the target parameter.

4. The method according to claim 3, wherein obtaining the impact effect value of the target impact parameter on the target parameter based on the target correlation coefficient and the target weight comprises: In at least one first sub - period within a first period, based on the target correlation coefficient and the target weight, obtaining at least one impact effect value of each target impact parameter on the target parameter.

5. The method according to claim 4, wherein the determining the delay impact value of the target impact parameter on the target parameter based on the target impact effect value comprises: Based on the interval between the first time corresponding to obtaining the target impact effect value and the start time of the first period, determining the delay impact value corresponding to the first period.

6. The method according to any one of claims 1 - 5, wherein the determining the target impact parameter comprises: Obtaining candidate impact parameters; Performing a first process on the candidate impact parameters to obtain an impact parameter sequence; Based on the impact parameter sequence, determining the target impact parameter.

7. The method according to claim 6, wherein the target weight is obtained by the following method: Based on a normalization algorithm, processing the target correlation coefficient to determine a target weight corresponding to the target correlation coefficient.

8. The method according to any one of claims 4 - 5, before the determining the target impact parameter, the method further comprises: Monitoring whether an update condition is satisfied; The update condition is related to a second period; The second period is greater than the first period; When the update condition is satisfied, performing the: determining the target impact parameter.

9. A data processing device, comprising: A parameter determination unit for determining a target impact parameter, where the target impact parameter is an impact parameter that affects a target parameter; A target determination unit, configured to determine a target correlation coefficient of the target impact parameter and a target weight corresponding to the target correlation coefficient based on the correlation relationship between the target impact parameter and the target parameter; A time delay determination unit, configured to determine a time delay impact value of the target impact parameter on the target parameter based on the target correlation coefficient and the target weight.

10. An electronic device, comprising: A memory, configured to store a computer program and data generated by running the computer program; A processor, configured to execute the computer program to implement: determining a target impact parameter, where the target impact parameter is an impact parameter that affects a target parameter; determining a target correlation coefficient of the target impact parameter and a target weight corresponding to the target correlation coefficient based on the correlation relationship between the target impact parameter and the target parameter; Determining a time delay impact value of the target impact parameter on the target parameter based on the target correlation coefficient and the target weight.

Images