CN111240874B

CN111240874B - Data processing method and electronic equipment

Info

Publication number: CN111240874B
Application number: CN202010006318.6A
Authority: CN
Inventors: 朱砡赐; 罗德里戈·罗莎; 保罗·奥利维拉; 胡长建
Original assignee: Lenovo Beijing Ltd
Current assignee: Lenovo Beijing Ltd
Priority date: 2020-01-03
Filing date: 2020-01-03
Publication date: 2022-01-14
Anticipated expiration: 2040-01-03
Also published as: CN111240874A

Abstract

The embodiment of the application provides a data processing method and electronic equipment, wherein the data processing method comprises the following steps: obtaining first diagnostic data characterizing an analysis result of a user on a target event and second diagnostic data characterizing an analysis result of a machine on the target event, the second diagnostic data being different from the first diagnostic data; determining, from the first and second diagnostic data, condition data to which the machine refers the second diagnostic data rather than the first diagnostic data; and outputting the condition data. The data processing method can acquire the condition data referred to during machine diagnosis and output the condition data for the user to refer to.

Description

Data processing method and electronic equipment

Technical Field

The embodiment of the application relates to the field of intelligent equipment, in particular to a data processing method and electronic equipment.

Background

Equipment maintenance, particularly in the IT industry, is a very important department that requires maintenance engineers to be able to quickly and accurately diagnose the crux of the equipment and to perform correct maintenance to ensure that the equipment can operate normally. However, since the diagnosis and maintenance of the fault of the equipment are manually performed by a maintenance engineer, it is difficult to ensure the uniformity of the diagnosis and maintenance standards and levels. At present, many manufacturers adopt machines to diagnose the symptom of the equipment in order to solve the technical problem, but the precision is poor, when the obtained diagnosis data is wrong, an operator does not know the reason of the machine diagnosis mistake, and the diagnosis program cannot be refined, so that the technical problem cannot be solved well.

Content of application

The embodiment of the application provides a data processing method capable of obtaining condition data referred to in machine diagnosis and electronic equipment applying the method.

In order to solve the above technical problem, an embodiment of the present application provides a data processing method, including:

obtaining first diagnostic data characterizing an analysis result of a user on a target event and second diagnostic data characterizing an analysis result of a machine on the target event, the second diagnostic data being different from the first diagnostic data;

determining, from the first and second diagnostic data, condition data to which the machine refers the second diagnostic data rather than the first diagnostic data;

and outputting the condition data.

Preferably, the method further comprises the following steps:

obtaining judgment data representing the acceptance of the user to the condition data;

actual diagnostic data for the target event is determined based on the decision data, which may be at least the first diagnostic data or the second diagnostic data.

Preferably, the determining, from the first diagnostic data and second diagnostic data, each condition data to which the machine derives the second diagnostic data instead of the first diagnostic data comprises:

determining first condition data related to the first diagnostic data information in the analysis result of the machine based on the first diagnostic data;

second condition data related to the second diagnostic data information in the analysis results of the machine is determined based on the second diagnostic data.

Preferably, a plurality of diagnostic models are built in the machine and are respectively used for analyzing the target events to obtain different diagnostic data, and training data of each diagnostic model are at least partially different;

the determining the first and second condition data comprises:

determining a first diagnostic model in the machine for calculating that the target event has first diagnostic data based on the first diagnostic data;

determining a second diagnostic model in the machine for calculating that the target event has second diagnostic data based on the second diagnostic data;

determining a first significant factor based on the first diagnostic model and defining the first significant factor as the first condition data;

determining a second significance factor based on the second diagnostic model and defining the second significance factor as the second condition data.

Preferably, the obtaining of the determination data representing the user's acceptance of the condition data includes:

obtaining first judgment data, wherein the first judgment data represents the proportion of the sub data identified by the user in the first condition data to all the sub data;

and obtaining second judgment data, wherein the second judgment data represents the proportion of the user to all the subdata which is identified in the second condition data.

Preferably, the determining actual diagnostic data of the target event based on the decision data comprises:

determining that the first decision data is less than a first threshold;

determining that the second determination data is less than a second threshold;

determining the first diagnostic data to be actual diagnostic data; or

Determining that the first decision data is not less than a first threshold;

determining that the second determination data is not less than a second threshold;

the second diagnostic data is determined to be actual diagnostic data.

Preferably, the first determination data is determined to be smaller than a first threshold;

determining the magnitude relation of the first judgment data and the second judgment data;

and determining the first diagnostic data or the second diagnostic data corresponding to the first judgment data or the second judgment data with large values as actual diagnostic data.

determining that the first decision data is less than a first threshold, or that the second decision data is less than a second threshold;

processing the first condition data or the second condition data, and performing clustering calculation on a processing result;

determining a reference event in the historical data, which is similar to the target event, and diagnosis data thereof based on the calculation result;

outputting the reference event and the diagnosis data thereof to enable a user to judge the first condition data or the second condition data again based on the reference event and the diagnosis data thereof, and determining actual diagnosis data of the target event based on the new judgment data.

determining that the first decision data is not less than a first threshold;

determining that the second determination data is less than a second threshold;

and re-determining the weight of the second diagnostic model by using an analytic hierarchy process based on the second condition data and the second determination data, and updating the second diagnostic model.

The embodiment of the present application provides an electronic equipment simultaneously, wherein, include:

an obtaining module for obtaining first diagnostic data characterizing an analysis result of a user on a target event and second diagnostic data characterizing an analysis result of a machine on the target event, the second diagnostic data being different from the first diagnostic data;

a processing module for determining, from the first and second diagnostic data, condition data to which the machine refers the second diagnostic data instead of the first diagnostic data.

Based on the disclosure of the above embodiments, it can be known that the embodiments of the present application have the beneficial effects that when the diagnostic data of the user and the machine for the target syndrome are inconsistent, the second diagnostic data of the machine is determined based on the first diagnostic data of the user and the second diagnostic data of the machine, instead of the condition data referred by the first diagnostic data, and the condition data is output for the user to refer to, so that the user knows the reason why the machine makes the second diagnosis.

Drawings

Fig. 1 is a flowchart of a data processing method in an embodiment of the present application.

Fig. 2 is a flowchart of a data processing method in another embodiment of the present application.

Fig. 3 is a flowchart of a data processing method in another embodiment of the present application.

Fig. 4 is a flowchart of a data processing method in another embodiment of the present application.

Fig. 5 is a flowchart of a data processing method in another embodiment of the present application.

Fig. 6 is a flowchart of a data processing method in another embodiment of the present application.

FIG. 7 is a flowchart of a method implemented in the embodiments of the present application.

Fig. 8 is a block diagram of an electronic device in the embodiment of the present application.

Detailed Description

Specific embodiments of the present application will be described in detail below with reference to the accompanying drawings, but the present application is not limited thereto.

It will be understood that various modifications may be made to the embodiments disclosed herein. The following description is, therefore, not to be taken in a limiting sense, but is made merely as an exemplification of embodiments. Other modifications will occur to those skilled in the art within the scope and spirit of the disclosure.

The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the disclosure and, together with a general description of the disclosure given above, and the detailed description of the embodiments given below, serve to explain the principles of the disclosure.

These and other characteristics of the present application will become apparent from the following description of preferred forms of embodiment, given as non-limiting examples, with reference to the attached drawings.

It should also be understood that, although the present application has been described with reference to some specific examples, a person of skill in the art shall certainly be able to achieve many other equivalent forms of application, having the characteristics as set forth in the claims and hence all coming within the field of protection defined thereby.

The above and other aspects, features and advantages of the present disclosure will become more apparent in view of the following detailed description when taken in conjunction with the accompanying drawings.

Specific embodiments of the present disclosure are described hereinafter with reference to the accompanying drawings; however, it is to be understood that the disclosed embodiments are merely examples of the disclosure that may be embodied in various forms. Well-known and/or repeated functions and structures have not been described in detail so as not to obscure the present disclosure with unnecessary or unnecessary detail. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a basis for the claims and as a representative basis for teaching one skilled in the art to variously employ the present disclosure in virtually any appropriately detailed structure.

The specification may use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which may each refer to one or more of the same or different embodiments in accordance with the disclosure.

Hereinafter, embodiments of the present application will be described in detail with reference to the accompanying drawings.

As shown in fig. 1, an embodiment of the present application discloses a data processing method, including:

obtaining first diagnostic data representing an analysis result of a user on a target event and second diagnostic data representing an analysis result of a machine on the target event, the second diagnostic data being different from the first diagnostic data;

determining condition data referred by the second diagnostic data but not the first diagnostic data according to the first diagnostic data and the second diagnostic data;

and outputting each condition data.

For example, when a component of an automobile (corresponding to a target event) is faulty or a functional component in a computer is faulty and a maintenance engineer (corresponding to a user) intends to perform fault diagnosis on the faulty component, status data of the faulty component, data on a problem phenomenon, and the like may be input to a machine, and the machine may perform diagnosis to obtain second diagnosis data, and the maintenance engineer may perform diagnosis on the faulty component to obtain first diagnosis data. When the maintenance engineer finds that the first diagnostic data and the second diagnostic data are different, the first diagnostic data can be input into the machine, or the maintenance engineer directly inputs the first diagnostic data into the machine after making the diagnostic data, and the machine makes a judgment that the first diagnostic data is different from the second diagnostic data. When the machine obtains two diagnostic data and knows that the two diagnostic data are different, the machine may analyze the two diagnostic data, e.g., the machine may analyze the two diagnostic data to determine a faulty knot, e.g., the first diagnostic data indicates that the faulty component's knot is at knot a and the second diagnostic data indicates that the faulty component's knot is at knot B, and then the machine may determine the condition data it references when it obtains the second diagnostic data, i.e., the condition data that is referenced when it determines that the knot is at B, and analyze to derive the condition data based on which condition data the faulty component's knot is determined to be not at a. After all the condition data referred by the machine when making the diagnosis, the machine can output all the condition data or only output the condition data referred by the machine when the symptom is not in A for the reference of the maintenance engineer. The maintenance engineer can perform self-detection and self-analysis according to the condition data output by the machine to determine whether the misjudge is the machine or the machine, and if the misjudge is the machine, the maintenance engineer can know the diagnosis bug of the machine based on the condition data output by the machine and make corresponding adjustment on the diagnosis program of the machine so as to refine the diagnosis precision of the machine. Alternatively, the service engineer may inform the machine of the condition data that the service engineer does not recognize or determines to be incorrect, and input the correct diagnosis result and condition data into the machine, so that the machine automatically learns and updates the diagnosis program to improve the diagnosis accuracy.

Based on the disclosure of the above embodiments, it can be known that the embodiments of the present application have the beneficial effects that when the diagnostic data of the user and the machine for the target syndrome are inconsistent, the second diagnostic data of the machine is determined based on the first diagnostic data of the user and the second diagnostic data of the machine, instead of the condition data referred by the first diagnostic data, and the condition data is output for the user to refer to, so that the user knows the reason why the machine makes the second diagnosis, and based on the reason or self-detection, or the vulnerability of the diagnostic program of the machine is determined, and for the vulnerability, the correct diagnostic data and condition data of the machine are actively modified or notified to be automatically repaired by the machine, and the diagnostic program is updated, so as to improve the future diagnostic accuracy of the machine.

Further, as shown in fig. 2, the method in this embodiment further includes:

actual diagnostic data for the target event is determined based on the decision data, and the actual diagnostic data may be at least the first diagnostic data or the second diagnostic data.

That is, after the machine outputs at least the condition data of which the diagnosis result B is obtained instead of a, the service engineer may determine whether the reference of each condition data is wrong, that is, whether the reference of the condition data should be judged as wrong, determine whether the behavior of referring to the condition data is approved, and input the behavior into the machine, for example, the machine may list each condition data and set a check box in front, and the service engineer may check the condition data which is not approved or approved, so as to input the judgment data about the approval degree into the machine in this way. Alternatively, the machine may also set a plurality of checkboxes representing different acceptance percentages before each condition data after listing each condition data, and the service engineer may determine the acceptance, etc. based on the checkboxes. After the machine obtains the identification determination data of each condition data by the maintenance engineer, the actual diagnosis data of the target component may be determined based on the identification determination data, for example, the machine may analyze the identification determination data of the identification of the maintenance engineer, for example, if the data based on the historical similar events is analyzed and compared with the identification determination data of each condition data, to determine whether the identification data of the maintenance engineer is correct, if the machine is analyzed and the identification of each condition data by the maintenance engineer is determined, it is determined that the first diagnosis data made by the maintenance engineer is correct, and the first diagnosis data is defined as the actual diagnosis data, that is, the first diagnosis data is used as the final diagnosis result. Conversely, if the service engineer approves all or most of the condition data referenced by the machine, the machine may determine that the second diagnostic data is correct and define the second diagnostic data as the final actual diagnostic data. Alternatively, the service engineer may specify third diagnostic data different from the first diagnostic data and the second diagnostic data based on the condition data, or the machine may specify the third diagnostic data based on the determination data and output the same, and if the service engineer and the machine mutually approve the third diagnostic data, the third diagnostic data may be specified as actual diagnostic data. During the determination of the third diagnostic data, the machine may also output condition data, which is referred to when its diagnosis is the third diagnosis, for the service engineer to decide, and determine whether the service engineer approves the diagnosis based on the result of the decision.

Further, the determining, according to the first diagnostic data and the second diagnostic data, the condition data to which the machine refers to the second diagnostic data instead of the first diagnostic data includes:

second condition data related to the second diagnostic data information in the analysis result of the machine is determined based on the second diagnostic data.

For example, the machine may first determine the meaning of the first diagnostic data and the second diagnostic data based on the obtained first diagnostic data, such as the first diagnostic data indicating that the faulty component has a fault a and the second diagnostic data indicating that the faulty component has a fault B. The machine may then analyze the condition data that it would reference when it determines that the component has a fault a, the condition data that it would reference when it determines that the component has a fault B, and define the condition data that it references when it determines that fault a is the first condition data and the condition data that it references when it determines that fault B is the second condition data. The first condition data and the second condition data may be indicative of condition data to which the machine is referenced when the diagnostic component has a fault B rather than a fault a, and the machine may directly output the first condition data and the second condition data. Or the machine can compare the condition data of the fault A and the fault B to determine different condition data, then only output different condition data, and simultaneously output fault data which can be matched with the first condition data or the second condition data or the different condition data and is possessed by the fault component for the reference of a maintenance engineer.

Further, in order to improve the accuracy of fault diagnosis, a plurality of diagnostic models for analyzing the target event to obtain different diagnostic data are established in the machine in the embodiment, and the training data of each diagnostic model is at least partially different. That is, each diagnostic model is used to diagnose different fault problems of the target component, so each diagnostic model is targeted during training, for example, the first diagnostic model is used to diagnose whether the target component has a fault a, i.e., a problem a, the second diagnostic model is used to diagnose whether the target component has a fault B, i.e., a problem B, and so on, so when each diagnostic model is trained, the data that causes the target component to have the problem a is trained on the first diagnostic model, and the data that causes the target component to have the problem B is trained on the second diagnostic model, so the diagnostic accuracy of the trained diagnostic model for each fault problem is greatly improved, compared with a model trained based on all the diagnostic data of the problem, the diagnostic result of each diagnostic model in this embodiment is in accordance with the actual situation, i.e. higher accuracy.

Further, as shown in fig. 3, the determining of the first condition data and the second condition data in the present embodiment includes:

determining a second diagnosis model used for calculating that the target event has second diagnosis data in the machine based on the second diagnosis data;

determining a first significant factor based on the first diagnostic model, and defining the first significant factor as first condition data;

a second significance factor is determined based on the second diagnostic model and defined as second condition data.

Specifically, in the present embodiment, when determining the first condition data and the second condition data, first, a matching diagnostic model is determined according to the first diagnostic data and the second diagnostic data, for example, if it is known from the first diagnostic data that the maintenance engineer considers that the target component has the problem a, the second diagnostic data determines that the target component has the problem B, the diagnostic model for diagnosing whether the target component has the problem a is the first diagnostic model, and the diagnostic model for diagnosing whether the target component has the problem B is the second diagnostic model, so the diagnostic models respectively matching the two diagnostic data are determined as the first diagnostic model and the second diagnostic model through comparison. Next, the system determines a first significant factor of the first diagnostic model and a second significant factor of the second diagnostic model, the two significant factors being respectively a factor characterizing the first diagnostic model as determining that the target component has a problem a and a factor characterizing the significant impact on the target component as having a problem a, and a factor characterizing the second diagnostic model as determining that the target component has a problem B, the two factors being available during training of the respective models. The number of the significant factors is not unique, and may be one or more, and is not particularly limited. The machine, upon determining the first and second significant factors, defines the first and second significant factors as first and second condition data, respectively.

Further, when obtaining the determination data representing the user's acceptance of the condition data, the method includes:

obtaining first judgment data, wherein the first judgment data represents the proportion of the subdata identified by the user in the first condition data to all subdata;

For example, in combination with the above embodiment, the first condition data is the first significant factor, and if there are k first significant factors, the machine lists and outputs the k first significant factors, please the service engineer judge the k first significant factors, determine whether the k first significant factors cause the target component to have the problem a, and also determine whether the target component has the fault data matching with the first significant factor, when the k first significant factors are specifically output, the first significant factors and the fault data of the target component may be fused, so that when the k first significant factors are output, the specific content output represents the k significant factors through the fault data of the target component, or the matched fault data and the first significant factors are output at the same time, for example, output in a table comparison manner, so as to facilitate the reference of the service engineer, and (6) judging. When the maintenance engineer judges, the machine may set a check box before each significant factor for the maintenance engineer to check the significant factor approved by the maintenance engineer, then the machine determines the proportion of the checked significant factor based on the check result, and determines the first determination data based on the proportion data. Similarly, the second determination data is obtained by the same method as described above.

Further, determining actual diagnostic data for the target event based on the decision data after the machine obtains the decision data includes:

determining that the first decision data is less than a first threshold;

determining that the second determination data is less than a second threshold;

determining the first diagnostic data as actual diagnostic data; or

Determining that the first determination data is not less than a first threshold;

the second diagnostic data is determined to be actual diagnostic data.

Specifically, it is assumed that the first threshold and the second threshold are respectively a fixed limit, such as 50%, although the two limits may not be consistent, and are not specific. In this embodiment, two thresholds that are the same, and both of them are 50%, are taken as an example for explanation. When the machine judges that the first judgment data and the second judgment data are both less than 50%, the machine indicates that the maintenance engineer does not approve the first significant factor and the second significant factor determined by the first diagnosis model and the second diagnosis model in the machine, or approve the corresponding relationship between the fault data in the fault part and each significant factor by the machine, and at the moment, the machine automatically determines that the first diagnosis data determined by the maintenance engineer is the final actual diagnosis data. And when the machine determines that the first judgment data and the second judgment data are not less than or equal to 50%, if so, the service engineer recognizes the significant factors of each diagnosis model in the machine and also recognizes the corresponding relationship between the significant factors and the fault data of the fault component, and at the moment, the machine automatically determines that the second diagnosis data made by the machine is the actual diagnosis data.

Further, as shown in fig. 4, when the machine determines that the first determination data is smaller than the first threshold value and the second determination data is not smaller than the second threshold value based on the first determination data and the second determination data, then:

If the machine determines that the first judgment data is smaller than the first threshold value and the second judgment data is not smaller than the second threshold value, the machine can know that the maintenance engineer does not approve the machine to judge that the fault part has the significant factor of the problem A and approves the machine to judge that the fault part has the significant factor of the problem B, at the moment, the machine can compare the two judgment data to determine the size relationship of the two judgment data, and the diagnosis data corresponding to the judgment data of the part with a larger value is taken as final actual diagnosis data.

Further, as shown in fig. 5, if the machine determines that the first determination data is smaller than the first threshold or the second determination data is smaller than the second threshold, in order to prevent the service engineer from making a false determination, the machine may:

processing the first condition data or the second condition data, and performing clustering calculation on the processing result;

determining a reference event similar to the target event in the historical data and diagnosis data thereof based on the calculation result;

and outputting the reference event and the diagnosis data thereof so that the user judges the first condition data or the second condition data again based on the reference event and the diagnosis data thereof, and determines the actual diagnosis data of the target event based on the new judgment data.

For example, when the machine determines that the service engineer only approves the first significant factor or the second significant factor, in order to avoid erroneous judgment by the service engineer, the machine may process the first condition data or the second condition data which is not recognized, if the first condition data or the second condition data is analyzed and processed by PCA analysis method, the vector X is obtained, then carrying out cluster calculation on the processing result X, if the processing result X is input into a kmeans model for cluster calculation, determining a vector point with the shortest distance to the vector X in a spatial coordinate system of the historical data based on the calculation result, determining a historical case through the vector point, the historical case is a reference event with the highest similarity to the target event, and after the reference event is determined, diagnostic data about the reference event, including fault data, reference condition data and final actual diagnostic data, can be obtained based on a historical database and the like. Then, the machine outputs the diagnostic data of each reference event for the user to refer to, so as to re-determine the recognition degree of the denied first condition data or second condition data, and finally, the machine determines the final actual diagnostic data based on the relationship between the re-determined determination data and the corresponding threshold.

Further, as shown in fig. 6, in the present embodiment, the determining the actual diagnostic data of the target event based on the determination data further includes:

determining that the second determination data is less than a second threshold;

For example, when the machine determines that the first decision data for diagnosing that the target component has the first significant factor of the first diagnostic model of problem a is not less than the first threshold, it indicates that the service engineer approves the ability of the first diagnostic model to diagnose problem a and also approves it as deeming that the target component does not have a diagnosis of problem a. The machine then determines that second decision data for diagnosing that the target component has a second significant factor of the second diagnostic model of problem B is less than a second threshold, indicating that the service engineer does not recognize the ability of the second diagnostic model to diagnose problem B, does not recognize its marked significant factor, or does not recognize that the target component is considered to have a diagnosis of problem B. At this time, the machine re-determines the significant factor and the weight of the second diagnostic model by using an ahp (analytic Hierarchy process) analytic Hierarchy process based on the second condition data and the second determination data.

Specifically, as shown in fig. 7, in a specific application, the method may specifically include:

for the target component to diagnose, the engineer diagnosis is considered as problem a, and the machine diagnosis is considered as B problem rather than a problem, which is inconsistent.

Step 1, prompting an engineer of a machine diagnosis result, and asking the engineer to quickly check the problem A and the problem B, if the engineer judges that the problem A and the problem B are consistent with the machine learning result, ending the event, and if the engineer still insists on that the problem A and the problem B are inconsistent with the machine diagnosis result, entering step 2, and checking the machine diagnosis factors by the engineer.

And 2-1, the machine learns a diagnosis model for judging the problem A in advance according to historical data, lists the significant factors of the model for learning the problem A by the machine, and asks an engineer to check and confirm the k significant factors one by one to determine whether the event is approved or not based on the values of the k significant factors, so that the machine can judge that the event does not belong to the problem A. The number of significant factors approved by statistical engineers is m.

And detecting whether m is larger than k/2, and if so, regarding that the engineer approves the machine learning to consider that the target component has no A problem.

Step 2-2 lists the significant factors of the diagnosis model of the machine learning B problem, and if the number of the significant factors is L, an engineer is asked to check and confirm one by one, whether the event is approved to be based on the numerical values of the L significant factors can represent that the machine judges that the part has the B problem. The number of significant factors recognized by the statistical engineer is n.

And detecting whether n is greater than L/2, and if so, regarding that the engineer approves the machine diagnosis to consider that the target component has a B problem.

When the calculated value m is equal to k/2 in the step 2-1, the characterization engineer does not approve the machine learning to consider that the target component has no A problem, or the step 2-2, the calculated value n is equal to L/2, the engineer does not approve the machine learning to consider that the target component has B problem, and then the step 3-1 or 3-2 is carried out, and other historical records close to the values of all factors of the event are provided for reference. The engineer is then asked to repeat the evaluation of step 2-1 or step 2-2.

The method comprises the following steps: firstly, PCA is used for extracting a vector, then the vector is put into a kmeans model, and another historical event with the shortest distance to the event is measured and is used as a reference for an engineer.

When 5 engineers evaluate, compare m/k and n/L

(1) If the machine diagnosis is agreed to have no problem A (the original engineer judges the problem) and the machine diagnosis is agreed to have a problem B (the original machine judges the problem), the result is correct machine learning: there is a problem B (original machine determination problem).

(2) If the machine diagnosis is agreed to have no problem A (the original engineer judges the problem) and the machine diagnosis is not agreed to have a problem B (the original machine judges the problem), the result is analyzed by an AHP (analytic Hierarchy process) analytic Hierarchy process to determine the significant factors and weights of the diagnosis model for diagnosing the problem B again.

(3) If the machine diagnosis is not agreed to have the A problem (the original engineer judges the problem) and the machine diagnosis is agreed to have the B problem (the original machine judges the problem), the result takes the problem type with high agreement percentage as the diagnosis problem, for example, the A problem does not exist, the acceptance is m/k%, which indicates the existence of the A problem, the acceptance is 1-m/k, the acceptance of the B problem agrees to the machine diagnosis is n/L%, for example, 1-m/k > n/L, the diagnosis result is the A problem, for example, 1-m/k < n/L, the diagnosis result is the B problem.

(4) If the diagnosis of the unapproved machine is not the problem A (the original engineer judges the problem) and the unapproved machine is diagnosed with the problem B (the original machine judges the problem), the diagnosis result is that the engineer is correct and the problem A exists.

By adopting the data processing method of the embodiment, machine learning can be combined with the intelligence of a maintenance engineer, so that the machine learning is more flexible, the learning process can be considered to be participated, and the efficiency and the capability of the machine learning are effectively improved.

As shown in fig. 8, an embodiment of the present application also provides an electronic device, which includes:

an obtaining module, configured to obtain first diagnostic data representing an analysis result of a user on a target event, and second diagnostic data representing an analysis result of an electronic device on the target event, where the second diagnostic data is different from the first diagnostic data;

and the processing module is used for determining the condition data referred by the electronic equipment to obtain the second diagnosis data instead of the first diagnosis data according to the first diagnosis data and the second diagnosis data.

For example, when a component of an automobile (corresponding to a target event) is faulty or a functional component in a computer is faulty and a maintenance engineer (corresponding to a user) intends to perform fault diagnosis on the faulty component, status data of the faulty component, data on a problem phenomenon, and the like may be input into the electronic device, the electronic device performs diagnosis and the obtaining module obtains second diagnosis data, and the maintenance engineer may also perform diagnosis on the faulty component and obtain first diagnosis data. When the maintenance engineer finds that the first diagnostic data and the second diagnostic data are different, the first diagnostic data can be input into the electronic device, or the maintenance engineer directly inputs the first diagnostic data into the electronic device after making the diagnostic data, and the electronic device makes a judgment that the first diagnostic data is different from the second diagnostic data. When the electronic device obtains two diagnostic data and knows that the two diagnostic data are different, the processing module of the electronic device may analyze the two diagnostic data, for example, the processing module may analyze the two diagnostic data to determine a fault knot, such as a first diagnostic data indicating that the knot of the faulty component is a knot a and a second diagnostic data indicating that the knot of the faulty component is a knot B, and then the electronic device may determine condition data that it refers to when obtaining the second diagnostic data, i.e., determine the condition data that the knot is B and analyze to derive condition data that determines that the knot of the faulty component is not a based on which condition data. After all the condition data referred by the electronic equipment when making diagnosis, the electronic equipment can output all the condition data or only output the condition data referred by the electronic equipment when the symptom knot is not in A for reference of a maintenance engineer. The maintenance engineer can perform self-detection and self-analysis according to the condition data output by the electronic equipment to determine whether the misjudge party is the electronic equipment or not, and if the misjudge party is the electronic equipment, the maintenance engineer can know the diagnosis vulnerability of the electronic equipment based on the condition data output by the electronic equipment and correspondingly adjust the diagnosis program of the electronic equipment so as to refine the diagnosis precision of the electronic equipment. Alternatively, the service engineer may inform the electronic device of the condition data that the electronic device does not recognize or determines the condition data to be incorrect, and input the correct diagnosis result and condition data into the electronic device, so that the electronic device automatically learns and updates the diagnosis program to improve the diagnosis accuracy.

Based on the disclosure of the above embodiment, it can be known that the electronic device according to the embodiment of the present application has the beneficial effects that when the diagnostic data of the user and the electronic device for the target syndrome are inconsistent, the electronic device can determine, based on the first diagnostic data of the user and the second diagnostic data of the electronic device, that the electronic device obtains the second diagnostic data, instead of each condition data referred by the first diagnostic data, and output the condition data for the user to refer to, so that the user knows the reason why the electronic device makes the second diagnosis, and based on the reason or self-detection, or determines the vulnerability of the diagnostic program of the electronic device, and actively modifies or informs the electronic device that the correct diagnostic data and condition data are automatically repaired by the electronic device, and updates the diagnostic program to improve the future diagnostic accuracy of the electronic device.

Further, the obtaining module in this embodiment is further configured to:

the processing module is further configured to:

That is, when the obtaining module outputs at least the condition data of which the diagnosis result B is obtained instead of a, the service engineer may judge whether the reference of each condition data is wrong, that is, whether the reference of the condition data should be wrong, determine whether the behavior of referring to the condition data is approved, and input the behavior into the electronic device, for example, the electronic device may list each condition data and set a check box in front of it, and the service engineer may check the condition data which is not approved or approved, so as to input the determination data about the degree of approval into the obtaining module in this way. Alternatively, after listing the condition data, the electronic device may set a plurality of checkboxes representing different acceptance percentages before the condition data, and the service engineer may determine the acceptance degree based on the checkboxes. After the obtaining module obtains the identification determination data of each condition data by the maintenance engineer, the actual diagnosis data of the target component may be determined based on the identification determination data, for example, the processing module of the electronic device may analyze the identification determination data of the identification of the maintenance engineer, for example, if the data based on the historical similar events is analyzed and compared with the identification determination data of each condition data, to determine whether the identification data of the maintenance engineer is correct, and if the processing module identifies the identification of each condition data by the maintenance engineer after analysis, it is determined that the first diagnosis data made by the maintenance engineer is correct, and the first diagnosis data is defined as the actual diagnosis data, that is, the first diagnosis data is taken as the final diagnosis result. On the contrary, if the service engineer approves all or most of the condition data referred to by the electronic device, the electronic device may determine that the second diagnostic data is correct, and define the second diagnostic data as the final actual diagnostic data. Alternatively, the service engineer may determine third diagnostic data different from the first diagnostic data and the second diagnostic data based on the condition data, or the processing module may determine the third diagnostic data based on the determination data and output the third diagnostic data, and if the service engineer and the electronic device mutually approve the third diagnostic data, the third diagnostic data may be determined as actual diagnostic data. During the determination of the third diagnostic data, the electronic device may also output condition data, which is referred to when its diagnosis is the third diagnosis, for the service engineer to decide, and determine whether the service engineer approves the diagnosis based on the result of the decision.

Further, the determining, according to the first diagnostic data and the second diagnostic data, that the electronic device derives the second diagnostic data instead of the condition data referred to by the first diagnostic data in the present embodiment includes:

determining first condition data related to the first diagnosis data information in the analysis result of the electronic equipment based on the first diagnosis data;

second condition data related to the second diagnosis data information in the analysis result of the electronic equipment is determined based on the second diagnosis data.

For example, a processing module of the electronic device may first determine, based on the obtained first diagnostic data, the meaning of the first diagnostic data and the second diagnostic data characterizing the faulty component, such as the first diagnostic data characterizing the faulty component as having fault a, and the second diagnostic data characterizing the faulty component as having fault B. The electronic device may then analyze the condition data that it would refer to when it determines that the component has the failure a, and the condition data that it would refer to when it determines that the component has the failure B, and define the condition data that is referred to when it determines the failure a as the first condition data and the condition data that is referred to when it determines the failure B as the second condition data. The first condition data and the second condition data can represent condition data referred by a processing module of the electronic equipment when the diagnosis component has the fault B instead of the fault A, and the electronic equipment can directly output the first condition data and the second condition data. Or, the processing module of the electronic device may compare the condition data of the fault a and the fault B to determine different condition data, and then output only the different condition data, and the electronic device may also output fault data that the faulty component has and can be matched with the first condition data or the second condition data or the different condition data, for reference by a maintenance engineer.

Further, in order to improve the accuracy of fault diagnosis, a plurality of diagnostic models respectively used for analyzing the target event to obtain different diagnostic data are built in the processing module in the embodiment, and the training data of each diagnostic model is at least partially different. That is, each diagnostic model is used to diagnose different fault problems of the target component, so each diagnostic model is targeted during training, for example, the first diagnostic model is used to diagnose whether the target component has a fault a, i.e., a problem a, the second diagnostic model is used to diagnose whether the target component has a fault B, i.e., a problem B, and so on, so when each diagnostic model is trained, the data that causes the target component to have the problem a is trained on the first diagnostic model, and the data that causes the target component to have the problem B is trained on the second diagnostic model, so the diagnostic accuracy of the trained diagnostic model for each fault problem is greatly improved, compared with a model trained based on all the diagnostic data of the problem, the diagnostic result of each diagnostic model in this embodiment is in accordance with the actual situation, i.e. higher accuracy.

Further, the determining of the first condition data and the second condition data in this embodiment includes:

determining a first diagnosis model used for calculating that the target event has first diagnosis data in the electronic equipment based on the first diagnosis data;

determining a second diagnosis model used for calculating that the target event has second diagnosis data in the electronic equipment based on the second diagnosis data;

Specifically, in the present embodiment, when determining the first condition data and the second condition data, first, a matching diagnostic model is determined according to the first diagnostic data and the second diagnostic data, for example, if it is known from the first diagnostic data that the maintenance engineer considers that the target component has the problem a, the second diagnostic data determines that the target component has the problem B, the diagnostic model for diagnosing whether the target component has the problem a is the first diagnostic model, and the diagnostic model for diagnosing whether the target component has the problem B is the second diagnostic model, so the diagnostic models respectively matching the two diagnostic data are determined as the first diagnostic model and the second diagnostic model through comparison. Next, the system determines a first significant factor of the first diagnostic model and a second significant factor of the second diagnostic model, the two significant factors being respectively a factor characterizing the first diagnostic model as determining that the target component has a problem a and a factor characterizing the significant impact on the target component as having a problem a, and a factor characterizing the second diagnostic model as determining that the target component has a problem B, the two factors being available during training of the respective models. The number of the significant factors is not unique, and may be one or more, and is not particularly limited. The processing module defines the first and second significant factors as first and second condition data, respectively, upon determining the first and second significant factors.

Further, the obtaining module, when obtaining the decision data representing the user's acceptance of the condition data, includes:

For example, in combination with the above embodiment, the first condition data is the first significant factor, assuming that there are k first significant factors, the electronic device outputs the k first significant factors in a list, please the service engineer to judge the k first significant factors, determine whether the k first significant factors cause the target component to have the problem a, and also determine whether the target component has fault data matching with the first significant factor, when the k first significant factors are specifically output, the first significant factors and the fault data of the target component may be fused, so that when the k first significant factors are output, the specific content output represents the k significant factors through the fault data of the target component, or the matched fault data and the first significant factors are output at the same time, for example, output in a table comparison manner, so as to facilitate the reference of the service engineer, and (6) judging. When the maintenance engineer judges, the electronic device may set a check box before each significant factor for the maintenance engineer to check the significant factor approved by the maintenance engineer, and then the electronic device determines the proportion of the checked significant factor based on the check result, and determines the first determination data based on the proportion data. Similarly, the second determination data is obtained by the same method as described above.

Further, determining actual diagnostic data for the target event based on the decision data after the processing module obtains the decision data includes:

determining that the first decision data is less than a first threshold;

determining that the second determination data is less than a second threshold;

determining the first diagnostic data as actual diagnostic data; or

the second diagnostic data is determined to be actual diagnostic data.

Specifically, it is assumed that the first threshold and the second threshold are respectively a fixed limit, such as 50%, although the two limits may not be consistent, and are not specific. In this embodiment, two thresholds that are the same, and both of them are 50%, are taken as an example for explanation. When the processing module judges that the first judgment data and the second judgment data are both less than 50%, the processing module indicates that a maintenance engineer does not approve the first significant factor and the second significant factor determined by the first diagnosis model and the second diagnosis model in the processing module, or approves the corresponding relationship between the fault data in the fault component and each significant factor made by the processing module, and at this moment, the processing module automatically determines that the first diagnosis data determined by the maintenance engineer is the final actual diagnosis data. And when the processing module determines that the first judgment data and the second judgment data are not less than or equal to 50%, if the first judgment data and the second judgment data are not less than or equal to 50%, the service engineer approves the significant factors of each diagnosis model in the processing module, and also approves the corresponding relationship between each significant factor and the fault data of the fault component by the processing module, and at the moment, the processing module automatically determines that the second diagnosis data made by the processing module by the electronic equipment is actual diagnosis data.

Further, if the processing module determines that the first determination data is smaller than the first threshold and the second determination data is not smaller than the second threshold based on the first determination data and the second determination data, then:

If the processing module determines that the first determination data is smaller than the first threshold value and the second determination data is not smaller than the second threshold value, the processing module can know that the maintenance engineer does not approve the significant factor of the electronic equipment that the failure component has the problem A, and the approval processing module determines that the failure component has the significant factor of the problem B, at this moment, the processing module can compare the two determination data to determine the size relationship of the two determination data, and the diagnosis data corresponding to the determination data of the party with larger value is taken as the final actual diagnosis data.

Further, if the processing module determines that the first determination data is smaller than the first threshold or the second determination data is smaller than the second threshold, in order to prevent the service engineer from making a false determination, the processing module may:

For example, when the processing module determines that the service engineer only approves the first significant factor or the second significant factor, in order to avoid the misjudgment phenomenon of the maintenance engineer, the processing module can process the first condition data or the second condition data which is not approved, if the first condition data or the second condition data is analyzed and processed by PCA analysis method, the vector X is obtained, then carrying out cluster calculation on the processing result X, if the processing result X is input into a kmeans model for cluster calculation, determining a vector point with the shortest distance to the vector X in a spatial coordinate system of the historical data based on the calculation result, determining a historical case through the vector point, the historical case is a reference event with the highest similarity to the target event, and after the reference event is determined, diagnostic data about the reference event, including fault data, reference condition data and final actual diagnostic data, can be obtained based on a historical database and the like. Then, the processing module controls the electronic device to output the diagnostic data of each reference event for the user to refer, so as to judge the recognition degree of the denied first condition data or the denied second condition data again, and finally, the processing module determines the final actual diagnostic data based on the relation between the judged data and the corresponding threshold value.

Further, in this embodiment, the determining, at the processing module, the actual diagnostic data of the target event based on the determination data further includes:

determining that the second determination data is less than a second threshold;

For example, when the processing module determines that the first decision data for diagnosing that the target component has a first significant factor of the first diagnostic model of problem A is not less than the first threshold, it indicates that the service engineer approves the ability of the first diagnostic model to diagnose problem A and also approves that the target component does not have a diagnosis of problem A. The processing module then determines that second decision data for diagnosing a second significant factor of a second diagnostic model of the target component having problem B is less than a second threshold, indicating that the service engineer does not recognize the ability of the second diagnostic model to diagnose problem B, does not recognize its marked significant factor, or does not recognize that the target component is considered to have a diagnosis of problem B. At this time, the processing module re-determines the significant factor and the weight of the second diagnosis model by using an ahp (analytic Hierarchy process) analytic Hierarchy process based on the second condition data and the second determination data.

Specifically, as shown in fig. 7, in a specific application, the method executed by the electronic device may specifically include:

When 5 engineers evaluate, compare m/k and n/L

After the electronic equipment adopts the data processing method of the embodiment, the learning of the electronic equipment can be combined with the intelligence of a maintenance engineer, so that the learning of the electronic equipment is more flexible, the learning process can be considered to participate, and the learning efficiency and capability of the electronic equipment are effectively improved.

The above embodiments are only exemplary embodiments of the present application, and are not intended to limit the present application, and the protection scope of the present application is defined by the claims. Various modifications and equivalents may be made by those skilled in the art within the spirit and scope of the present application and such modifications and equivalents should also be considered to be within the scope of the present application.

Claims

1. A data processing method, comprising:

determining from the first and second diagnostic data at least condition data to which the machine refers to derive the second diagnostic data but not the first diagnostic data;

outputting the condition data;

wherein determining from the first and second diagnostic data at least condition data to which the machine derived the second diagnostic data but not the first diagnostic data refers comprises:

2. The method of claim 1, further comprising:

determining actual diagnostic data for the target event based on the decision data, the actual diagnostic data being at least the first diagnostic data or the second diagnostic data.

3. The method of claim 2, wherein a plurality of diagnostic models are built into the machine for analyzing the target events to derive different diagnostic data, respectively, the training data of each diagnostic model being at least partially different;

the determining the first and second condition data comprises:

4. A method according to claim 3, wherein said obtaining decision data characterizing user acceptance of said condition data comprises:

5. The method of claim 4, wherein said determining actual diagnostic data for the target event based on the decision data comprises:

determining that the first decision data is less than a first threshold;

determining that the second determination data is less than a second threshold;

determining the first diagnostic data to be actual diagnostic data; or

Determining that the first decision data is not less than a first threshold;

the second diagnostic data is determined to be actual diagnostic data.

6. The method of claim 4, wherein,

determining that the first decision data is less than a first threshold;

7. The method of claim 4, wherein said determining actual diagnostic data for the target event based on the decision data comprises:

8. The method of claim 4, wherein said determining actual diagnostic data for the target event based on the decision data comprises:

determining that the first decision data is not less than a first threshold;

determining that the second determination data is less than a second threshold;

9. An electronic device, comprising:

a processing module for determining, from the first and second diagnostic data, respective condition data to which the machine refers the second diagnostic data instead of the first diagnostic data, comprising: