CN117857311A - Data mining task execution method and system in Internet of things scene - Google Patents

Abstract

The invention provides a data mining task execution method and a system in an Internet of things scene, wherein an edge equipment work log of the Internet of things is analyzed, edge equipment meeting data mining conditions is screened, an optimal network path between the screened edge equipment and a platform end of the Internet of things is constructed, and optimal communication connection between the edge equipment and the platform end is ensured; identifying a data mining task initiated by a platform end, determining attribute information of an applicable algorithm model, determining edge equipment matched with the data mining task, and distributing the data mining task to the matched edge equipment, so that the data mining task is ensured to be executed efficiently preferentially; and based on the real-time execution result information of the data mining task, determining occurrence time information of the occurrence of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution data mining task for continuous processing, realizing flexible switching execution of the data mining task and improving the execution efficiency and reliability of the data mining task.

Description

Data mining task execution method and system in Internet of things scene

Technical Field

The invention relates to the field of data processing of the Internet of things, in particular to a data mining task execution method and system in the scene of the Internet of things.

Background

As an important link of program development, data mining needs high-performance hardware configuration, the existing internet of things platform does not provide a data mining function, only can execute a data mining task by means of special data mining equipment, and the data mining task can only be executed on one data mining equipment from beginning to end, when the data mining equipment fails, the data mining task cannot be normally executed, the execution efficiency and reliability of the data mining task are seriously affected, and therefore the execution time cost of the data mining task is increased, and the execution switching flexibility of the data mining task is reduced.

Disclosure of Invention

The invention aims to provide a data mining task execution method and a system in an Internet of things scene, which analyze an edge equipment work log of the Internet of things, screen edge equipment meeting data mining conditions, construct an optimal network path between the screened edge equipment and a platform end of the Internet of things, and ensure optimal communication connection between the edge equipment and the platform end; identifying a data mining task initiated by a platform end, determining attribute information of an applicable algorithm model, determining edge equipment matched with the data mining task, and distributing the data mining task to the matched edge equipment, so that the data mining task is ensured to be executed efficiently preferentially; and based on the real-time execution result information of the data mining task, determining occurrence time information of the occurrence of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution data mining task for continuous processing, realizing flexible switching execution of the data mining task and improving the execution efficiency and reliability of the data mining task.

The invention is realized by the following technical scheme:

the data mining task execution method in the scene of the Internet of things comprises the following steps:

analyzing the edge equipment work logs of the Internet of things, and screening a plurality of edge equipment meeting the data mining conditions; based on the network positions of all the screened edge devices in the Internet of things, constructing an optimal network path between each screened edge device and a platform end of the Internet of things;

identifying the data mining task initiated by the platform end to obtain the attribute information of the applicable algorithm model corresponding to the data mining task; determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and distributing the data mining task to one of the matched edge devices;

judging whether an abnormal execution event occurs to the data mining task or not based on real-time execution result information of the data mining task; and transferring the data mining task to other edge devices matched with the execution of the data mining task to continue processing based on the occurrence time information of the execution abnormal event.

Optionally, analyzing the edge equipment work log of the Internet of things, and screening a plurality of edge equipment meeting the data mining condition; and constructing an optimal network path between each screened edge device and a platform end of the internet of things based on network positions of all screened edge devices in the internet of things, wherein the optimal network path comprises the following steps:

acquiring the work logs of all the edge devices in the active state in the Internet of things, and analyzing the work logs to obtain the workload of data mining to be processed of the edge devices in the active state; comparing the data mining workload to be processed with a preset workload threshold, and judging that the corresponding edge equipment meets the data mining condition if the data mining workload to be processed is smaller than the preset workload threshold; otherwise, judging that the corresponding edge equipment does not meet the data mining condition;

and acquiring network address information of all edge devices meeting data mining conditions in the Internet of things, and determining a network path with minimum transmission delay between each edge device and a platform end of the Internet of things based on the network address information of each edge device and the network address information of the platform end of the Internet of things, wherein the network path is used as an optimal network path between each edge device and the platform end.

Optionally, identifying a data mining task initiated by the platform end to obtain applicable algorithm model attribute information corresponding to the data mining task; determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and distributing the data mining task to one of the matched edge devices, wherein the method comprises the following steps:

identifying an original data code contained in a data mining task initiated by the platform end to obtain a code structure and code type information of the original data code; determining applicable algorithm model type information matched with the data mining task based on the code structure and the code type information;

based on the applicable algorithm model type information, determining all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices as edge devices which are matched with the data mining task; and then distributing the data mining task to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located.

Optionally, based on real-time execution result information of the data mining task, judging whether an abnormal event of execution occurs in the data mining task; based on the occurrence time information of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution of the data mining task for continuous processing, wherein the method comprises the following steps:

Acquiring real-time execution result information corresponding to the data mining task in each execution stage, analyzing the real-time execution result information, and judging whether an error data result exists in the real-time execution result; if yes, judging that the data mining task generates an execution abnormal event in a corresponding execution stage; if the data mining task does not exist, judging that the data mining task does not have an abnormal execution event in the corresponding execution stage;

based on the execution time of the execution stage of the execution abnormal event, transferring all execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task; the algorithm operation accuracy of the algorithm model matched with the data mining task and contained in other edge devices matched with the data mining task is only smaller than that of the algorithm model corresponding to the edge device currently executing the data mining task.

The data mining task execution system in the scene of the Internet of things comprises:

the edge equipment screening module is used for analyzing the edge equipment work logs of the Internet of things and screening a plurality of edge equipment meeting the data mining conditions;

The network path construction module is used for constructing an optimal network path between each screened edge device and the platform end of the Internet of things based on the network positions of all the screened edge devices in the Internet of things;

the algorithm model identification module is used for identifying the data mining task initiated by the platform end to obtain the applicable algorithm model attribute information corresponding to the data mining task;

the data mining task allocation module is used for determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and allocating the data mining task to one of the matched edge devices;

the task execution state identification module is used for judging whether the data mining task generates an execution abnormal event or not based on the real-time execution result information of the data mining task;

and the data mining task transfer module is used for transferring the data mining task to other edge equipment matched with the execution of the data mining task to continue processing based on the occurrence time information of the execution abnormal event.

Optionally, the edge device screening module is configured to analyze an edge device work log of the internet of things, and screen a plurality of edge devices that meet a data mining condition, including:

Acquiring the work logs of all the edge devices in the active state in the Internet of things, and analyzing the work logs to obtain the workload of data mining to be processed of the edge devices in the active state; comparing the data mining workload to be processed with a preset workload threshold, and judging that the corresponding edge equipment meets the data mining condition if the data mining workload to be processed is smaller than the preset workload threshold; otherwise, judging that the corresponding edge equipment does not meet the data mining condition;

the network path construction module is configured to construct an optimal network path between each screened edge device and a platform end of the internet of things based on network positions of all screened edge devices in the internet of things, and includes:

and acquiring network address information of all edge devices meeting data mining conditions in the Internet of things, and determining a network path with minimum transmission delay between each edge device and a platform end of the Internet of things based on the network address information of each edge device and the network address information of the platform end of the Internet of things, wherein the network path is used as an optimal network path between each edge device and the platform end.

Optionally, the algorithm model identifying module is configured to identify a data mining task initiated by the platform end, and obtain applicable algorithm model attribute information corresponding to the data mining task, where the applicable algorithm model attribute information includes:

identifying an original data code contained in a data mining task initiated by the platform end to obtain a code structure and code type information of the original data code; determining applicable algorithm model type information matched with the data mining task based on the code structure and the code type information;

the data mining task allocation module is configured to determine, based on the applicable algorithm model attribute information, a plurality of edge devices that match and execute the data mining task from all the screened edge devices, and allocate the data mining task to one of the matched edge devices, including:

based on the applicable algorithm model type information, determining all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices as edge devices which are matched with the data mining task; and then distributing the data mining task to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located.

Optionally, the task execution state identifying module is configured to determine, based on real-time execution result information of the data mining task, whether an execution exception event occurs in the data mining task, including:

acquiring real-time execution result information corresponding to the data mining task in each execution stage, analyzing the real-time execution result information, and judging whether an error data result exists in the real-time execution result; if yes, judging that the data mining task generates an execution abnormal event in a corresponding execution stage; if the data mining task does not exist, judging that the data mining task does not have an abnormal execution event in the corresponding execution stage;

the data mining task transferring module is configured to transfer the data mining task to other edge devices that execute the data mining task in a matching manner based on occurrence time information of the execution exception event, where the processing is further performed, and includes:

based on the execution time of the execution stage of the execution abnormal event, transferring all execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task; the algorithm operation accuracy of the algorithm model matched with the data mining task and contained in other edge devices matched with the data mining task is only smaller than that of the algorithm model corresponding to the edge device currently executing the data mining task.

Compared with the prior art, the invention has the following beneficial effects:

according to the data mining task execution method and system in the Internet of things scene, the edge equipment working logs of the Internet of things are analyzed, edge equipment meeting data mining conditions is screened, an optimal network path between the screened edge equipment and the platform end of the Internet of things is constructed, and optimal communication connection between the edge equipment and the platform end is ensured; identifying a data mining task initiated by a platform end, determining attribute information of an applicable algorithm model, determining edge equipment matched with the data mining task, and distributing the data mining task to the matched edge equipment, so that the data mining task is ensured to be executed efficiently preferentially; and based on the real-time execution result information of the data mining task, determining occurrence time information of the occurrence of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution data mining task for continuous processing, realizing flexible switching execution of the data mining task and improving the execution efficiency and reliability of the data mining task.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. Wherein:

Fig. 1 is a flow chart of a data mining task execution method in an internet of things scene.

Fig. 2 is a schematic structural diagram of a data mining task execution system in an internet of things scenario provided by the invention.

Detailed Description

In order to make the above objects, features and advantages of the present application more comprehensible, embodiments accompanied with figures are described in detail below. It is to be understood that the specific embodiments described herein are for purposes of illustration only and are not limiting. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present application are shown in the drawings. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.

The terms "comprising" and "having" and any variations thereof herein are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.

Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present application. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.

Referring to fig. 1, the method for executing a data mining task in an internet of things scenario provided in an embodiment of the present application includes:

analyzing the edge equipment work logs of the Internet of things, and screening a plurality of edge equipment meeting the data mining conditions; based on the network positions of all the screened edge devices in the Internet of things, constructing an optimal network path between each screened edge device and a platform end of the Internet of things;

identifying a data mining task initiated by the platform end to obtain attribute information of an applicable algorithm model corresponding to the data mining task; determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and distributing the data mining task to one of the matched edge devices;

Judging whether the data mining task generates an execution abnormal event or not based on real-time execution result information of the data mining task; and transferring the data mining task to other edge devices matched with the execution of the data mining task to continue processing based on the occurrence time information of the execution abnormal event.

The method for executing the data mining task in the scene of the Internet of things has the advantages that the method for executing the data mining task in the scene of the Internet of things analyzes the working log of the edge equipment of the Internet of things, screens the edge equipment meeting the data mining condition, constructs the optimal network path between the screened edge equipment and the platform end of the Internet of things, and ensures the optimal communication connection between the edge equipment and the platform end; identifying a data mining task initiated by a platform end, determining attribute information of an applicable algorithm model, determining edge equipment matched with the data mining task, and distributing the data mining task to the matched edge equipment, so that the data mining task is ensured to be executed efficiently preferentially; and based on the real-time execution result information of the data mining task, determining occurrence time information of the occurrence of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution data mining task for continuous processing, realizing flexible switching execution of the data mining task and improving the execution efficiency and reliability of the data mining task.

In another embodiment, an edge device work log of the internet of things is analyzed, and a plurality of edge devices meeting data mining conditions are screened; and constructing an optimal network path between each screened edge device and a platform end of the internet of things based on the network positions of all the screened edge devices in the internet of things, wherein the optimal network path comprises:

acquiring the work logs of all the edge devices in the active state in the Internet of things, and analyzing the work logs to obtain the workload of data mining to be processed of the edge devices in the active state; comparing the data mining workload to be processed with a preset workload threshold, and judging that the corresponding edge equipment meets the data mining condition if the data mining workload to be processed is smaller than the preset workload threshold; otherwise, judging that the corresponding edge equipment does not meet the data mining condition;

and acquiring network address information of all edge devices meeting the data mining condition in the Internet of things, and determining a network path with minimum transmission delay between each edge device and a platform end of the Internet of things based on the network address information of each edge device and the network address information of the platform end of the Internet of things, wherein the network path is used as an optimal network path between each edge device and the platform end.

The Internet of things has the beneficial effects that the Internet of things comprises the platform end and a plurality of edge devices, the platform end is used for being connected with an external terminal and used for receiving data mining tasks uploaded by the external terminal, and then the data mining tasks are distributed to the corresponding edge devices, so that the edge devices can execute the corresponding data mining tasks, and the edge devices execute the data mining tasks by utilizing an algorithm model in the edge devices. The types of algorithm models inside different edge devices are different, and the available computing capacities of different edge devices are correspondingly different. In order to ensure that the data mining task can be distributed to the proper edge equipment, the work logs of all the edge equipment in an active state in the Internet of things are analyzed to obtain the to-be-processed data mining workload of the edge equipment in the active state (namely, the incomplete data mining workload in the data mining task received by the edge equipment); and comparing the data mining workload to be processed with a preset workload threshold value, and judging whether the edge equipment meets corresponding data mining conditions, so as to perform primary screening on the edge equipment in an active state in the Internet of things. And determining a network path with minimum transmission delay between each edge device and the platform end of the Internet of things based on the network address information of all edge devices meeting the data mining conditions inside the Internet of things and the network address information of the platform end of the Internet of things, so that when the platform end distributes data mining tasks to the edge devices through the corresponding network paths, the situation that delay distortion occurs to the data mining tasks is avoided.

In another embodiment, identifying a data mining task initiated by the platform end to obtain applicable algorithm model attribute information corresponding to the data mining task; determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and distributing the data mining task to one of the matched edge devices, wherein the method comprises the following steps:

identifying an original data code contained in a data mining task initiated by the platform end to obtain a code structure and code type information of the original data code; determining applicable algorithm model type information matched with the data mining task based on the code structure and the code type information;

based on the applicable algorithm model type information, determining all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices as edge devices which are matched with the data mining task; and then the data mining task is distributed to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located.

The method has the advantages that the original data codes contained in the data mining task initiated by the platform end are identified, the code structure and the code type information of the original data codes are determined, and the code structure and the code type information are used as references to determine the type information of the applicable algorithm model matched with the data mining task, so that the data mining task can be conveniently and accurately selected by a proper type of algorithm model to be executed and processed. In addition, all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices are determined to be matched with the edge device which executes the data mining task based on the type information of the applicable algorithm model; and then the data mining task is distributed to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located, so that the data mining task is preferentially distributed to the edge equipment which contains the corresponding algorithm model and can carry out the highest algorithm operation accuracy, and the optimal execution processing of the data mining task is ensured.

In another embodiment, based on real-time execution result information of the data mining task, judging whether an execution abnormal event occurs to the data mining task; based on the occurrence time information of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution of the data mining task for processing, wherein the method comprises the following steps of:

acquiring real-time execution result information corresponding to the data mining task in each execution stage, analyzing the real-time execution result information, and judging whether an error data result exists in the real-time execution result; if yes, judging that the data mining task generates an execution abnormal event in a corresponding execution stage; if the data mining task does not exist, judging that the data mining task does not have an abnormal execution event in the corresponding execution stage;

based on the execution time of the execution stage of the execution abnormal event, transferring all execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task; the algorithm operation accuracy of the algorithm model matched with the data mining task and contained in other edge devices matched with the data mining task is only smaller than that of the algorithm model corresponding to the edge device currently executing the data mining task.

The method and the device have the advantages that the corresponding real-time execution result information of the data mining task in each execution stage is obtained and analyzed, whether the real-time execution result of each execution stage has an error data result or not is judged, the execution correctness of the data mining task in each execution stage is conveniently identified, whether an execution abnormal event occurs in each execution stage or not is judged, and therefore the execution problem of each execution stage is timely and accurately found. And based on the execution time of the execution stage of the execution abnormal event, transferring all the execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task, so that the corresponding edge devices after transfer can keep the execution result of the data mining task correctly executed before, and continue to execute the transferred data mining task without processing the data mining task again, and the execution efficiency and reliability of the data mining task are improved.

Referring to fig. 2, a system for executing a data mining task in an internet of things scenario according to an embodiment of the present application includes:

The edge equipment screening module is used for analyzing the edge equipment work logs of the Internet of things and screening a plurality of edge equipment meeting the data mining conditions;

the network path construction module is used for constructing an optimal network path between each screened edge device and a platform end of the Internet of things based on the network positions of all screened edge devices in the Internet of things;

the algorithm model identification module is used for identifying the data mining task initiated by the platform end to obtain the applicable algorithm model attribute information corresponding to the data mining task;

the data mining task allocation module is used for determining a plurality of edge devices which are matched and execute the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and allocating the data mining task to one of the matched edge devices;

the task execution state identification module is used for judging whether the data mining task generates an execution abnormal event or not based on the real-time execution result information of the data mining task;

and the data mining task transferring module is used for transferring the data mining task to other edge equipment matched with the execution of the data mining task to continue processing based on the occurrence time information of the execution abnormal event.

The data mining task execution system in the scene of the Internet of things analyzes the edge equipment work log of the Internet of things, screens edge equipment meeting the data mining condition, constructs an optimal network path between the screened edge equipment and the platform end of the Internet of things, and ensures optimal communication connection between the edge equipment and the platform end; identifying a data mining task initiated by a platform end, determining attribute information of an applicable algorithm model, determining edge equipment matched with the data mining task, and distributing the data mining task to the matched edge equipment, so that the data mining task is ensured to be executed efficiently preferentially; and based on the real-time execution result information of the data mining task, determining occurrence time information of the occurrence of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution data mining task for continuous processing, realizing flexible switching execution of the data mining task and improving the execution efficiency and reliability of the data mining task.

In another embodiment, the edge device screening module is configured to analyze an edge device work log of the internet of things, and screen a plurality of edge devices that meet a data mining condition, including:

Acquiring the work logs of all the edge devices in the active state in the Internet of things, and analyzing the work logs to obtain the workload of data mining to be processed of the edge devices in the active state; comparing the data mining workload to be processed with a preset workload threshold, and judging that the corresponding edge equipment meets the data mining condition if the data mining workload to be processed is smaller than the preset workload threshold; otherwise, judging that the corresponding edge equipment does not meet the data mining condition;

the network path construction module is used for constructing an optimal network path between each screened edge device and a platform end of the internet of things based on the network positions of all screened edge devices in the internet of things, and comprises the following steps:

and acquiring network address information of all edge devices meeting the data mining condition in the Internet of things, and determining a network path with minimum transmission delay between each edge device and a platform end of the Internet of things based on the network address information of each edge device and the network address information of the platform end of the Internet of things, wherein the network path is used as an optimal network path between each edge device and the platform end.

The Internet of things has the beneficial effects that the Internet of things comprises the platform end and a plurality of edge devices, the platform end is used for being connected with an external terminal and used for receiving data mining tasks uploaded by the external terminal, and then the data mining tasks are distributed to the corresponding edge devices, so that the edge devices can execute the corresponding data mining tasks, and the edge devices execute the data mining tasks by utilizing an algorithm model in the edge devices. The types of algorithm models inside different edge devices are different, and the available computing capacities of different edge devices are correspondingly different. In order to ensure that the data mining task can be distributed to the proper edge equipment, the work logs of all the edge equipment in an active state in the Internet of things are analyzed to obtain the to-be-processed data mining workload of the edge equipment in the active state (namely, the incomplete data mining workload in the data mining task received by the edge equipment); and comparing the data mining workload to be processed with a preset workload threshold value, and judging whether the edge equipment meets corresponding data mining conditions, so as to perform primary screening on the edge equipment in an active state in the Internet of things. And determining a network path with minimum transmission delay between each edge device and the platform end of the Internet of things based on the network address information of all edge devices meeting the data mining conditions inside the Internet of things and the network address information of the platform end of the Internet of things, so that when the platform end distributes data mining tasks to the edge devices through the corresponding network paths, the situation that delay distortion occurs to the data mining tasks is avoided.

In another embodiment, the algorithm model identification module is configured to identify a data mining task initiated by the platform end, and obtain applicable algorithm model attribute information corresponding to the data mining task, where the applicable algorithm model attribute information includes:

identifying an original data code contained in a data mining task initiated by the platform end to obtain a code structure and code type information of the original data code; determining applicable algorithm model type information matched with the data mining task based on the code structure and the code type information;

the data mining task allocation module is configured to determine, based on the applicable algorithm model attribute information, a plurality of edge devices that match and execute the data mining task from all the screened edge devices, and allocate the data mining task to one of the matched edge devices, including:

based on the applicable algorithm model type information, determining all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices as edge devices which are matched with the data mining task; and then the data mining task is distributed to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located.

The method has the advantages that the original data codes contained in the data mining task initiated by the platform end are identified, the code structure and the code type information of the original data codes are determined, and the code structure and the code type information are used as references to determine the type information of the applicable algorithm model matched with the data mining task, so that the data mining task can be conveniently and accurately selected by a proper type of algorithm model to be executed and processed. In addition, all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices are determined to be matched with the edge device which executes the data mining task based on the type information of the applicable algorithm model; and then the data mining task is distributed to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located, so that the data mining task is preferentially distributed to the edge equipment which contains the corresponding algorithm model and can carry out the highest algorithm operation accuracy, and the optimal execution processing of the data mining task is ensured.

In another embodiment, the task execution status identifying module is configured to determine, based on real-time execution result information of the data mining task, whether an execution exception event occurs in the data mining task, including:

acquiring real-time execution result information corresponding to the data mining task in each execution stage, analyzing the real-time execution result information, and judging whether an error data result exists in the real-time execution result; if yes, judging that the data mining task generates an execution abnormal event in a corresponding execution stage; if the data mining task does not exist, judging that the data mining task does not have an abnormal execution event in the corresponding execution stage;

the data mining task transferring module is configured to transfer the data mining task to other edge devices that are matched to execute the data mining task to continue processing based on occurrence time information of the execution exception event, where the data mining task transferring module includes:

based on the execution time of the execution stage of the execution abnormal event, transferring all execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task; the algorithm operation accuracy of the algorithm model matched with the data mining task and contained in other edge devices matched with the data mining task is only smaller than that of the algorithm model corresponding to the edge device currently executing the data mining task.

The method and the device have the advantages that the corresponding real-time execution result information of the data mining task in each execution stage is obtained and analyzed, whether the real-time execution result of each execution stage has an error data result or not is judged, the execution correctness of the data mining task in each execution stage is conveniently identified, whether an execution abnormal event occurs in each execution stage or not is judged, and therefore the execution problem of each execution stage is timely and accurately found. And based on the execution time of the execution stage of the execution abnormal event, transferring all the execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task, so that the corresponding edge devices after transfer can keep the execution result of the data mining task correctly executed before, and continue to execute the transferred data mining task without processing the data mining task again, and the execution efficiency and reliability of the data mining task are improved.

In general, the data mining task execution method and the system in the Internet of things scene analyze the edge equipment work log of the Internet of things, screen edge equipment meeting the data mining conditions, and construct an optimal network path between the screened edge equipment and the platform end of the Internet of things, so as to ensure optimal communication connection between the edge equipment and the platform end; identifying a data mining task initiated by a platform end, determining attribute information of an applicable algorithm model, determining edge equipment matched with the data mining task, and distributing the data mining task to the matched edge equipment, so that the data mining task is ensured to be executed efficiently preferentially; and based on the real-time execution result information of the data mining task, determining occurrence time information of the occurrence of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution data mining task for continuous processing, realizing flexible switching execution of the data mining task and improving the execution efficiency and reliability of the data mining task.

The foregoing is merely one specific embodiment of the invention, and any modifications made in light of the above teachings are intended to fall within the scope of the invention.

Claims (8)

1. The method for executing the data mining task in the scene of the Internet of things is characterized by comprising the following steps:

analyzing the edge equipment work logs of the Internet of things, and screening a plurality of edge equipment meeting the data mining conditions; based on the network positions of all the screened edge devices in the Internet of things, constructing an optimal network path between each screened edge device and a platform end of the Internet of things; identifying the data mining task initiated by the platform end to obtain the attribute information of the applicable algorithm model corresponding to the data mining task; determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and distributing the data mining task to one of the matched edge devices;

judging whether an abnormal execution event occurs to the data mining task or not based on real-time execution result information of the data mining task; and transferring the data mining task to other edge devices matched with the execution of the data mining task to continue processing based on the occurrence time information of the execution abnormal event.

2. The method for executing the data mining task in the internet of things scene as set forth in claim 1, wherein:

analyzing the edge equipment work logs of the Internet of things, and screening a plurality of edge equipment meeting the data mining conditions; and constructing an optimal network path between each screened edge device and a platform end of the internet of things based on network positions of all screened edge devices in the internet of things, wherein the optimal network path comprises the following steps:

acquiring the work logs of all the edge devices in the active state in the Internet of things, and analyzing the work logs to obtain the workload of data mining to be processed of the edge devices in the active state; comparing the data mining workload to be processed with a preset workload threshold, and judging that the corresponding edge equipment meets the data mining condition if the data mining workload to be processed is smaller than the preset workload threshold; otherwise, judging that the corresponding edge equipment does not meet the data mining condition;

and acquiring network address information of all edge devices meeting data mining conditions in the Internet of things, and determining a network path with minimum transmission delay between each edge device and a platform end of the Internet of things based on the network address information of each edge device and the network address information of the platform end of the Internet of things, wherein the network path is used as an optimal network path between each edge device and the platform end.

3. The method for executing the data mining task in the internet of things scene as set forth in claim 1, wherein:

identifying the data mining task initiated by the platform end to obtain the attribute information of the applicable algorithm model corresponding to the data mining task; determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and distributing the data mining task to one of the matched edge devices, wherein the method comprises the following steps:

identifying an original data code contained in a data mining task initiated by the platform end to obtain a code structure and code type information of the original data code; determining applicable algorithm model type information matched with the data mining task based on the code structure and the code type information; based on the applicable algorithm model type information, determining all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices as edge devices which are matched with the data mining task; and then distributing the data mining task to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located.

4. The method for executing the data mining task in the internet of things scene as set forth in claim 1, wherein:

judging whether an abnormal execution event occurs to the data mining task or not based on real-time execution result information of the data mining task; based on the occurrence time information of the execution abnormal event, transferring the data mining task to other edge devices matched with the execution of the data mining task for continuous processing, wherein the method comprises the following steps:

acquiring real-time execution result information corresponding to the data mining task in each execution stage, analyzing the real-time execution result information, and judging whether an error data result exists in the real-time execution result; if yes, judging that the data mining task generates an execution abnormal event in a corresponding execution stage; if the data mining task does not exist, judging that the data mining task does not have an abnormal execution event in the corresponding execution stage;

based on the execution time of the execution stage of the execution abnormal event, transferring all execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task; the algorithm operation accuracy of the algorithm model matched with the data mining task and contained in other edge devices matched with the data mining task is only smaller than that of the algorithm model corresponding to the edge device currently executing the data mining task.

5. The data mining task execution system in the scene of the Internet of things is characterized by comprising:

the edge equipment screening module is used for analyzing the edge equipment work logs of the Internet of things and screening a plurality of edge equipment meeting the data mining conditions;

the network path construction module is used for constructing an optimal network path between each screened edge device and the platform end of the Internet of things based on the network positions of all the screened edge devices in the Internet of things;

the algorithm model identification module is used for identifying the data mining task initiated by the platform end to obtain the applicable algorithm model attribute information corresponding to the data mining task;

the data mining task allocation module is used for determining a plurality of edge devices matched with the data mining task from all the screened edge devices based on the attribute information of the applicable algorithm model, and allocating the data mining task to one of the matched edge devices;

the task execution state identification module is used for judging whether the data mining task generates an execution abnormal event or not based on the real-time execution result information of the data mining task;

and the data mining task transfer module is used for transferring the data mining task to other edge equipment matched with the execution of the data mining task to continue processing based on the occurrence time information of the execution abnormal event.

6. The data mining task execution system in an internet of things scenario of claim 5, wherein:

the edge equipment screening module is used for analyzing the edge equipment work logs of the Internet of things, screening a plurality of edge equipment meeting the data mining condition, and comprises:

acquiring the work logs of all the edge devices in the active state in the Internet of things, and analyzing the work logs to obtain the workload of data mining to be processed of the edge devices in the active state; comparing the data mining workload to be processed with a preset workload threshold, and judging that the corresponding edge equipment meets the data mining condition if the data mining workload to be processed is smaller than the preset workload threshold; otherwise, judging that the corresponding edge equipment does not meet the data mining condition;

the network path construction module is configured to construct an optimal network path between each screened edge device and a platform end of the internet of things based on network positions of all screened edge devices in the internet of things, and includes:

and acquiring network address information of all edge devices meeting data mining conditions in the Internet of things, and determining a network path with minimum transmission delay between each edge device and a platform end of the Internet of things based on the network address information of each edge device and the network address information of the platform end of the Internet of things, wherein the network path is used as an optimal network path between each edge device and the platform end.

7. The data mining task execution system in an internet of things scenario of claim 5, wherein:

the algorithm model identification module is used for identifying the data mining task initiated by the platform end to obtain the applicable algorithm model attribute information corresponding to the data mining task, and comprises the following steps:

identifying an original data code contained in a data mining task initiated by the platform end to obtain a code structure and code type information of the original data code; determining applicable algorithm model type information matched with the data mining task based on the code structure and the code type information; the data mining task allocation module is configured to determine, based on the applicable algorithm model attribute information, a plurality of edge devices that match and execute the data mining task from all the screened edge devices, and allocate the data mining task to one of the matched edge devices, including:

based on the applicable algorithm model type information, determining all edge devices which contain the algorithm model of the corresponding type in all the screened edge devices as edge devices which are matched with the data mining task; and then distributing the data mining task to the edge equipment where the algorithm model with the highest algorithm operation accuracy is located.

8. The data mining task execution system in an internet of things scenario of claim 5, wherein:

the task execution state identification module is configured to determine, based on real-time execution result information of the data mining task, whether an execution exception event occurs in the data mining task, including:

acquiring real-time execution result information corresponding to the data mining task in each execution stage, analyzing the real-time execution result information, and judging whether an error data result exists in the real-time execution result; if yes, judging that the data mining task generates an execution abnormal event in a corresponding execution stage; if the data mining task does not exist, judging that the data mining task does not have an abnormal execution event in the corresponding execution stage;

the data mining task transferring module is configured to transfer the data mining task to other edge devices that execute the data mining task in a matching manner based on occurrence time information of the execution exception event, where the processing is further performed, and includes:

based on the execution time of the execution stage of the execution abnormal event, transferring all execution result information of the data mining task before the execution time to other edge devices matched with the execution of the data mining task, so as to continue to process the data mining task; the algorithm operation accuracy of the algorithm model matched with the data mining task and contained in other edge devices matched with the data mining task is only smaller than that of the algorithm model corresponding to the edge device currently executing the data mining task.