CN113705896A - Target equipment determination method and device and electronic equipment - Google Patents
Target equipment determination method and device and electronic equipment Download PDFInfo
- Publication number
- CN113705896A CN113705896A CN202111004041.4A CN202111004041A CN113705896A CN 113705896 A CN113705896 A CN 113705896A CN 202111004041 A CN202111004041 A CN 202111004041A CN 113705896 A CN113705896 A CN 113705896A
- Authority
- CN
- China
- Prior art keywords
- data
- maintenance
- value
- trigger threshold
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07F—COIN-FREED OR LIKE APPARATUS
- G07F19/00—Complete banking systems; Coded card-freed arrangements adapted for dispensing or receiving monies or the like and posting such transactions to existing accounts, e.g. automatic teller machines
- G07F19/20—Automatic teller machines [ATMs]
Abstract
The specification discloses a method and a device for determining target equipment and electronic equipment, wherein the method comprises the following steps: acquiring characteristic data of a plurality of devices; according to the characteristic data of the plurality of devices, obtaining the prediction probability of the failure of each device in a preset time period from the characteristic data acquisition time; establishing a total cost target function with a maintenance trigger threshold as a variable according to the prediction probability of the fault of each device in the plurality of devices, the maintenance cost under the condition of the fault of each device and the maintenance cost of each device; solving the total cost objective function, and taking the maintenance trigger threshold value when the value of the total cost objective function is minimum as a target trigger threshold value; and taking the equipment with the failure prediction probability larger than or equal to the target trigger threshold value as the target equipment needing maintenance. The scheme can minimize the total maintenance cost of all equipment, reduce the failure rate of the equipment and the time which cannot be used when the equipment fails, and reduce the maintenance workload.
Description
Technical Field
The present application relates to the field of artificial intelligence technologies, and in particular, to a method and an apparatus for determining a target device, and an electronic device.
Background
Maintenance of many devices is currently performed regularly, for example, four times a year for permanent maintenance. The total cost of strengthening the maintenance cost after the equipment fails is higher.
When equipment breaks down and needs to be maintained outside fixed maintenance time, a maintenance unit is often required to be contacted, maintenance personnel are sent by the maintenance unit, and the maintenance personnel arrange maintenance time according to the self workload condition and the distance. Therefore, it often takes a long time from the generation of a fault to the troubleshooting, especially in an enterprise where the fault repair requires progressive feedback. For example, the regular maintenance and temporary fault maintenance mechanisms of the ATM of the bank are all in a step-by-step feedback mode, and a network node, a higher network node and a maintenance provider need to participate together. When the ATM fails, the fault is reported to a maintenance provider step by step through the network points, and the maintenance provider sends maintenance personnel to maintain the fault, so that the fault machine cannot be used for a long time.
Under the background of intense competition and high customer requirements among different enterprises and in the same enterprise, the availability of equipment is high. Current equipment repair and maintenance mechanisms have been unable to meet this requirement.
Disclosure of Invention
The embodiment of the application aims to provide a method and a device for determining target equipment and electronic equipment, so as to solve the problems that the existing equipment maintenance and repair mechanism is high in maintenance cost and cannot be used due to long time of fault equipment.
To solve the above technical problem, a first aspect of the present specification provides a method for determining a target device, including: acquiring characteristic data of a plurality of devices; according to the characteristic data of the plurality of devices, obtaining the prediction probability of the failure of each device in a preset time period from the characteristic data acquisition time; establishing a total cost target function with a maintenance trigger threshold as a variable according to the prediction probability of the fault of each device in the plurality of devices, the maintenance cost under the condition of the fault of each device and the maintenance cost of each device; solving the total cost objective function, and taking the maintenance trigger threshold value when the value of the total cost objective function is minimum as a target trigger threshold value; and taking the equipment with the failure prediction probability larger than or equal to the target trigger threshold value as the target equipment needing maintenance.
In some embodiments, the total cost objective function with the maintenance trigger threshold as a variable includes one of:
where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykThe maintenance cost is the maintenance cost under the condition that the equipment fails;
where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykThe maintenance cost under the condition that equipment fails is obtained, and n is the maximum failure frequency corresponding to the non-zero prediction probability;
wherein, 0 < pk< 1 and pkλ is more than or equal to λ, and λ is a maintenance trigger threshold; alpha is a predetermined coefficient, K is the number of devices with a predicted probability greater than or equal to a maintenance trigger threshold, paykThe maintenance cost is the maintenance cost under the condition that the equipment fails;
wherein, 0 < pk< 1 and pkλ is more than or equal to λ, and λ is a maintenance trigger threshold; alpha is a predetermined coefficient, K is the number of devices with a predicted probability greater than or equal to a maintenance trigger threshold, paykAnd n is the maximum failure frequency corresponding to the non-zero prediction probability, which is the maintenance cost under the condition that the equipment fails.
In some embodiments, solving the total cost objective function, and taking the maintenance trigger threshold when the value of the total cost objective function is minimum as the target trigger threshold includes: taking 0 to 1 as a target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the target interval, substituting the numerical value into the total cost target function, and obtaining a total cost target function value; selecting a maintenance trigger threshold corresponding to the minimum total cost target function value as a reference value; the following steps are executed in a loop until the determined current reference value is equal to the current reference value determined in the previous loop: taking an interval with a preset length taking a current reference value as a center as a current target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the current target interval, substituting the numerical value into the total cost target function, solving a total cost target function value, and selecting a maintenance trigger threshold value corresponding to the minimum total cost target function value as a current reference value; and taking the determined current reference value as the target trigger threshold.
In some embodiments, obtaining the predicted probability of failure of each device within a predetermined time period from the time of collecting the characteristic data according to the characteristic data of the plurality of devices includes: inputting the characteristic data of the plurality of devices into a pre-established prediction model; and taking the output of the prediction model as the prediction probability of the fault of each device in a preset time period from the characteristic data acquisition moment.
In some embodiments, the predictive model is trained as follows: obtaining a sample pair of a plurality of devices, wherein the sample pair comprises characteristic data and a marking value, and the marking value is used for marking whether the device has a fault within a preset time period from the characteristic data acquisition time; training the predictive model based on the sample pairs for the plurality of devices.
In some embodiments, in a case that the prediction model is a lightGBM model, selecting a data set on a leaf node to be segmented in the lightGBM model includes: calculating the sum of the mutual distances among all data in the data set on each leaf node; and taking the data set with the maximum mutual distance between the data as a data set on the leaf node to be segmented.
In some embodiments, when the prediction model is a lightGBM model, the first data set on a leaf node is segmented to obtain a second data set on the leaf node according to the following method: calculating the sum of the distances between each data in the first data set and the other data; taking the data with the largest sum of the distances between the data and the rest data as the data in the second data set; repeatedly executing the following steps until the difference value is a negative value: calculating the mean value of the distances between each data point in the first data set and the rest data in the first data set as first data, calculating the mean value of the distances between each data point in the first data set and all data points in the second data set as second data, and calculating the difference value of subtracting the second data from the first data; and dividing the data point with the maximum difference value into a second data set.
In some embodiments, before training the lightGBM model based on the sample pairs of the plurality of devices, the method further comprises: inputting the sample pairs into a lightGBM model; determining the times of splitting of the characteristic data of each type as leaf nodes; screening out a plurality of types of feature data according to the splitting times; correspondingly, when the lightGBM model is trained, the lightGBM model is trained by adopting a sample consisting of the screened multiple types of feature data and the labeled values.
A second aspect of the present specification provides an apparatus for determining a target device, including: the acquisition module acquires characteristic data of a plurality of devices; the prediction module is used for obtaining the prediction probability of the failure of each device in a preset time period from the characteristic data acquisition time according to the characteristic data of the plurality of devices; the establishing module is used for establishing a total cost objective function taking a maintenance trigger threshold value as a variable according to the prediction probability of the fault of each device in the plurality of devices, the maintenance cost under the condition of the fault of each device and the maintenance cost of each device; the solving module is used for solving the total cost objective function and taking the maintenance trigger threshold value when the value of the total cost objective function is minimum as a target trigger threshold value; and the first determining module is used for taking the equipment with the failure prediction probability larger than or equal to the target trigger threshold value as the target equipment needing to be maintained.
In some embodiments, the total cost objective function with the maintenance trigger threshold as a variable includes one of:
where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykThe maintenance cost is the maintenance cost under the condition that the equipment fails;
where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykThe maintenance cost under the condition that equipment fails is obtained, and n is the maximum failure frequency corresponding to the non-zero prediction probability;
wherein, 0 < pk< 1 and pkλ is more than or equal to λ, and λ is a maintenance trigger threshold; alpha is a predetermined coefficient, K is the number of devices with a predicted probability greater than or equal to a maintenance trigger threshold, paykThe maintenance cost is the maintenance cost under the condition that the equipment fails;
wherein, 0 < pk< 1 and pkλ is more than or equal to λ, and λ is a maintenance trigger threshold; alpha is a predetermined coefficient, K is the number of devices with a predicted probability greater than or equal to a maintenance trigger threshold, paykAnd n is the maximum failure frequency corresponding to the non-zero prediction probability, which is the maintenance cost under the condition that the equipment fails.
In some embodiments, the solving module comprises: the obtaining submodule is used for taking 0-1 as a target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the target interval, substituting the numerical value into the total cost target function, and obtaining a total cost target function value; the first determining submodule is used for selecting a maintenance triggering threshold value corresponding to the minimum total cost target function value as a reference value; the obtaining submodule and the first determining submodule are further configured to perform the following steps in a loop until the determined current reference value is equal to the current reference value determined in the previous loop: taking an interval with a preset length taking a current reference value as a center as a current target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the current target interval, substituting the numerical value into the total cost target function, solving a total cost target function value, and selecting a maintenance trigger threshold value corresponding to the minimum total cost target function value as a current reference value; and the second determination submodule is used for taking the determined current reference value as the target trigger threshold.
In some embodiments, the prediction module comprises: the input submodule is used for inputting the characteristic data of the plurality of devices into a pre-established prediction model; and the third determining submodule is used for taking the output of the prediction model as the prediction probability of the fault of each device within a preset time period from the characteristic data acquisition moment.
In some embodiments, a training module is further included for training the predictive model; the training module comprises: the acquisition submodule is used for acquiring sample pairs of a plurality of devices, the sample pairs comprise characteristic data and a marking value, and the marking value is used for marking whether the devices have faults within a preset time period from the characteristic data acquisition moment; a training sub-module to train the predictive model based on the sample pairs of the plurality of devices.
In some embodiments, in a case that the prediction model is a lightGBM model, the training module further includes a selecting sub-module, configured to select a data set on a leaf node to be segmented in the lightGBM model; the selection submodule comprises: the first calculation submodule is used for calculating the sum of the mutual distances among all data in the data set on each leaf node; and the fourth determining submodule is used for taking the data set with the maximum mutual distance between the data as the data set on the leaf node to be segmented.
In some embodiments, in the case that the prediction model is a lightGBM model, the training module further includes a partitioning sub-module, configured to partition the first data set on the leaf node to obtain a second data set on the leaf node according to the following method; the partitioning submodule includes: the second calculation submodule is used for calculating the sum of the distances between each datum in the first data set and the rest of data; a fifth determining submodule, configured to use the data with the largest sum of distances to the remaining data as the data in the second data set; the second calculating submodule and the fifth determining submodule are further configured to repeatedly perform the following steps until the difference value is a negative value: calculating the mean value of the distances between each data point in the first data set and the rest data in the first data set as first data, calculating the mean value of the distances between each data point in the first data set and all data points in the second data set as second data, and calculating the difference value of subtracting the second data from the first data; and dividing the data point with the maximum difference value into a second data set.
In some embodiments, in the case that the prediction model is a lightGBM model, further comprising: an input module for inputting the sample pairs into a lightGBM model; the second determining module is used for determining the times of splitting of the characteristic data of each type as leaf nodes; the screening module is used for screening out a plurality of types of feature data according to the splitting times; correspondingly, when the lightGBM model is trained, the lightGBM model is trained by adopting a sample consisting of the screened multiple types of feature data and the labeled values.
A third aspect of the present specification provides an electronic apparatus comprising: a memory and a processor, the processor and the memory being communicatively connected to each other, the memory having stored therein computer instructions, the processor implementing the steps of the method of the first aspect or any implementation thereof by executing the computer instructions.
A fourth aspect of the present specification provides a computer storage medium storing computer program instructions which, when executed, implement the steps of the method of the first aspect or any implementation thereof.
According to the determining method and device for the target equipment and the electronic equipment, the prediction probability that each piece of equipment fails within a preset time period from the characteristic data acquisition time is predicted according to the characteristic data of a plurality of pieces of equipment, the equipment with the prediction probability being greater than or equal to the target maintenance trigger threshold value is used as the target equipment, and the target equipment obtained according to the target maintenance trigger threshold value is maintained in advance to enable the total maintenance cost of all pieces of equipment to be the lowest due to the fact that the target maintenance trigger threshold value corresponds to the situation that the total cost target function value is the smallest; the failure rate of the equipment and the time which cannot be used when the equipment fails can be reduced; during maintenance at every turn, only need maintain the target equipment can, need not to carry out inspection one by one to every equipment, reduced the work load of maintenance.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 shows a flowchart of a method for determining a target device according to an embodiment of the present description;
FIG. 2 illustrates a flowchart of a method of one embodiment of determining a target service trigger threshold;
FIG. 3 shows a flow chart of a method of deriving a prediction probability;
FIG. 4 shows a flow chart of a method of training a predictive model;
FIG. 5 illustrates a flow chart of a method of selecting a data set on a leaf node to be partitioned;
FIG. 6 illustrates a flowchart of a method of partitioning a first data set on a leaf node into a second data set on the leaf node;
FIG. 7 illustrates a flow chart of a method of screening feature data types;
FIG. 8 is a schematic block diagram of a target device determination apparatus according to an embodiment of the present disclosure;
FIG. 9 shows a functional block diagram of an electronic device according to an embodiment of the invention.
Detailed Description
In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art without any inventive work based on the embodiments in the present application shall fall within the scope of protection of the present application.
The present specification proposes a new equipment maintenance mechanism, that is, before the equipment is out of order, the equipment is maintained in advance, so as to reduce the failure rate and maintenance cost of the equipment, that is, regular maintenance is changed into time-flexible maintenance.
To this end, the present disclosure also provides a method of determining target equipment for determining which equipment needs to be serviced in advance in a future period of time. As shown in fig. 1, the method for determining the target device includes the following steps.
S110: feature data for a plurality of devices is acquired.
The devices described in the embodiments of the present specification may be self-service teller machines (e.g., ATM machines of banks), vehicles (e.g., taxis, tool cars, trains, subways, airplanes, etc.), household appliances (e.g., televisions, refrigerators, washing machines, air conditioners, etc.), electronic devices (e.g., computers, mobile phones, earphones, etc.), and may also be other devices.
In some embodiments, the "equipment" in this specification may also be each module or component contained in a large-scale equipment, and what scope the "equipment" specifically covers may vary according to actual situations.
The specific meaning of "device" can be determined according to the actual situation. Or a component or module in a device. This description is not repeated.
The critical components or modules of the various devices are different, the types of faults are different, some types of faults are predictable, and some types of faults are not predictably feasible. The determination method of the target device provided by the embodiment of the specification only focuses on the predictable failure type of the device. For example, the ATM of a bank has a line fault, a cassette fault, a receipt fault, a card reader fault, etc., and some line faults are considered to be not predictive feasibility, and the method can pay attention to the faults of the cash-out and cash-in modules.
The characteristic data of one device may include various types.
In some embodiments, the characteristic data may include length information such as the length of time the data was collected from the date the device was manufactured, the length of time the data was collected from the date the device was initially used, and the like.
In some embodiments, the characteristic data may include usage frequency information of the device. This information may be obtained by averaging, maximizing, and summing the usage frequency for each time unit for a predetermined length of time prior to the time of data acquisition. For example, the average value of the usage frequency per month in 6 months before the data acquisition time, the maximum value, and the sum of the usage frequency in 6 months.
In some embodiments, the feature data may also include brand information for the device.
In some embodiments, the frequency of use also includes environmental information of device usage. For example, the average value, the maximum value, and the like of each time unit within a predetermined time period before the data collection time, such as the average temperature, the maximum temperature, the average humidity, the maximum humidity, and the like of the environment in which the outdoor unattended sensor device is located every day within 6 months before the data collection time.
The difference between the acquisition moments of these characteristic data should be within a predetermined time difference range. For example, the difference between the acquisition times of any two feature data is within 1 hour, or within 10 minutes. Accordingly, the average value of the respective characteristic data acquisition timings or any one of the characteristic data acquisition timings may be used as the acquisition timing of the characteristic data.
S120: and obtaining the prediction probability of the fault of each device in a preset time period from the characteristic data acquisition time according to the characteristic data of the devices.
In some embodiments, the "multiple devices" may be of the same type, e.g., the devices are all bank's ATM machines.
In some embodiments, the "plurality of devices" may also be of different types, for example, the devices may include bank ATM machines and self-service payment machines, wherein the self-service payment machines may use bank cards to pay for water, electricity, gas, etc.
The "predetermined period of time" may be one week, i.e., a predicted probability of each device failing within one week from the time of feature data collection. Accordingly, the prediction method of the target device shown in fig. 1 may be performed once per week.
In some embodiments, the predicted probability obtained by the model may be a probability value indicating whether a fault has occurred.
In some embodiments, the predicted probability obtained by the model may also be two or more non-zero probability values, each of which may correspond to a number of times, representing a probability that the device failed a corresponding number of times. The number of non-zero probability values for different devices may be different. For example, device a corresponds to a non-zero probability value of: the probability of 1 fault is 0.6, and the probability of 2 faults is 0.1; the non-zero probability value corresponding to device B is: the probability of 1 failure occurrence is 0.4.
S130: and establishing a total cost target function with a maintenance trigger threshold as a variable according to the prediction probability of the fault of each device in the plurality of devices, the maintenance cost under the fault condition of each device and the maintenance cost of each device.
When the maintenance triggering threshold value, that is, the prediction probability is greater than or equal to the maintenance triggering threshold value, it indicates that the equipment corresponding to the prediction probability should be maintained.
In some embodiments, the total cost of repair in the event of a device failure may be factored into a total cost objective function. Wherein whether the equipment fails within a predetermined time period from the time of the characteristic data acquisition is determined by whether the predicted probability of the equipment is greater than or equal to a maintenance trigger threshold.
For example, a total cost objective function with a maintenance trigger threshold as a variableCan be prepared as
Where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykThe maintenance cost in case of equipment failure.
As another example, the total cost objective function with maintenance trigger threshold as a variable may be
Where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykAnd n is the maximum failure frequency corresponding to the non-zero prediction probability, which is the maintenance cost under the condition that the equipment fails.
In some embodiments, the multiple devices may be taken as a whole, and the prediction probabilities corresponding to the failed devices are summed to obtain P, and the value of P represents the overall performance of the multiple devices. Specifically, the smaller the value of P, the lower the degree of maintenance required, the better the overall performance of the multiple devices, and the lower the maintenance cost; conversely, a larger value of P indicates a higher degree of maintenance, and the overall performance of the plurality of devices is poorer, resulting in a higher maintenance cost. In this case, P, which represents performance, maintenance costs, can be converted to a cost that is factored into the total cost objective function. Wherein whether the equipment fails within a predetermined time period from the time of the characteristic data acquisition is determined by whether the predicted probability of the equipment is greater than or equal to a maintenance trigger threshold.
For example, the total cost objective function with the maintenance trigger threshold as a variable may be
As another example, the total cost objective function with maintenance trigger threshold as a variable may be
S140: and solving the total cost objective function, and taking the maintenance trigger threshold value when the value of the total cost objective function is minimum as the target trigger threshold value.
In the total cost objective function, the maintenance trigger threshold is a variable, and the value of the maintenance trigger threshold can be adjusted in the range of 0 to 1 to obtain different values of the total cost objective function, so that the corresponding target maintenance trigger threshold when the value of the total cost objective function is minimum can be determined.
S150: and taking the equipment with the failure prediction probability larger than or equal to the target trigger threshold value as the target equipment needing maintenance.
The target device is a device that needs to be maintained from the time of feature data acquisition, and preferably, the maintenance work is completed within a predetermined time period from the time of feature data acquisition.
According to the method for determining the target equipment, the prediction probability of failure of each equipment in a preset time period from the characteristic data acquisition time is predicted according to the characteristic data of the plurality of equipment, the equipment with the prediction probability larger than or equal to the target maintenance trigger threshold value is used as the target equipment, and the target equipment obtained according to the target maintenance trigger threshold value is maintained in advance to enable the total maintenance cost of all the equipment to be the lowest due to the fact that the target maintenance trigger threshold value corresponds to the situation that the value of the total cost target function is the minimum; the failure rate of the equipment and the time which cannot be used when the equipment fails can be reduced; during maintenance at every turn, only need maintain the target equipment can, need not to carry out inspection one by one to every equipment, reduced the work load of maintenance.
In some embodiments, as shown in fig. 2, step S140 may include the following steps.
S141: and taking 0 to 1 as a target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the target interval, substituting the numerical value into a total cost target function, and solving a total cost standard function numerical value.
S142: and selecting the maintenance trigger threshold corresponding to the minimum total cost target function value as a reference value.
S143: and taking an interval with a preset length taking the current reference value as a center as a current target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the current target interval, substituting the numerical value into the total cost target function, solving the total cost target function value, and selecting the maintenance trigger threshold value corresponding to the minimum total cost target function value as the current reference value.
The predetermined interval is different every time step S143 is performed. Typically, the predetermined interval is less than half of the current target interval.
S144: and judging whether the current reference value is equal to the current reference value determined in the last cycle. In the case of equality, step S145 is executed; otherwise, the step S143 is continuously performed.
When the determination of step S144 is performed for the first time, the reference value determined at step S142 may be used as the current reference value determined in the previous cycle.
S145: and taking the determined current reference value as a target trigger threshold value.
In some embodiments, as shown in fig. 3, the step S120 may obtain the prediction probability by the following steps.
S310: characteristic data of a plurality of devices is input into a pre-established prediction model.
S320: and taking the output of the prediction model as the prediction probability of the fault of each device in a preset time period from the characteristic data acquisition moment.
In some embodiments, as shown in fig. 4, the prediction model in step S310 may be trained as follows.
S410: a sample pair of a plurality of devices is obtained, the sample pair comprising characteristic data and a flag value for marking whether the device has a fault within a predetermined time period from the time of characteristic data acquisition.
Please refer to S110 for description of the feature data.
In some embodiments, the prediction model may also predict the number of times each device failed within a predetermined time period from the time of feature data acquisition and a corresponding prediction probability, and accordingly, when training the prediction model, the sample pair may further include the number of times each device failed within a predetermined time period from the time of feature data acquisition.
S420: the predictive model is trained based on the sample pairs for the plurality of devices.
The prediction model may be any machine learning model, such as an RNN model, a decision tree model, or the like. The decision tree model may be, for example, a lightGBM model.
Under the condition of adopting the lightGBM model, a histogram algorithm can be adopted to select a data set on a leaf node to be segmented, and the data set on the leaf node is segmented.
An embodiment of the present specification provides a method for clustering, as shown in fig. 5, to select a data set on a leaf node to be segmented, where the method includes the following steps.
S510: the sum of the mutual distances between all data in the data set on each leaf node is calculated.
I.e. calculate the distance between every two data in the set and sum all the distances together.
For example, there are two sets of data on leaf nodes U1: (1, 5, 7, 4) and U2: (1, 8, 4, 20).
Calculating the sum of distances d1 corresponding to U1: the sum of the distances from 1 to the rest of the data is 4+6+3 to 13, the sum of the distances from 5 to the rest of the data is 4+2+1 to 7, the sum of the distances from 7 to the rest of the data is 6+2+3 to 11, the sum of the distances from 4 to the rest of the data is 3+1+3 to 7, and the sum of the distances from U1 to d1 is 13+7+11+7 to 38.
Calculating the sum of distances d2 corresponding to U2: the sum of the distances between 1 and the other data is 7+3+19 to 29, the sum of the distances between 8 and the other data is 7+4+12 to 23, the sum of the distances between 4 and the other data is 3+4+16 to 23, the sum of the distances between 20 and the other data is 19+12+16 to 47, and the sum of the distances corresponding to U2 is d2 to 29+23 +47 to 122.
S520: and taking the data set with the maximum mutual distance between the data as the data set on the leaf node to be segmented.
Following the above example, since d2> d1, U2 was chosen as the leaf node to be split.
An embodiment of the present specification provides a method based on clustering, which divides a first data set on a leaf node to obtain a second data set on the leaf node, as shown in fig. 6, and includes the following steps.
S610: the sum of the distances between each data in the first data set and the remaining data is calculated.
S620: and taking the data with the largest sum of the distances from the rest data as the data in the second data set.
S630: and calculating the mean value of the distances between each data point in the first data set and the rest data in the first data set as first data, calculating the mean value of the distances between each data point in the first data set and all data points in the second data set as second data, and subtracting the second data from the first data.
S640: and dividing the data point with the maximum difference value into a second data set.
S650: and judging whether the difference value is a negative value. If so, ending the segmentation; in the case of no, execution proceeds to step S630.
Following the above example, for the first data set U2: (1, 8, 4, 20) the segmentation was performed as follows:
and calculating to obtain: the sum of the distances from 1 to the rest of the data is 7+3+19 to 29, the sum of the distances from 8 to the rest of the data is 7+4+12 to 23, the sum of the distances from 4 to the rest of the data is 3+4+16 to 23, and the sum of the distances from 20 to the rest of the data is 19+12+16 to 47; the sum of the distances 47 is maximal, so that the data 20 corresponding to 47 is regarded as the data in the new segmented data set, i.e. the second data set U3. After this division, the first data set U2: (1, 8, 4), a second data set U3 (20).
In calculating the first data set U2: the distance between 1 and the rest of the data is 7+ 3-10; the distance between 8 and each of the other data is 7+ 4-11, and the distance between 4 and each of the other data is 3+ 4-7; the mean value of the distances is (10+11+7) ÷ 3 ═ 9.333. In calculating the first data set U2: the distance between 1 and 20 in the second data set U3 is 19, the distance between 8 and 20 in the second data set U3 is 12, and the distance between 4 and 20 in the second data set U3 is 16; the mean value of the distances is (19+12+16) ÷ 3 ═ 15.666.
Since 9.333-15.666<0, the segmentation is ended, and the first data set U2 is obtained after the final segmentation: (1, 8, 4), a second data set U3 (20).
In some embodiments, before training the lightGBM model based on a plurality of device sample pairs, feature data needs to be screened to screen out a feature data type having a large influence on the value of the prediction probability, and a feature data type having a small influence on the value of the prediction probability is discarded, so that the calculation amount of the model can be reduced. A manual experience screening method may be adopted, or an embodiment of the present specification provides a screening method for a feature data type, as shown in fig. 7, the screening method includes the following steps:
s710: the sample pairs are input into the lightGBM model.
The lightGBM model in this step uses initial parameter values.
After the feature data are screened out, the process of training the model is the process of adjusting parameter values in the model.
S720: the number of times the feature data for each type is split as a leaf node is determined.
S730: and screening out a plurality of types of characteristic data according to the splitting times.
The screening method of the feature data takes the splitting times of each type of feature data as leaf nodes as the importance of the type of feature data, and can arrange the importance from large to small to screen out a preset number (for example, the first 10) of feature data in the front row; the importance may also be ranked from large to small, and the ratio of the number of feature data screened from the previous column to the number of all feature data is a predetermined ratio, for example, the feature data screened 50% of the importance is screened.
Corresponding to the above steps S710 to S730, when the lightGBM model is trained, the lightGBM model is trained by using a sample composed of the selected multiple types of feature data and the labeled values.
The embodiment of the present specification provides a target device determination apparatus, which may be used to implement the target device determination method described in fig. 1. As shown in fig. 8, the apparatus includes an obtaining module 10, a predicting module 20, an establishing module 30, a solving module 40, and a first determining module 50.
The acquisition module 10 includes acquiring feature data of a plurality of devices.
The prediction module 20 obtains the prediction probability of the failure of each device within a predetermined time period from the characteristic data acquisition time according to the characteristic data of the plurality of devices.
The establishing module 30 is configured to establish a total cost objective function with a maintenance trigger threshold as a variable according to the predicted failure probability of each device in the plurality of devices, the maintenance cost in the case of failure of each device, and the maintenance cost of each device.
The solving module 40 is configured to solve the total cost objective function, and set the maintenance trigger threshold when the total cost objective function takes the minimum value as the target trigger threshold.
The first determination module 50 is configured to determine, as a target device to be maintained, a device having a failure prediction probability greater than or equal to a target trigger threshold.
In some embodiments, the total cost objective function with the maintenance trigger threshold as a variable includes one of:
where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykThe maintenance cost is the maintenance cost under the condition that the equipment fails;
where K is the number of devices with a prediction probability greater than or equal to the maintenance trigger threshold, paykThe maintenance cost under the condition that equipment fails is obtained, and n is the maximum failure frequency corresponding to the non-zero prediction probability;
wherein, 0 < pk< 1 and pkλ is more than or equal to λ, and λ is a maintenance trigger threshold; alpha is a predetermined coefficient, K is the number of devices with a predicted probability greater than or equal to a maintenance trigger threshold, paykThe maintenance cost is the maintenance cost under the condition that the equipment fails;
wherein, 0 < pk< 1 and pkλ is more than or equal to λ, and λ is a maintenance trigger threshold; alpha is a predetermined coefficient, K is the number of devices with a predicted probability greater than or equal to a maintenance trigger threshold, paykAnd n is the maximum failure frequency corresponding to the non-zero prediction probability, which is the maintenance cost under the condition that the equipment fails.
In some embodiments, the solving module 40 includes a solving submodule 41, a first determining submodule 42, and a second determining submodule 43.
The obtaining submodule 41 is configured to use 0 to 1 as a target interval, select a value as a maintenance trigger threshold at predetermined intervals in the target interval, substitute the value into the total cost target function, and obtain a total cost target function value.
The first determining submodule 42 is configured to select the maintenance trigger threshold corresponding to the smallest total cost objective function value as the reference value.
The finding submodule 41 and the first determining submodule 42 are also arranged to cyclically carry out the following steps until the determined current reference value is equal to the current reference value determined in the previous cycle: and taking an interval with a preset length taking the current reference value as a center as a current target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the current target interval, substituting the numerical value into the total cost target function, solving the total cost target function value, and selecting the maintenance trigger threshold value corresponding to the minimum total cost target function value as the current reference value.
The second determination submodule 43 is configured to use the determined current reference value as the target trigger threshold.
In some embodiments, the prediction module 20 includes an input submodule 21 and a third determination submodule 22.
The input sub-module 21 is used to input the characteristic data of a plurality of devices into a pre-established prediction model. The third determination submodule 22 is arranged to use the output of the predictive model as a predictive probability of each device failing within a predetermined time period from the time of feature data acquisition.
In some embodiments, the apparatus further comprises a training module 60 for training the predictive model. Training module 60 may include an acquisition sub-module 61 and a training sub-module 62.
The obtaining submodule 61 is configured to obtain a plurality of sample pairs of the device, where each sample pair includes feature data and a flag value, and the flag value is used to flag whether the device has a fault within a predetermined time period from the time of feature data acquisition. The training submodule 62 is configured to train the predictive model based on the sample pairs for the plurality of devices.
In some embodiments, in the case that the prediction model is a lightGBM model, the training module 60 further includes a selecting submodule 63 for selecting a data set on a leaf node to be segmented in the lightGBM model.
The selection submodule 63 may include a first calculation submodule 631 and a fourth determination submodule 632.
The first calculation submodule 631 is configured to calculate a sum of mutual distances between all data in the data set on each leaf node. The fourth determining submodule 632 is configured to use the data set with the largest mutual distance between the data as the data set on the leaf node to be segmented.
In some embodiments, in the case that the prediction model is a lightGBM model, the training module 60 further includes a partitioning submodule 64 for partitioning the first data set on the leaf node into the second data set on the leaf node according to the following method.
The segmentation submodule 64 may include a second calculation submodule 641 and a fifth determination submodule 642.
The second calculating submodule 641 is configured to calculate a sum of distances between each data in the first data set and the remaining data.
The fifth determining submodule 642 is configured to determine, as data in the second data set, data having a largest sum of distances to the remaining data.
The second calculating sub-module 641 and the fifth determining sub-module 642 are further configured to repeatedly perform the following steps until the difference value is a negative value: calculating the mean value of the distances between each data point in the first data set and the rest data in the first data set as first data, calculating the mean value of the distances between each data point in the first data set and all data points in the second data set as second data, and calculating the difference value of subtracting the second data from the first data; and dividing the data point with the maximum difference value into a second data set.
In some embodiments, in case the predictive model is a lightGBM model, the apparatus further comprises an input module 70, a second determination module 80 and a screening module 90.
The input module 70 is used to input the sample pairs into the lightGBM model.
The second determination module 80 is used to determine the number of times each type of feature data is split as a leaf node.
The filtering module 90 is configured to filter out a plurality of types of feature data according to the number of splits.
Correspondingly, when the lightGBM model is trained, the lightGBM model is trained by adopting a sample consisting of the screened multiple types of feature data and the labeled values.
The details of the above-mentioned apparatus can be understood by referring to the corresponding related descriptions and effects in the embodiments of fig. 1 to fig. 7, which are not described herein again.
An embodiment of the present invention further provides an electronic device, as shown in fig. 9, the electronic device may include a processor 91 and a memory 92, where the processor 91 and the memory 92 may be connected by a bus or in another manner, and fig. 9 takes the example of connection by a bus as an example.
The processor 91 may be a Central Processing Unit (CPU). The Processor 91 may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, or any combination thereof.
The memory 92, as a non-transitory computer readable storage medium, may be used for storing non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the determination method of the target device in the embodiment of the present invention (for example, the obtaining module 10, the predicting module 20, the establishing module 30, the solving module 40, and the first determining module 50 shown in fig. 8). The processor 91 executes various functional applications and data classification of the processor by executing non-transitory software programs, instructions and modules stored in the memory 92, namely, realizes the determination method of the target device in the above method embodiment.
The memory 92 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created by the processor 91, and the like. Further, memory 92 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 92 may optionally include memory located remotely from the processor 91, and such remote memory may be connected to the processor 91 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The one or more modules are stored in the memory 92 and, when executed by the processor 91, perform a method of determining a target device as in the embodiment of fig. 1-3.
The details of the electronic device may be understood by referring to the corresponding descriptions and effects in the embodiments of fig. 1 to fig. 7, which are not described herein again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk (Hard Disk Drive, abbreviated as HDD), a Solid State Drive (SSD), or the like; the storage medium may also comprise a combination of memories of the kind described above.
In the 90 s of the 20 th century, improvements in a technology could clearly distinguish between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate a dedicated integrated circuit chip 2. Furthermore, nowadays, instead of manually making an Integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Language Description Language), traffic, pl (core unified Programming Language), HDCal, JHDL (Java Hardware Description Language), langue, Lola, HDL, laspam, hardbyscript Description Language (vhr Description Language), and the like, which are currently used by Hardware compiler-software (Hardware Description Language-software). It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from the other embodiments.
The systems, devices, modules or units described in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions.
For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functionality of the units may be implemented in one or more software and/or hardware when implementing the present application.
From the above description of the embodiments, it is clear to those skilled in the art that the present application can be implemented by software plus necessary general hardware platform. Based on such understanding, the technical solutions of the present application may be essentially or partially implemented in the form of software products, which may be stored in a storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and include instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods of some parts of the embodiments of the present application.
The application is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
Although the present application has been described in terms of embodiments, those of ordinary skill in the art will recognize that there are numerous variations and permutations of the present application without departing from the spirit of the application, and it is intended that the appended claims encompass such variations and permutations without departing from the spirit of the application.
Claims (11)
1. A method for determining a target device, comprising:
acquiring characteristic data of a plurality of devices;
according to the characteristic data of the plurality of devices, obtaining the prediction probability of the failure of each device in a preset time period from the characteristic data acquisition time;
establishing a total cost target function with a maintenance trigger threshold as a variable according to the prediction probability of the fault of each device in the plurality of devices, the maintenance cost under the condition of the fault of each device and the maintenance cost of each device;
solving the total cost objective function, and taking the maintenance trigger threshold value when the value of the total cost objective function is minimum as a target trigger threshold value;
and taking the equipment with the failure prediction probability larger than or equal to the target trigger threshold value as the target equipment needing maintenance.
2. The method of claim 1, wherein the total cost objective function in terms of maintenance trigger threshold as a variable comprises:
3. The method of claim 1, wherein solving the total cost objective function to obtain the target trigger threshold as the maintenance trigger threshold when the total cost objective function has the minimum value comprises:
taking 0 to 1 as a target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the target interval, substituting the numerical value into the total cost target function, and obtaining a total cost target function value;
selecting a maintenance trigger threshold corresponding to the minimum total cost target function value as a reference value;
the following steps are executed in a loop until the determined current reference value is equal to the current reference value determined in the previous loop: taking an interval with a preset length taking a current reference value as a center as a current target interval, selecting a numerical value as a maintenance trigger threshold value at preset intervals in the current target interval, substituting the numerical value into the total cost target function, solving a total cost target function value, and selecting a maintenance trigger threshold value corresponding to the minimum total cost target function value as a current reference value;
and taking the determined current reference value as the target trigger threshold.
4. The method of claim 1, wherein obtaining the predicted probability of failure of each device within a predetermined time period from the time of feature data acquisition based on the feature data of the plurality of devices comprises:
inputting the characteristic data of the plurality of devices into a pre-established prediction model;
and taking the output of the prediction model as the prediction probability of the fault of each device in a preset time period from the characteristic data acquisition moment.
5. The method of claim 4, wherein the predictive model is trained as follows:
obtaining a sample pair of a plurality of devices, wherein the sample pair comprises characteristic data and a marking value, and the marking value is used for marking whether the device has a fault within a preset time period from the characteristic data acquisition time;
training the predictive model based on the sample pairs for the plurality of devices.
6. The method of claim 4, wherein in the case that the prediction model is a lightGBM model, selecting the data set on the leaf node to be partitioned in the lightGBM model comprises:
calculating the sum of the mutual distances among all data in the data set on each leaf node;
and taking the data set with the maximum mutual distance between the data as a data set on the leaf node to be segmented.
7. The method of claim 5, wherein in the case that the prediction model is a lightGBM model, the first data set at a leaf node is segmented into the second data sets at the leaf nodes according to the following method:
calculating the sum of the distances between each data in the first data set and the other data;
taking the data with the largest sum of the distances between the data and the rest data as the data in the second data set;
repeatedly executing the following steps until the difference value is a negative value:
calculating the mean value of the distances between each data point in the first data set and the rest data in the first data set as first data, calculating the mean value of the distances between each data point in the first data set and all data points in the second data set as second data, and calculating the difference value of subtracting the second data from the first data;
and dividing the data point with the maximum difference value into a second data set.
8. The method of claim 5, further comprising, prior to training the lightGBM model based on the sample pairs of the plurality of devices:
inputting the sample pairs into a lightGBM model;
determining the times of splitting of the characteristic data of each type as leaf nodes;
screening out a plurality of types of feature data according to the splitting times;
correspondingly, when the lightGBM model is trained, the lightGBM model is trained by adopting a sample consisting of the screened multiple types of feature data and the labeled values.
9. An apparatus for determining a target device, comprising:
the acquisition module acquires characteristic data of a plurality of devices;
the prediction module is used for obtaining the prediction probability of the failure of each device in a preset time period from the characteristic data acquisition time according to the characteristic data of the plurality of devices;
the establishing module is used for establishing a total cost objective function taking a maintenance trigger threshold value as a variable according to the prediction probability of the fault of each device in the plurality of devices, the maintenance cost under the condition of the fault of each device and the maintenance cost of each device;
the solving module is used for solving the total cost objective function and taking the maintenance trigger threshold value when the value of the total cost objective function is minimum as a target trigger threshold value;
and the first determining module is used for taking the equipment with the failure prediction probability larger than or equal to the target trigger threshold value as the target equipment needing to be maintained.
10. An electronic device, comprising:
a memory and a processor, the processor and the memory being communicatively connected to each other, the memory having stored therein computer instructions, the processor implementing the steps of the method of any one of claims 1 to 8 by executing the computer instructions.
11. A computer storage medium storing computer program instructions which, when executed, implement the steps of the method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111004041.4A CN113705896A (en) | 2021-08-30 | 2021-08-30 | Target equipment determination method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111004041.4A CN113705896A (en) | 2021-08-30 | 2021-08-30 | Target equipment determination method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113705896A true CN113705896A (en) | 2021-11-26 |
Family
ID=78656802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111004041.4A Pending CN113705896A (en) | 2021-08-30 | 2021-08-30 | Target equipment determination method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113705896A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114185375A (en) * | 2021-12-06 | 2022-03-15 | 安徽新科水处理设备有限公司 | Water temperature monitoring and controlling system for drinking water table |
| CN117409495A (en) * | 2023-12-11 | 2024-01-16 | 北汽利戴工业技术服务(北京)有限公司 | Optimal maintenance time acquisition method and system based on equipment maintenance data |
-
2021
- 2021-08-30 CN CN202111004041.4A patent/CN113705896A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114185375A (en) * | 2021-12-06 | 2022-03-15 | 安徽新科水处理设备有限公司 | Water temperature monitoring and controlling system for drinking water table |
| CN117409495A (en) * | 2023-12-11 | 2024-01-16 | 北汽利戴工业技术服务(北京)有限公司 | Optimal maintenance time acquisition method and system based on equipment maintenance data |
| CN117409495B (en) * | 2023-12-11 | 2024-02-20 | 北汽利戴工业技术服务(北京)有限公司 | Optimal maintenance time acquisition method and system based on equipment maintenance data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11403164B2 (en) | Method and device for determining a performance indicator value for predicting anomalies in a computing infrastructure from values of performance indicators | |
| US10069684B2 (en) | Core network analytics system | |
| US8365019B2 (en) | System and method for incident management enhanced with problem classification for technical support services | |
| US11532056B2 (en) | Deep convolutional neural network based anomaly detection for transactive energy systems | |
| US20160212007A1 (en) | Distributed map reduce network | |
| CN103116531A (en) | Storage system failure predicting method and storage system failure predicting device | |
| CN108334997B (en) | Standby optimization method and device based on support fault event constraint unit combination | |
| CN113705896A (en) | Target equipment determination method and device and electronic equipment | |
| CN104021264A (en) | Defect prediction method and device | |
| US11307916B2 (en) | Method and device for determining an estimated time before a technical incident in a computing infrastructure from values of performance indicators | |
| US11860721B2 (en) | Utilizing automatic labelling, prioritizing, and root cause analysis machine learning models and dependency graphs to determine recommendations for software products | |
| CN105488539A (en) | Generation method and device of classification method, and estimation method and device of system capacity | |
| CN109885456A (en) | A kind of polymorphic type event of failure prediction technique and device based on system log cluster | |
| CN114580263A (en) | Knowledge graph-based information system fault prediction method and related equipment | |
| Bogojeska et al. | Classifying server behavior and predicting impact of modernization actions | |
| CN111930526B (en) | Load prediction method, load prediction device, computer equipment and storage medium | |
| US11736363B2 (en) | Techniques for analyzing a network and increasing network availability | |
| CN111949429A (en) | Server fault monitoring method and system based on density clustering algorithm | |
| CN105471647A (en) | Power communication network fault positioning method | |
| CN111027591B (en) | Node fault prediction method for large-scale cluster system | |
| US20210097432A1 (en) | Gpu code injection to summarize machine learning training data | |
| CN101495978B (en) | Reduction of message flow between bus-connected consumers and producers | |
| CN113837635A (en) | Risk detection processing method, device and equipment | |
| CN113569955A (en) | Model training method, user portrait generation method, device and equipment | |
| WO2021130298A1 (en) | System, apparatus and method for managing energy consumption at a technical installation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |