Production process monitoring method and system based on big data analysis

Abstract

The production process monitoring method and system based on big data analysis provided in this application, based on the embodiments of this application, obtains multi-dimensional behavioral description content through relative distribution adjustment operations and key user behavior description mining operations. This is beneficial for monitoring online user activity information in the production process. Furthermore, by obtaining multi-dimensional behavioral description content through multi-dimensional mining operations, the multi-dimensional behavioral description content can be obtained more accurately, improving the accuracy of the mining operations. Thus, the accuracy and credibility of the description content can be effectively improved through multi-dimensional mining operations.

Description

Technical Field

[0001] This application relates to the field of data monitoring and analysis technology, and more specifically, to a method and system for monitoring production processes based on big data analysis. Background Technology

[0002] Big data analytics refers to the analysis of massive amounts of data. Big data can be summarized as the 4 Vs: Volume, Velocity, Variety, and Value. As one of the hottest terms in the IT industry today, big data has led to the development of data warehousing, data security, data analysis, data mining, and other related technologies, all centered around leveraging the commercial value of big data, which have become highly sought-after profit drivers.

[0003] Currently, the scope of big data analytics technology is constantly expanding, such as its integration with smart healthcare (Wise Information Technology of Medicine) and smart home (smart home automation) data analysis and processing. Furthermore, with the intelligent upgrading of production processes, the integration of big data with production processes is becoming increasingly mature. Even so, the various scenarios in process monitoring data can significantly interfere with the accurate analysis of relevant data. Without considering every possible scenario, it is difficult to guarantee the accuracy and reliability of production process monitoring data. Summary of the Invention

[0004] In view of this, this application provides a production process monitoring method and system based on big data analysis.

[0005] Firstly, a production process monitoring method based on big data analysis is provided, including:

[0006] From the online user activity information of the business user big data to be analyzed, several target online user activity information is identified; key user behavior description mining operations are performed on the several target online user activity information for several production process states to obtain the first key user behavior description distribution of the target online user activity information for several production process states.

[0007] A multi-dimensional mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain multi-dimensional behavior description content of the target online user activity information. The multi-dimensional mining operation includes a relative distribution adjustment operation and a key user behavior description mining operation. The relative distribution adjustment operation includes adjusting the relative distribution of the original key user behavior description distribution and the processed key user behavior description distribution of one or more key user behavior description mining operations. Based on the multi-dimensional behavior description content, production process monitoring is performed on the business user big data to be analyzed to obtain the production process monitoring status.

[0008] In one standalone embodiment, a multi-dimensional mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain multi-dimensional behavior description content of the target online user activity information, including:

[0009] A first relative distribution adjustment operation is performed on the first key user behavior description distribution of the target online user activity information to obtain a plurality of first key user behavior description distribution sets, wherein the order of the first key user behavior description distributions in the first key user behavior description distribution sets is consistent with the order of the corresponding target online user activity information.

[0010] Perform key user behavior description mining operation on the first key user behavior description distribution in the first key user behavior description distribution set to obtain a second key user behavior description distribution set, wherein the second key user behavior description distribution set includes a second key user behavior description distribution corresponding to the first key user behavior description distribution;

[0011] A second relative distribution adjustment operation is performed on the second key user behavior description distribution set to obtain the multi-dimensional behavior description content.

[0012] In one standalone embodiment, a first relative distribution adjustment operation is performed on the first key user behavior description distribution of the target online user activity information to obtain a plurality of first key user behavior description distribution sets, including:

[0013] For each target online user activity information, several first key user behavior descriptions of the target online user activity information are clustered to obtain several first key user behavior description clusters of the target online user activity information. Each first key user behavior description cluster carries a semantic label, and the first key user behavior description clusters of each target online user activity information carrying the same semantic label correspond to the same group of production process states in several production process states.

[0014] In the first key user behavior description clusters of each target online user activity information, the first key user behavior description clusters with consistent semantic labels are concatenated based on the chronological relationship of the target online user activity information to obtain the first key user behavior description distribution set corresponding to the semantic labels.

[0015] In one standalone embodiment, a second relative distribution adjustment operation is performed on the second key user behavior description distribution set to obtain the multi-dimensional behavior description content, including:

[0016] The second key user behavior description distribution in the second key user behavior description distribution set is clustered based on the corresponding target online user activity information to obtain the second key user behavior description cluster. The second key user behavior description distribution in the same second key user behavior description cluster corresponds to the same target online user activity information.

[0017] The clusters of each second key user behavior description are concatenated to obtain multi-dimensional behavior descriptions corresponding to the online user activity information of each target.

[0018] In one standalone embodiment, a multi-dimensional mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain multi-dimensional behavior description content of the target online user activity information, including:

[0019] A first key user behavior description mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain a third key user behavior description distribution corresponding to the first key user behavior description distribution.

[0020] A third relative distribution adjustment operation is performed on the third key user behavior description distribution to obtain several sets of third key user behavior description distributions, wherein the order of the third key user behavior description distributions in the set of third key user behavior description distributions is consistent with the order of the corresponding target online user activity information.

[0021] Perform a second key user behavior description mining operation on the third key user behavior description distribution in the third key user behavior description distribution set to obtain a fourth key user behavior description distribution set, wherein the fourth key user behavior description distribution set includes a fourth key user behavior description distribution corresponding to the third key user behavior description distribution;

[0022] A fourth relative distribution adjustment operation is performed on the fourth key user behavior description distribution set to obtain the third key user behavior description cluster corresponding to the target online user activity information;

[0023] A third key user behavior description mining operation is performed on the fourth key user behavior description distribution in the third key user behavior description cluster to obtain multi-dimensional behavior description content corresponding to the target online user activity information.

[0024] In one standalone embodiment, from a number of online user activity information items in the business user big data to be analyzed, several target online user activity information items are identified, including:

[0025] The system filters several online user activity information from the business user big data to be analyzed, and obtains several business user big data fragments to be analyzed.

[0026] The data fragments of business users to be analyzed are subjected to information parsing operations to obtain the activity information of the target online users.

[0027] In one standalone embodiment, the production process monitoring method based on big data analytics is implemented using a machine learning model, and the method further includes:

[0028] From the online user activity information of the example business user big data, the example online user activity information was identified;

[0029] The online user activity information of the example is loaded into the machine learning model to obtain the test production process monitoring status of the big data of the example business users;

[0030] Based on the prior annotations of the test production process monitoring and the example business user big data, the model quality evaluation of the machine learning model is determined.

[0031] Based on the model quality evaluation, the machine learning model is optimized.

[0032] Secondly, a production process monitoring system based on big data analysis is provided, including a processor and a memory that communicate with each other. The processor is used to retrieve a computer program from the memory and implement the above-mentioned method by running the computer program.

[0033] The production process monitoring method and system based on big data analysis provided in this application, based on the embodiment of this application, obtains multi-dimensional behavioral description content through relative distribution adjustment operation and key user behavior description mining operation. This is beneficial for monitoring production processes based on online user activity information. Furthermore, by obtaining multi-dimensional behavioral description content through multi-dimensional mining operation, the multi-dimensional behavioral description content can be obtained more accurately, improving the accuracy of the mining operation. Thus, the accuracy and credibility of the description content can be effectively improved through multi-dimensional mining operation. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart illustrating a production process monitoring method based on big data analysis, provided as an embodiment of this application.

[0036] Figure 2 This is a block diagram of a production process monitoring device based on big data analysis, provided as an embodiment of this application.

[0037] Figure 3 This is an architecture diagram of a production process monitoring system based on big data analysis, provided as an embodiment of this application. Detailed Implementation

[0038] To better understand the above technical solutions, the technical solutions of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this application and the specific features in the embodiments are detailed descriptions of the technical solutions of this application, rather than limitations on the technical solutions of this application. In the absence of conflict, the embodiments of this application and the technical features in the embodiments can be combined with each other.

[0039] Please see Figure 1 This paper presents a production process monitoring method based on big data analysis, which may include the technical solutions described in steps 11-14 below.

[0040] Step 11: From the online user activity information of the business user big data to be analyzed, identify several target online user activity information.

[0041] For example, target online user activity information can characterize relevant information that is monitored in real time during the production process (such as production process data monitored by monitoring equipment).

[0042] Step 12: Perform key user behavior description mining operations on the target online user activity information for several production process states to obtain the first key user behavior description distribution of the target online user activity information for several production process states.

[0043] For example, the production process status is used to characterize the state of the relevant production equipment during production.

[0044] Step 13: Perform multi-dimensional mining operations on the first key user behavior description distribution of the target online user activity information to obtain the multi-dimensional behavior description content of the target online user activity information.

[0045] For example, the multi-dimensional mining operation includes a relative distribution adjustment operation and a key user behavior description mining operation, wherein the relative distribution adjustment operation includes adjusting the relative distribution of the original key user behavior description and the processed key user behavior description distribution of one or more key user behavior description mining operations.

[0046] Step 14: Based on the multi-dimensional behavioral description content, perform production process monitoring on the business user big data to be analyzed to obtain the production process monitoring status.

[0047] For example, production process monitoring can be understood as the results obtained by monitoring and analyzing the production process.

[0048] It is understood that the production process monitoring method based on big data analysis in the embodiments of this application obtains multi-dimensional behavioral descriptions through relative distribution adjustment operations and key user behavior description mining operations. This is beneficial for monitoring online user activity information in the production process. Furthermore, by obtaining multi-dimensional behavioral descriptions through multi-dimensional mining operations, the multi-dimensional behavioral descriptions can be obtained more accurately, improving the accuracy of the mining operations. Thus, the accuracy and credibility of the descriptions can be effectively improved through multi-dimensional mining operations.

[0049] In an alternative embodiment, the business user big data to be analyzed can be any online user activity information, such as online user activity information recorded by a process monitoring system. The business user big data to be analyzed can cover one of these contents. The business user big data to be analyzed may include several online user activity information. In the monitoring step of the production process monitoring of the content of the business user big data to be analyzed, the global online user activity information of the business user big data to be analyzed can be identified and processed, or local online user activity information can be filtered and processed. For example, local online user activity information with significant characteristics can be filtered and processed.

[0050] In an alternative embodiment, the step of filtering local online user activity information for identification may specifically include the content described in step 11.

[0051] The system performs a filtering operation on several online user activity information segments of the business user big data to be analyzed, thereby obtaining several business user big data segments to be analyzed; and performs an information parsing operation on the business user big data segments to be analyzed, thereby obtaining the several target online user activity information segments.

[0052] Understandably, when analyzing a large dataset of online user activity information, it is necessary to improve the accuracy of the filtering process and thus accurately identify the target online user activity information.

[0053] In an alternative embodiment, a filtering operation can be performed on several online user activity information segments of the business user big data to be analyzed. For example, data filtering can be performed on a search interface or in a data dashboard to obtain M (M is a positive integer) business user big data segments to be analyzed. This application does not limit the filtering method to any one of them.

[0054] In an alternative embodiment, information parsing operations can be performed on several business user big data segments to be analyzed.

[0055] In an alternative embodiment, in step 12, a machine learning model can be used to perform key user behavior description mining operations on several production process states of the target online user activity information to obtain a first key user behavior description distribution for several production process states of each target online user activity information. For example, the machine learning model can be an AI machine learning model.

[0056] In an alternative embodiment, in step 13, since there is spatial order information among the online user activity information of the business user big data to be analyzed, multi-dimensional behavioral descriptions of the target online user activity information can be obtained to represent the spatial order information among the online user activity information of the business user big data to be analyzed, thereby providing more complete information on the types of online user activity information for production process monitoring.

[0057] In an alternative embodiment, multi-dimensional behavioral descriptions of target online user activity information can be obtained by performing relative distribution adjustment operations and key user behavior description mining operations on the key user behavior description distributions of each production process state. The multi-dimensional behavioral description content is key information used to characterize the matching situation between online user activity information, such as a key user behavior description distribution based on the chronological relationship of online user activity information. The multi-dimensional behavioral description content can characterize the location of online user activity information; that is, it can represent the description of online user activity information based on the chronological relationship of online user activity information. For example, a piece of information in online user activity information appears in region Q in the first online user activity information, appears in region P in the second online user activity information, and so on. The multi-dimensional behavioral description content can represent the description in each online user activity information based on the chronological relationship of online user activity information. The relative distribution adjustment operation covers the relative distribution adjustment of the original key user behavior description distribution and the processed key user behavior description distribution of one or more key user behavior description mining operations.

[0058] The steps of the multi-dimensional mining operation based on the embodiments of this application, specifically step 13 may include the following:

[0059] A first relative distribution adjustment operation is performed on the first key user behavior description distribution of the target online user activity information to obtain several sets of first key user behavior description distributions, wherein the order of the first key user behavior description distributions in the first key user behavior description distribution sets is consistent with the order of the corresponding target online user activity information; a key user behavior description mining operation is performed on the first key user behavior description distributions in the first key user behavior description distribution sets to obtain a second set of key user behavior description distributions, wherein the second set of key user behavior description distributions includes a second key user behavior description distribution corresponding to the first key user behavior description distribution (i.e., the second key user behavior description distribution obtained through the key user behavior description mining operation on the first key user behavior description distribution); a second relative distribution adjustment operation is performed on the second set of key user behavior description distributions to obtain the multi-dimensional behavior description content.

[0060] It is understandable that, in performing the above steps, by carrying out the first relative distribution adjustment operation, the accuracy of the first key user behavior description distribution set can be effectively improved, thereby improving the completeness and accuracy of the second key user behavior description distribution set.

[0061] In an alternative embodiment, the machine learning model may include several units. At least one unit of the machine learning model can perform key user behavior description mining operations on the target online user activity information to obtain a first key user behavior description distribution of several production process states. Furthermore, the remaining units of the machine learning model can continuously process the first key user behavior description distribution of several production process states.

[0062] In an alternative embodiment, a first relative distribution adjustment operation is performed on the first key user behavior description distribution of the target online user activity information to obtain a plurality of first key user behavior description distribution sets, specifically including the following:

[0063] For each target online user activity information, several first key user behavior description distributions of the target online user activity information are clustered to obtain several first key user behavior description clusters of the target online user activity information. Each first key user behavior description cluster carries a semantic label, and the first key user behavior description clusters of the target online user activity information carrying the same semantic label correspond to the same group of production process states among several production process states. The first key user behavior description clusters with consistent semantic labels in each of the target online user activity information clusters are concatenated according to the chronological relationship of the target online user activity information to obtain a set of first key user behavior description distributions corresponding to the semantic labels.

[0064] Understandably, when performing a first relative distribution adjustment operation on the first key user behavior description distribution of the target online user activity information, the problem of inaccurate clustering is improved, thereby enabling the accurate acquisition of several sets of first key user behavior description distributions.

[0065] For example, the distribution of several first key user behavior descriptions for each group of target online user activity information can be represented by {I, O, P}, where I (I is a positive integer) represents the production process status description result, O (O is a positive integer) represents the distribution of the first key user behavior descriptions, and P (P is a positive integer) represents the layout of the first key user behavior description distributions. That is, each group of target online user activity information corresponds to I first key user behavior description distributions for production process statuses.

[0066] For example, the distributions of first key user behavior descriptions for I production process states can be clustered, or the production process states can be clustered. The number of first key user behavior description distributions or the number of production process states covered by each group can be determined by the number M of target online user activity information and the production process state description result I. For example, each group can cover I / M first key user behavior description distributions or production process states. For example, if the production process state description result I is 128 and the number M of target online user activity information is 16, then each group can cover 16 first key user behavior description distributions, or each group can cover 16 production process states. Each first key user behavior description cluster can carry semantic labels.

[0067] In an alternative embodiment, the distributions of first key user behavior descriptions with consistent semantic tags in the first key user behavior description clusters of each target online user activity information are concatenated according to the chronological relationship of the target online user activity information to obtain a set of first key user behavior description distributions corresponding to the semantic tags.

[0068] For example, the distribution of the first key user behavior descriptions for several production process states of each target online user activity information can be clustered according to the above method. For instance, the distribution of the first key user behavior descriptions for each target online user activity information can be divided into 16 groups. Furthermore, the clustering criteria for the first key user behavior descriptions of M target online user activity information can be combined using the chronological relationship between the target online user activity information.

[0069] For example, the target online user activity information is a set of online user activity information carrying a time stamp or frame number, that is, there is a chronological relationship between each target online user activity information. The chronological relationship of the target online user activity information can be concatenated using the first key user behavior description clustering credential with consistent semantic tags.

[0070] For example, the clusters of the first key user behavior descriptions of the first group of the first target online user activity information, the first key user behavior descriptions of the second target online user activity information, the first key user behavior descriptions of the third target online user activity information, ... the first key user behavior descriptions of the sixteenth target online user activity information are concatenated to obtain the distribution set of the first key user behavior descriptions corresponding to semantic label 1.

[0071] For example, the clusters of the second group of first key user behavior descriptions for the first target online user activity information, the second group of first key user behavior descriptions for the second target online user activity information, the second group of first key user behavior descriptions for the third target online user activity information, and so on, up to the 16th target online user activity information, are concatenated to obtain the distribution set of first key user behavior descriptions corresponding to semantic label 2. This method can be used to obtain the distribution set of first key user behavior descriptions corresponding to each semantic label. In other words, by changing the order of the distributions of first key user behavior descriptions using target online user activity information as clustering credentials, the distribution set of first key user behavior descriptions using semantic labels as clustering credentials is obtained. Within each set of first key user behavior description distributions, each cluster of first key user behavior descriptions is distributed according to the chronological relationship of the target online user activity information. That is, the distribution is based on the chronological relationship of the first key user behavior description cluster of the first target online user activity information, the second target online user activity information, ..., the 16th target online user activity information, so as to obtain the spatial order information of the target online user activity information in subsequent processing.

[0072] In an alternative embodiment, a key user behavior description mining operation can be performed on the first key user behavior description distribution in the first key user behavior description distribution set. For example, descriptions of each first key user behavior description distribution in the first key user behavior description distribution set can be extracted based on the chronological relationship of the target online user activity information; the obtained descriptions can cover the spatial order information among the target online user activity information.

[0073] For example, key user behavior description mining operations can be performed on each of the first key user behavior description distributions in the first key user behavior description distribution set using three units: 2*2, 6*6, and 2*2. This yields the second key user behavior description distribution bound to each first key user behavior description distribution, and these second key user behavior description distributions can constitute a second key user behavior description distribution set. That is, the second key user behavior description distribution set corresponding to each semantic tag can be obtained using the above method.

[0074] In an alternative embodiment, after obtaining the second key user behavior description distribution set, a second relative distribution adjustment operation can be performed on the second key user behavior description distribution set. That is, the second key user behavior description distribution set based on semantic tags as clustering credentials is adjusted to optimize each second key user behavior description distribution in each second key user behavior description distribution set into a clustering method based on target online user activity information as clustering credentials, thereby obtaining multi-dimensional behavior description content.

[0075] In an alternative embodiment, a second relative distribution adjustment operation is performed on the second key user behavior description distribution set to obtain the multi-dimensional behavior description content, including:

[0076] The second key user behavior description distributions in the second key user behavior description distribution set are clustered based on the corresponding target online user activity information to obtain second key user behavior description clusters. The second key user behavior description distributions in the same second key user behavior description cluster correspond to the same target online user activity information. The second key user behavior description clusters are then concatenated to obtain multi-dimensional behavior description content corresponding to each target online user activity information.

[0077] It is understandable that when performing a second relative distribution adjustment operation on the second key user behavior description distribution set, the unreliability of clustering target online user activity information is improved, thereby enabling the reliable acquisition of the multi-dimensional behavior description content.

[0078] For example, each of the second key user behavior description distributions in the second key user behavior description distribution set is obtained by mining key user behavior descriptions from the first key user behavior description distribution. Therefore, each second key user behavior description distribution also corresponds to target online user activity information. Clustering can be performed based on the target online user activity information corresponding to the second key user behavior description distribution to obtain second key user behavior description clusters. For example, for a certain second key user behavior description distribution set, the second key user behavior description distributions in the set can be clustered into second key user behavior description clusters for the first target online user activity information, second key user behavior description clusters for the second target online user activity information, ..., and so on, up to the 16th target online user activity information. Clustering can be performed on the second key user behavior description distributions in each second key user behavior description distribution set in this way to obtain second key user behavior description clusters in each second key user behavior description distribution set.

[0079] For example, clusters of second key user behavior descriptions that correspond to the target online user activity information in several second key user behavior description distribution sets can be concatenated. For instance, the clusters of second key user behavior descriptions corresponding to the first target online user activity information in the first second key user behavior description distribution set, the clusters of second key user behavior descriptions corresponding to the first target online user activity information in the second second key user behavior description distribution set, ..., the clusters of second key user behavior descriptions corresponding to the first target online user activity information in the 16th second key user behavior description distribution set can be concatenated to obtain multi-dimensional behavior description content corresponding to the first target online user activity information. The clusters of second key user behavior descriptions corresponding to the second target online user activity information in the first second key user behavior description distribution set, the second key user behavior description clusters corresponding to the second target online user activity information in the second second key user behavior description distribution set, ..., the 16th second key user behavior description distribution set corresponding to the second target online user activity information can be concatenated to obtain multi-dimensional behavior description content corresponding to the second target online user activity information. The above method allows for the re-assembly of the second key user behavior description distribution within the second key user behavior description distribution set; that is, a second relative distribution adjustment operation is performed to obtain multi-dimensional behavior description content corresponding to each target online user activity information. This multi-dimensional behavior description content encompasses the spatial order information among the target online user activity information, which is beneficial for monitoring production processes using the business user big data to be analyzed. For example, the types of online user activity information for production process monitoring can be categorized, such as the indication of a target within the online user activity information for production process monitoring.

[0080] Through the above method, after the first relative distribution adjustment operation, the machine learning model can obtain multi-dimensional behavioral descriptions consistent with the temporal sequence, which is beneficial for expressing the content of online user activity information in a temporal sequence. Then, through the second relative distribution adjustment operation, the relative distribution of the key user behavior description distribution is adjusted to a chronological relationship based on the corresponding target online user activity information. That is, it is represented as the multi-dimensional behavioral description content corresponding to each target online user activity information. This ensures that the obtained multi-dimensional behavioral description content includes the location information of each target online user activity information, which is beneficial for monitoring the production process of the business user big data to be analyzed.

[0081] The steps of the multi-dimensional mining operation based on the embodiments of this application, specifically step 13 may include:

[0082] A first key user behavior description mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain a third key user behavior description distribution corresponding to the first key user behavior description distribution; a third relative distribution adjustment operation is performed on the third key user behavior description distribution to obtain several sets of third key user behavior description distributions, wherein the order of the third key user behavior description distributions in the sets of third key user behavior description distributions is consistent with the order of the corresponding target online user activity information; a second key user behavior description mining operation is performed on the third key user behavior description distributions in the sets of third key user behavior description distributions to obtain a fourth key user behavior description distribution set, wherein the fourth key user behavior description distribution set includes the fourth key user behavior description distribution corresponding to the third key user behavior description distribution; a fourth relative distribution adjustment operation is performed on the fourth key user behavior description distribution set to obtain a third key user behavior description cluster corresponding to the target online user activity information;

[0083] A third key user behavior description mining operation is performed on the fourth key user behavior description distribution in the third key user behavior description cluster to obtain multi-dimensional behavior description content corresponding to the target online user activity information.

[0084] It is understandable that when performing multi-dimensional mining operations on the distribution of the first key user behavior description of the target online user activity information, the problem of inaccurate relative distribution adjustment operations can be improved, thereby enabling the accurate acquisition of multi-dimensional behavioral description content of the target online user activity information.

[0085] In an alternative embodiment, a first key user behavior description mining operation can be performed on the first key user behavior description distribution to obtain a third key user behavior description distribution corresponding to each first key user behavior description distribution. For example, the first key user behavior description distribution can be processed in 2*2 to delete irrelevant information and obtain the third key user behavior description distribution bound to each first key user behavior description distribution.

[0086] In an alternative embodiment, a third relative distribution adjustment operation can be performed on the third key user behavior description distribution to obtain several sets of third key user behavior description distributions. The third relative distribution adjustment operation is similar to the first relative distribution adjustment operation; the third key user behavior description distribution corresponding to each target online user activity information can be clustered, for example, into 16 groups. Then, the clustering credentials of the third key user behavior descriptions with consistent semantic tags are concatenated to represent the chronological relationship of the target online user activity information. For example, the clusters of the first group of third key user behavior descriptions for the first target online user activity information, the first group of third key user behavior descriptions for the second target online user activity information, the first group of third key user behavior descriptions for the third target online user activity information, ..., the first group of third key user behavior descriptions for the sixteenth target online user activity information are concatenated to obtain the distribution set of third key user behavior descriptions corresponding to semantic label 1. Within this distribution set, the distributions can be concatenated according to the order of the third key user behavior descriptions for the first target online user activity information, the second target online user activity information, the third target online user activity information, ... . Furthermore, the distribution set of third key user behavior descriptions corresponding to each semantic label can be obtained using the above method.

[0087] In an alternative embodiment, a second key user behavior description mining operation can be performed on the third key user behavior description distribution in the third key user behavior description distribution set to obtain a fourth key user behavior description distribution set. For example, each third key user behavior description distribution can be processed in a 6*6 manner to obtain the fourth key user behavior description distribution, and each fourth key user behavior description distribution can constitute the fourth key user behavior description distribution set. That is, according to the above method, the fourth key user behavior description distribution set corresponding to each semantic tag can be obtained, and the fourth key user behavior description distribution set can cover the spatial order information among the target online user activity information.

[0088] In an alternative embodiment, a fourth relative distribution adjustment operation can be performed on the fourth key user behavior description distribution set. Similar to the second relative distribution adjustment operation, the fourth key user behavior description distributions in the fourth key user behavior description distribution set can be clustered according to their corresponding target online user activity information, and the fourth key user behavior description distributions in each fourth key user behavior description distribution set that correspond to the same target online user activity information can be concatenated. For example, the fourth key user behavior description distributions corresponding to the first target online user activity information in the first fourth key user behavior description distribution set, the fourth key user behavior description distributions corresponding to the first target online user activity information in the second fourth key user behavior description distribution set, ..., the fourth key user behavior description distributions corresponding to the first target online user activity information in the 16th fourth key user behavior description distribution set can be concatenated to obtain a third key user behavior description cluster corresponding to the first target online user activity information. Further, a third key user behavior description cluster corresponding to each target online user activity information can be obtained in this manner.

[0089] In an alternative embodiment, a third key user behavior description mining operation can be performed on the fourth key user behavior description distribution within the third key user behavior description cluster to obtain multi-dimensional behavior description content corresponding to each target online user activity information. For example, a 2*2 processing can be performed on the fourth key user behavior description distribution within the third key user behavior description cluster to obtain multi-dimensional behavior description content corresponding to each target online user activity information. This multi-dimensional behavior description content encompasses the spatial order information among the target online user activity information, which is beneficial for monitoring production processes using the business user big data to be analyzed. For example, the types of online user activity information can be monitored for production processes.

[0090] In an alternative embodiment, in addition to the multi-dimensional mining operation method, other multi-dimensional mining operations can be performed. For example, the first relative distribution adjustment operation can be performed between the first 2*2 and 6*6 operations, or the second relative distribution adjustment operation can be performed between 6*6 and the second 2*2 operations. Alternatively, the third relative distribution adjustment operation can be performed before the first 2*2 operation, or the fourth relative distribution adjustment operation can be performed after the second 2*2 operation. This application does not limit the processing order of the multi-dimensional mining operations.

[0091] In an alternative embodiment, in step 14, the production process monitoring status of the business user big data to be analyzed can be determined based on the first key user behavior description distribution and the multi-dimensional behavior description content. For example, assuming the machine learning model is an evaluation thread, the first key user behavior description distribution and the multi-dimensional behavior description content can be continuously processed by subsequent thread units of the machine learning model, and the production process monitoring status of the business user big data to be analyzed can be output. For example, the production process monitoring status can be the type of business user big data to be analyzed, and the type can be expressed by percentage or other means. This application does not limit the form of the type of business user big data to be analyzed.

[0092] For example, the machine learning model may not be an evaluation thread. Instead, it can be processed by subsequent thread units of the machine learning model to continuously analyze multi-dimensional behavioral descriptions and output the production process monitoring status of the business user big data to be analyzed.

[0093] In an alternative embodiment, performing a relative distribution adjustment operation before or after at least one unit of the machine learning model enables the model to extract multi-dimensional behavioral descriptions. This assists the machine learning model in monitoring the business user big data to be analyzed during production process monitoring, such as identifying the types of business user big data to be analyzed. The production process monitoring method based on big data analysis disclosed in this embodiment can also perform relative distribution adjustment operations before or after at least one unit based on actual processing requirements. This enables the unit with the added relative distribution adjustment operation to acquire multi-dimensional behavioral descriptions. The machine learning model can be optimized to carry the performance of acquiring multi-dimensional behavioral descriptions, thereby enhancing the production process monitoring performance of the machine learning model for the business user big data to be analyzed.

[0094] In an alternative embodiment, the machine learning model can be optimized before using it to process online user activity information for production process monitoring. This optimization may be performed on at least one unit of the machine learning model, for example, a machine learning model whose relative distribution can be adjusted before or after it. The method may also include the following.

[0095] From several online user activity information in the example business user big data, the example online user activity information is identified; the example online user activity information is loaded into the machine learning model to obtain the test production process monitoring status of the example business user big data; based on the test production process monitoring status and the prior annotations of the example business user big data, the model quality evaluation of the machine learning model is determined; based on the model quality evaluation, the machine learning model is optimized.

[0096] Understandably, accurately identifying exemplary online user activity information can improve the credibility of optimizing machine learning models.

[0097] For example, exemplary business user big data can be used to optimize machine learning models. This involves filtering the exemplary business user big data and parsing the selected online user activity information to obtain exemplary online user activity information. This exemplary online user activity information can then be loaded into the machine learning model for processing. The machine learning model can perform key user behavior description mining, relative distribution adjustment, and other processing on the exemplary online user activity information, and obtain the test production process monitoring status of the exemplary business user big data, such as the types of exemplary business user big data. Furthermore, based on the production process monitoring status obtained by the machine learning model (e.g., the percentage of exemplary business user big data types, which may contain anomalies) and the prior annotations of the exemplary business user big data (e.g., the precise percentage of exemplary business user big data types), a model quality evaluation of the machine learning model can be obtained, and the machine learning model can be optimized through this evaluation.

[0098] For example, optimization can be paused and the optimized machine learning model obtained when the model quality evaluation is no greater than the judgment value or the number of optimizations meets the target value. The optimized machine learning model can be used for business user big data to be analyzed in production process monitoring, such as the types of business user big data to be analyzed in production process monitoring.

[0099] Based on the steps of the production process monitoring method based on big data analysis in the embodiments of this application, the online user activity information of the business user big data to be analyzed can be filtered and parsed to obtain target online user activity information, and key user behavior description mining operation can be performed on the target online user activity information to obtain the first key user behavior description distribution of each target online user activity information.

[0100] In an alternative embodiment, a first relative distribution adjustment operation is performed on the first key user behavior description distribution of each target online user activity information to obtain several sets of first key user behavior description distributions. For example, each set of target online user activity information corresponds to the first key user behavior description distribution of I production process states.

[0101] In an alternative embodiment, the distributions of the first key user behavior descriptions for each production process state can be clustered. For example, each group may include I / M first key user behavior description distributions or production process states. For instance, if the production process state description result I is 128 and the number of target online user activity information M is 16, then each group may cover 16 first key user behavior description distributions, or each group may cover 16 production process states.

[0102] In an alternative embodiment, the first group of first key user behavior description clusters for the first target online user activity information, the first group of first key user behavior description clusters for the second target online user activity information, the first group of first key user behavior description clusters for the third target online user activity information, ..., the first group of first key user behavior description clusters for the sixteenth target online user activity information are concatenated to obtain the first key user behavior description distribution set corresponding to semantic label 1. The second group of first key user behavior description clusters for the first target online user activity information, the second group of first key user behavior description clusters for the second target online user activity information, the second group of first key user behavior description clusters for the third target online user activity information, ..., the second group of first key user behavior description clusters for the sixteenth target online user activity information are concatenated to obtain the first key user behavior description distribution set corresponding to semantic label 2, ... thus obtaining the first key user behavior description distribution set with semantic labels as clustering credentials.

[0103] In an alternative embodiment, key user behavior description mining operations can be performed on each of the first key user behavior description distributions in the first key user behavior description distribution set. For example, key user behavior description mining operations can be performed on each of the first key user behavior description distributions in the first key user behavior description distribution set using three units: 2*2, 6*6, and 2*2, respectively, to obtain a second key user behavior description distribution corresponding to each first key user behavior description distribution. These second key user behavior description distributions can constitute a second key user behavior description distribution set. The second key user behavior description distribution sets corresponding to each first key user behavior description distribution set can be obtained according to the above method.

[0104] In an alternative embodiment, a second relative distribution adjustment operation is performed on the second key user behavior description distribution set. That is, the second relative distribution adjustment operation is performed on the second key user behavior description distribution set based on semantic tags as clustering credentials, and the clustering method of each second key user behavior description distribution in each second key user behavior description distribution set is optimized to use the online user activity information of the target as the clustering credentials to obtain multi-dimensional behavior description content.

[0105] In an alternative embodiment, the first key user behavior description distribution and multi-dimensional behavior description content can be continuously processed by subsequent thread units of the machine learning model, and the types of business user big data to be analyzed can be output.

[0106] Based on the above, please refer to the following: Figure 2 A production process monitoring device 200 based on big data analysis is provided, which is applied to a production process monitoring system based on big data analysis. The device includes:

[0107] The description acquisition module 210 is used to identify several target online user activity information from several online user activity information in the big data of business users to be analyzed; and to perform key user behavior description mining operations on several production process states on the several target online user activity information to obtain the first key user behavior description distribution of several production process states of the target online user activity information.

[0108] The process monitoring module 220 is used to perform multi-dimensional mining operations on the first key user behavior description distribution of the target online user activity information to obtain multi-dimensional behavior description content of the target online user activity information. The multi-dimensional mining operations include relative distribution adjustment operations and key user behavior description mining operations. The relative distribution adjustment operations include relative distribution adjustment of the original key user behavior description distribution and the processing key user behavior description distribution of one or more key user behavior description mining operations. Based on the multi-dimensional behavior description content, the module performs production process monitoring on the business user big data to be analyzed to obtain the production process monitoring status.

[0109] Based on the above, please refer to the following: Figure 3 The present invention illustrates a production process monitoring system 300 based on big data analysis, comprising a processor 310 and a memory 320 that communicate with each other. The processor 310 is used to read computer programs from the memory 320 and execute them to implement the above-mentioned method.

[0110] Based on the above, a computer-readable storage medium is also provided, on which a computer program stored implements the above method during runtime.

[0111] In summary, based on the above scheme and the production process monitoring method based on big data analysis in the embodiments of this application, multi-dimensional behavioral description content is obtained through relative distribution adjustment operations and key user behavior description mining operations. This is beneficial for monitoring online user activity information in the production process. Furthermore, by obtaining multi-dimensional behavioral description content through multi-dimensional mining operations, multi-dimensional behavioral description content can be obtained more accurately, improving the accuracy of mining operations. Thus, multi-dimensional mining operations effectively improve the accuracy and credibility of the description content.

[0112] It should be understood that the systems and modules described above can be implemented in various ways. For example, in some embodiments, the systems and modules can be implemented by hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the methods and systems described above can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The systems and modules of this application can be implemented not only by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., but also by software executed by various types of processors, or by a combination of the aforementioned hardware circuits and software (e.g., firmware).

[0113] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects may be any one or a combination of the above, or any other possible beneficial effects.

[0114] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore remain within the spirit and scope of the exemplary embodiments of this application.

[0115] Furthermore, this application uses specific terms to describe embodiments of the application. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of the application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the application can be appropriately combined.

[0116] Furthermore, those skilled in the art will understand that aspects of this application can be described and illustrated through several patentable types or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Accordingly, aspects of this application can be implemented entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. All of the above hardware or software may be referred to as a “data block,” “module,” “engine,” “unit,” “component,” or “system.” Furthermore, aspects of this application may manifest as a computer product located on one or more computer-readable media, the product including computer-readable program code.

[0117] Computer storage media may contain a propagated data signal containing computer program code, for example, on baseband or as part of a carrier wave. This propagated signal may take various forms, including electromagnetic, optical, and suitable combinations thereof. Computer storage media can be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, apparatus, or device to enable communication, propagation, or transmission of a program for use. The program code located on the computer storage medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above media.

[0118] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages ​​such as C, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages ​​such as Python, Ruby, and Groovy, or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0119] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this application are not intended to limit the order of the processes and methods of this application. Although the foregoing disclosure has discussed some currently considered useful embodiments of the invention through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments of this application. For example, while the system components described above can be implemented using hardware devices, they can also be implemented solely through software solutions, such as installing the described system on existing servers or mobile devices.

[0120] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this disclosure method does not imply that the subject matter of the application requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of the single embodiments disclosed above.

[0121] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of embodiments are modified in some examples with the terms "approximately," "approximately," or "generally." Unless otherwise stated, "approximately," "approximately," or "generally" indicates that the numbers are open to adaptive variation. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may be changed depending on the characteristics required by individual embodiments. In some embodiments, numerical parameters are taken into account a specified number of significant digits and employ a general method of digit reservation. Although the numerical ranges and parameters used to confirm their breadth of application in some embodiments of this application are approximate values, in specific embodiments, such values ​​are set as precisely as feasible.

[0122] For each patent, patent application, patent application publication, and other material such as articles, books, specifications, publications, and documents referenced in this application, the entire contents of that patent are incorporated herein by reference. This excludes historical application documents that are inconsistent with or conflict with the content of this application, as well as documents that limit the broadest scope of the claims in this application (currently or subsequently appended to this application). It should be noted that if there are any inconsistencies or conflicts between the descriptions, definitions, and / or terminology used in the supplementary materials of this application and the content of this application, the descriptions, definitions, and / or terminology used in this application shall prevail.

[0123] Finally, it should be understood that the embodiments described in this application are merely illustrative of the principles of the embodiments of this application. Other modifications may also fall within the scope of this application. Therefore, alternative configurations of the embodiments of this application are considered as examples and not limitations, and are regarded as consistent with the teachings of this application. Accordingly, the embodiments of this application are not limited to the embodiments explicitly described and illustrated in this application.

[0124] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A production process monitoring method based on big data analysis, characterized in that, include: From the online user activity information of the business user big data to be analyzed, several target online user activity information are identified; Perform key user behavior description mining operations on several production process states for the target online user activity information to obtain the first key user behavior description distribution of several production process states of the target online user activity information. A multi-dimensional mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain multi-dimensional behavior description content of the target online user activity information. The multi-dimensional mining operation includes a relative distribution adjustment operation and a key user behavior description mining operation. The relative distribution adjustment operation includes adjusting the relative distribution of the original key user behavior description distribution and the processed key user behavior description distribution of one or more key user behavior description mining operations. Based on the multi-dimensional behavior description content, production process monitoring is performed on the business user big data to be analyzed to obtain the production process monitoring status. A multi-dimensional mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain multi-dimensional behavior description content of the target online user activity information, including: A first relative distribution adjustment operation is performed on the first key user behavior description distribution of the target online user activity information to obtain a plurality of first key user behavior description distribution sets, wherein the order of the first key user behavior description distributions in the first key user behavior description distribution sets is consistent with the order of the corresponding target online user activity information. Perform key user behavior description mining operation on the first key user behavior description distribution in the first key user behavior description distribution set to obtain a second key user behavior description distribution set, wherein the second key user behavior description distribution set includes a second key user behavior description distribution corresponding to the first key user behavior description distribution; A second relative distribution adjustment operation is performed on the second key user behavior description distribution set to obtain the multi-dimensional behavior description content.

2. The method as described in claim 1, characterized in that, A first relative distribution adjustment operation is performed on the first key user behavior description distribution of the target online user activity information to obtain several sets of first key user behavior description distributions, including: For each target online user activity information, several first key user behavior descriptions of the target online user activity information are clustered to obtain several first key user behavior description clusters of the target online user activity information. Each first key user behavior description cluster carries a semantic label, and the first key user behavior description clusters of each target online user activity information carrying the same semantic label correspond to the same group of production process states in several production process states. In the first key user behavior description clusters of each target online user activity information, the first key user behavior description clusters with consistent semantic labels are concatenated based on the chronological relationship of the target online user activity information to obtain the first key user behavior description distribution set corresponding to the semantic labels.

3. The method as described in claim 1, characterized in that, Perform a second relative distribution adjustment operation on the second key user behavior description distribution set to obtain the multi-dimensional behavior description content, including: The second key user behavior description distribution in the second key user behavior description distribution set is clustered based on the corresponding target online user activity information to obtain the second key user behavior description cluster. The second key user behavior description distribution in the same second key user behavior description cluster corresponds to the same target online user activity information. The clusters of each second key user behavior description are concatenated to obtain multi-dimensional behavior descriptions corresponding to the online user activity information of each target.

4. The method as described in claim 1, characterized in that, A multi-dimensional mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain multi-dimensional behavior description content of the target online user activity information, including: A first key user behavior description mining operation is performed on the first key user behavior description distribution of the target online user activity information to obtain a third key user behavior description distribution corresponding to the first key user behavior description distribution. A third relative distribution adjustment operation is performed on the third key user behavior description distribution to obtain several sets of third key user behavior description distributions, wherein the order of the third key user behavior description distributions in the set of third key user behavior description distributions is consistent with the order of the corresponding target online user activity information. Perform a second key user behavior description mining operation on the third key user behavior description distribution in the third key user behavior description distribution set to obtain a fourth key user behavior description distribution set, wherein the fourth key user behavior description distribution set includes a fourth key user behavior description distribution corresponding to the third key user behavior description distribution; A fourth relative distribution adjustment operation is performed on the fourth key user behavior description distribution set to obtain the third key user behavior description cluster corresponding to the target online user activity information; A third key user behavior description mining operation is performed on the fourth key user behavior description distribution in the third key user behavior description cluster to obtain multi-dimensional behavior description content corresponding to the target online user activity information.

5. The method as described in claim 1, characterized in that, From the online user activity information of the business user big data to be analyzed, several target online user activity information were identified, including: The system filters several online user activity information from the business user big data to be analyzed, and obtains several business user big data fragments to be analyzed. The data fragments of business users to be analyzed are subjected to information parsing operations to obtain the activity information of the target online users.

6. The method as described in claim 1, characterized in that, The production process monitoring method based on big data analysis is implemented through a machine learning model, and the method further includes: From the online user activity information of the example business user big data, the example online user activity information was identified; The online user activity information of the example is loaded into the machine learning model to obtain the test production process monitoring status of the big data of the example business users; Based on the prior annotations of the test production process monitoring and the example business user big data, the model quality evaluation of the machine learning model is determined. Based on the model quality evaluation, the machine learning model is optimized.

7. A production process monitoring system based on big data analysis, characterized in that, The method includes a processor and a memory that communicate with each other, the processor being configured to retrieve a computer program from the memory and to implement the method of any one of claims 1-6 by running the computer program.

Images