A method and device for predicting equipment status information

Abstract

This invention discloses a method, apparatus, readable medium, and electronic device for predicting device status information. The method includes: acquiring feature data of a target device; generating a hash table of relevant data for the target device based on a hash function and the feature data of the target device, wherein the relevant data includes feature data of detection points of non-target devices, and the feature data of the detection points is non-shared data; calculating the weight of the feature data of the detection points based on the hash table; establishing a joint model using the weight of the feature data of the detection points and the device status labels corresponding to the feature data of the detection points; and using the joint model to predict the device status of the target device. The technical solution provided by this invention can migrate data from non-target devices to the target device, establish the relationship between the feature data of the target device and the device status, and ensure data security by eliminating the need for shared feature data between devices.

Description

Technical Field

[0001] The present invention relates to the field of energy technology, and in particular to a method and device for predicting equipment status information. Background Art

[0002] A large number of industrial equipment are needed in the intelligent manufacturing system. In order to monitor the industrial equipment and understand the operation of the industrial equipment, and then ensure the normal operation of the intelligent manufacturing system, it is usually necessary to predict the equipment status of the industrial equipment. Here, equipment status prediction refers to predicting the probability of industrial equipment failure or the remaining service life of industrial equipment based on the characteristic data of industrial equipment.

[0003] At present, by collecting historical operating data of several industrial equipment, using machine learning methods to establish a mapping relationship between these historical operating data and the equipment operating status, an equipment status prediction model is obtained, and the equipment status prediction model is used to realize the equipment status prediction of other industrial equipment.

[0004] However, the above technical solution requires sharing of historical operation data of several industrial devices, thus resulting in low data security among the several industrial devices. Summary of the invention

[0005] In view of the above-mentioned technical problems existing in the prior art, the present invention provides a method, device, computer-readable storage medium and electronic device for predicting device status information, and the feature data of the device does not need to be shared between non-target devices and target devices, thereby ensuring data security.

[0006] In a first aspect, the present invention provides a method for predicting device status information, comprising:

[0007] Obtain characteristic data of the target device;

[0008] Generate a hash table of relevant data of the target device according to the hash function and the characteristic data of the target device, wherein the relevant data includes characteristic data of detection points of non-target devices, and the characteristic data of the detection points are non-shared data;

[0009] Calculating the weight of the feature data of the detection point according to the hash table;

[0010] Establishing a joint model by using the weights of the feature data of the detection points and the device status labels corresponding to the feature data of the detection points;

[0011] The joint model is used to predict the device state of the target device.

[0012] In a second aspect, the present invention provides a device for predicting device status information, comprising:

[0013] A data acquisition module, used to acquire characteristic data of a target device;

[0014] A generating module, configured to generate a hash table of relevant data of the target device according to a hash function and characteristic data of the target device, wherein the relevant data includes characteristic data of detection points of non-target devices, and the characteristic data of the detection points are non-shared data;

[0015] A weight calculation module, used for calculating the weight of the feature data of the detection point according to the hash table;

[0016] A model building module, used to build a joint model by using the weights of the feature data of the detection points and the device status labels corresponding to the feature data of the detection points;

[0017] The prediction module is used to use the joint model to predict the device state of the target device.

[0018] In a third aspect, the present invention provides a computer-readable storage medium comprising execution instructions. When a processor of an electronic device executes the execution instructions, the processor executes any method described in the first aspect or any method described in the second aspect.

[0019] In a fourth aspect, the present invention provides an electronic device comprising a processor and a memory storing execution instructions, wherein when the processor executes the execution instructions stored in the memory, the processor executes any method described in the first aspect or any method described in the second aspect.

[0020] The present invention provides a method, device, computer-readable storage medium and electronic device for predicting device status information. The method obtains feature data of a target device, and then generates a hash table of relevant data of the target device based on a hash function of the target device and the feature data of the target device, wherein the relevant data includes feature data of a detection point of a non-target device, and the feature data of the detection point is non-shared data. Then, according to the hash table, the weight of the feature data of the detection point is calculated, and then a joint model is established using the weight of the feature data of the detection point and the device status label corresponding to the feature data of the detection point. After that, the joint model is used to predict the device status of the target device. The technical solution provided by the present invention migrates the feature data of the non-target device and the device status label corresponding to the feature data to the target device, establishes a relationship between the feature data and the device status of the target device, and does not require shared feature data between devices, thereby ensuring data security.

[0021] The further effects of the above-mentioned non-conventional preferred manner will be described below in conjunction with specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the existing technical solutions, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0023] Figure 1 A schematic diagram of a flow chart of a method for predicting device status information provided by an embodiment of the present invention;

[0024] Figure 2 A schematic diagram of a flow chart of another method for predicting device status information provided by an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of the structure of a device for predicting device status information provided by an embodiment of the present invention;

[0026] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] like Figure 1 As shown, an embodiment of the present invention provides a method for predicting device status information, comprising the following steps:

[0029] Step 101: Acquire characteristic data of a target device.

[0030] Specifically, the target device is any industrial equipment of the intelligent manufacturing system, and the embodiment of the present invention does not specifically limit this. Preferably, the device to be predicted is an energy device, such as a gas boiler, an internal combustion engine, a steam turbine, a cogeneration device, a photovoltaic device, etc.

[0031] Specifically, there are multiple feature data of the target device, and each feature data includes several features corresponding to the feature values ​​of the target device. Among them, several features are influencing factors affecting the device status, which need to be determined in combination with the actual scenario. For example, if the target device is a gas boiler, the several features include but are not limited to gas flow, gas temperature, exhaust temperature, flue gas flow, gas pressure, start-up and shutdown status, flue gas humidity and flue gas pressure.

[0032] Specifically, this step can be implemented by the detection point of the target device in the Internet of Things, and the detection point obtains multiple feature data of the target device through the sensor installed on the target device. Among them, the detection point of the target device can be understood as the node closest to the target device and capable of data processing and data interaction, including but not limited to any one or more of the edge server, edge gateway and edge controller.

[0033] Step 102: Generate a hash table of relevant data of the target device according to a hash function and feature data of the target device, wherein the relevant data includes feature data of detection points of non-target devices, and the feature data of the detection points are non-shared data.

[0034] Specifically, for each feature data of the target device, the feature data is converted through a hash function to obtain a hash value of the feature data; and a hash table is obtained through the hash values ​​of all feature data of the target device. The hash value can be understood as the value obtained after the feature data of the target device is keylessly encrypted through a hash function. It should be noted that, considering that the hash function can realize keyless encryption processing of data relatively quickly and simply, the embodiment of the present invention selects a hash function to perform keyless encryption on the feature data of the target device.

[0035] Specifically, the target device and the non-target device have the same device type. For example, both the target device and the non-target device are gas boilers. However, information such as the production unit and device model may be different.

[0036] Specifically, the detection point of the non-target device is the node of the non-target device in the Internet of Things, and the detection point obtains multiple feature data of the non-target device through the sensor installed on the non-target device. Among them, the detection point can be understood as the node closest to the non-target device and capable of data processing and data interaction, including but not limited to any one or more of the edge server, edge gateway and edge controller. It should be noted that there can be multiple non-target devices, and each non-target device corresponds to a detection point.

[0037] Specifically, there are multiple feature data of non-target devices, each feature data includes several features corresponding to feature values ​​of the non-target devices. Here, the feature data of the non-target devices and the feature data of the target devices correspond to the same features.

[0038] It should be noted that the characteristic data of the detection point is non-shared data, which means that the characteristic data of the detection point does not leave the detection point, thereby ensuring the data security of the detection point. At the same time, there is no need to share data between the target device and the non-target device, ensuring data security.

[0039] Step 103: Calculate the weight of the feature data of the detection point according to the hash table.

[0040] In this embodiment, the data of the non-target device is migrated to the target device through the weight of the feature data of the detection point, and the data association between the non-target device and the target device is established. Here, the feature data of the detection point is non-shared data. In other words, the target device will not directly obtain the feature data of the detection point of the non-target device, that is, there is no data sharing between the target device and the non-target device, thereby ensuring data security.

[0041] In one embodiment, the weight of the feature data of the detection point can be determined in the following manner:

[0042] Determine the hash value of the characteristic data of the detection point; determine the distribution probability of the characteristic data of the detection point corresponding to the target device based on the hash table and the hash value of the characteristic data of the detection point; determine the distribution probability of the characteristic data of the detection point corresponding to the non-target device; determine the ratio of the distribution probability of the characteristic data of the detection point corresponding to the target device and the distribution probability corresponding to the non-target device as the weight of the characteristic data of the detection point.

[0043] In this embodiment, the weight of the feature data of the detection point is determined by the distribution probability of the feature data of the detection point corresponding to the target device and the distribution probability corresponding to the non-target device, so that the feature data of the detection point is migrated to the target device. At the same time, there is no need to share the feature data of the target device, thereby ensuring data security.

[0044] Specifically, the hash value of the characteristic data of the detection point can be understood as the value obtained after the characteristic data of the detection point is encrypted without a key, so there is no need to share the characteristic data of the detection point, and only the hash value of the characteristic data of the detection point needs to be shared, thereby ensuring data security. It should be noted that when the parameters in the hash function are different, the difficulty of analyzing the hash value of the characteristic data of the detection point and the hash table may increase. Therefore, the hash value of the characteristic data of the detection point is obtained by performing keyless encryption based on the hash function in step 101, thereby reducing the difficulty of analyzing the hash value and the hash table.

[0045] Optionally, the hash value of the feature data of the detection point of the non-target device is obtained by the detection point. The detection point can obtain the feature data of the non-target device, and then keylessly encrypt each feature data based on the hash function in step 101 to obtain the hash value corresponding to each feature data of the detection point.

[0046] Optionally, the detection point of the non-target device sends the hash value of the characteristic data of the non-target device to the detection point of the target device. The detection point of the target device determines the distribution probability of the characteristic data of the non-target device corresponding to the target device based on the hash table and the hash value of the characteristic data of the non-target device, and sends the distribution probability to the detection point of the non-target device. The detection point of the non-target device determines the distribution probability of the characteristic data corresponding to the non-target device for each characteristic data of the non-target device, and determines the weight of the characteristic data by the ratio of the distribution probability of the characteristic data corresponding to the target device to the distribution probability corresponding to the non-target device.

[0047] Optionally, according to the hash table and the hash value of the characteristic data of the detection point, the replacement data of the characteristic data of the detection point is determined from the characteristic data of the target device, and the distribution probability of the replacement data in the characteristic data of the target device is determined as the distribution probability of the characteristic data of the detection point corresponding to the target device. Here, through the hash value and the hash table, the replacement data corresponding to the characteristic data of the detection point is simply and quickly determined from the multiple characteristic data of the target device, thereby establishing a data association between the target device and the non-target device. As a possible implementation method, based on the local sensitive hashing algorithm, the hash table and the hash value of the characteristic data of the detection point are analyzed for similarity to determine the replacement data of the characteristic data of the detection point from the characteristic data of the target device. The local sensitive hashing algorithm can save computing time and improve computing efficiency. It should be noted that when multiple replacement data of the characteristic data of the detection point are determined from the characteristic data of the target device, one can be randomly selected as the replacement data.

[0048] Optionally, the distribution probability is determined based on a non-parametric estimation method. Here, considering that the non-parametric estimation method does not need to rely on any prior assumptions of the data, but is determined based on the data itself, and is simpler than the parameter estimation method, the embodiment of the present invention selects the non-parametric estimation method to determine the distribution probability. It should be further explained that the present invention is not intended to limit the non-parametric estimation method, which can be determined in combination with actual conditions. Considering that the amount of feature data in the embodiment of the present invention is relatively large, in order to more accurately estimate the data distribution, the kernel density estimation method is preferred. In other words, the distribution probability is determined based on the kernel density estimation method.

[0049] Step 104: Establish a joint model using the weights of the feature data of the detection points and the device status labels corresponding to the feature data of the detection points.

[0050] In this embodiment, a joint model is established by using the weights corresponding to the feature data of the detection points and the device status labels corresponding to the feature data of the detection points, without sharing the feature data of the target device, thereby ensuring data security.

[0051] It should be noted that the weight of the feature data of the detection point establishes a data connection between the feature data of the target device and the feature data of the detection point of the non-target device, thereby migrating the feature data of the detection point of the non-target device to the target device, ensuring data security.

[0052] It should also be noted that the data of the target device and the non-target device are distributed at different detection points in the Internet of Things. Sharing data for model training will cause data security issues. The embodiment of the present invention replaces the feature data of the target device with the feature data of the detection point of the non-target device, so that there is no data sharing between the non-target device and the target device, thereby avoiding the data security issues caused by directly sharing data.

[0053] In one embodiment, the joint model may be determined by the following method:

[0054] The local model of the detection point is determined by using the weight of the feature data of the detection point and the device status label of the feature data of the detection point; and a joint model is established based on the local models of each detection point.

[0055] Specifically, model training is performed based on the weights and device status labels corresponding to the feature data of the detection points to determine the local model. Then, the local models of multiple detection points are fused to establish a joint model.

[0056] Specifically, the feature data of the detection point and the weight corresponding to the feature data of the detection point can replace the feature data of the target device. Subsequently, the model parameters are adjusted by the weights corresponding to multiple feature data of the detection point, so that the adjusted model parameters can reflect the relationship between the feature data and the device status of the target device, without involving the sharing of the feature data of the detection point, thereby ensuring data security.

[0057] Specifically, the device status information label can be the fault information of the target device or the remaining service life of the target device. For example, the device status information label can be faulty or normal, or it can be the type of fault, or it can be the degree of fault. The embodiment of the present invention does not limit this, and it needs to be determined in combination with actual needs. It should be noted that the embodiment of the present invention does not intend to limit the method for obtaining the device status label corresponding to each characteristic data of the detection point. It can also be manual labeling, rule labeling, or cluster labeling. The above labeling methods are all existing technologies and will not be described in detail here.

[0058] In one embodiment, based on a joint learning algorithm, the local models of each detection point are iterated and then fused to establish a joint model.

[0059] Specifically, the detection points corresponding to the non-target devices perform model training based on the weights and device status labels corresponding to the feature data of the detection points to determine the local model. Then, the detection points corresponding to the target device and the detection points corresponding to the non-target devices iterate the local model of the target device based on the joint learning algorithm, and establish a joint model based on the local model after the iteration. Since the detection points corresponding to the non-target devices send the hash values ​​of the feature data of the detection points, and the joint learning algorithm is used to iterate the local model, there is no data sharing between the non-target devices and the target devices, thus ensuring data security.

[0060] Specifically, there are multiple non-target devices, each of which corresponds to a detection point. Each detection point is trained to obtain a local model. The structures of the local models of each detection point are the same, but the model parameters may be different. For example, they are all regression models. It should be noted that the standard machine learning method is to concentrate the feature data of the detection points on a machine or a data center. Joint learning enables the detection points corresponding to multiple non-target devices to collaboratively learn the prediction model of the detection point where the target device is located. At the same time, the detection point where the non-target device is located can also save its own feature data. In addition, there is no need to share the feature data of the detection point of the non-target device.

[0061] Specifically, model fusion can also be the averaging of parameters in the model.

[0062] Specifically, the local model can be a neural network model or a regression model, which needs to be determined based on actual needs.

[0063] Specifically, the local model can be iterated in the following way;

[0064] A1. Perform model training based on multiple feature data of the detection point, the device status label corresponding to each feature data, and the weight corresponding to each feature data to determine the local model;

[0065] A2, determine whether the error of the local model meets the iteration condition, if so, determine the local model as the final model and send it to the detection point of the target device, if not, execute A3;

[0066] A3. Send the model parameters of the local model to the detection point corresponding to the target device;

[0067] A4. Receive the updated model parameters sent by the detection point of the target device, and adjust the updated model parameters according to multiple feature data, the device status labels corresponding to the multiple feature data, and the weights corresponding to the multiple feature data to determine the adjusted model parameters, and replace the model parameters of the local model with the adjusted model parameters to execute A2.

[0068] Step 105: Use the joint model to predict the device status of the target device.

[0069] Specifically, the current feature data of the target device is collected, and the current feature data is substituted into the joint model to determine the device state of the target device.

[0070] In order to better understand the data processing process between the detection points of the target device and the detection points of the non-target device, for example, assume that the detection points corresponding to the three non-target devices are set to A, B and C, and the detection point of the target device is set to D. Taking A as an example, A sends the hash value of the feature data it obtains to D, and receives the distribution probability of the feature data sent by D corresponding to the target device, and determines the weight corresponding to the feature data in combination with the distribution probability of the feature data corresponding to the non-target device. B and C follow a similar processing process to A above to obtain the weights corresponding to the feature data in B and C. After that, A, B, and C perform model training respectively to obtain a local model of the target device. After that, D performs model fusion on the local models sent by A, B, and C to obtain the device state prediction model of the target device.

[0071] It can be seen from the above technical solutions that the embodiments of the present invention have at least the following effective effects: by weighting the feature data of the detection points of the non-target devices, a data association between the non-target devices and the target devices is established, and the data on the non-target devices are migrated to the target devices. At the same time, a joint model is established based on the non-shared data of the detection points, and the state prediction of the target device is achieved using the joint model. There is no need to share the feature data of the devices between the non-target devices and the target devices, thereby ensuring data security.

[0072] Figure 1 What is shown is only a basic embodiment of the method of the present invention. By performing certain optimization and expansion on this basis, other preferred embodiments of the method can be obtained.

[0073] In order to more clearly illustrate the technical solution of the present invention, please refer to Figure 2 The embodiment of the present invention provides another method for predicting device status information. This embodiment is based on the above embodiment and is further described in combination with a specific application scenario. In this embodiment, the following steps may be specifically included:

[0074] Step 201: Generate a hash table of relevant data of the target device according to a hash function and acquired characteristic data of the target device, wherein the relevant data includes characteristic data of detection points of non-target devices, and the characteristic data of the detection points are non-shared data.

[0075] The hash function is an initialized hash function. Assume that the target device corresponds to a detection point n in the Internet of Things, and there are N non-target devices, corresponding to detection points n1, n2, ..., nN in the Internet of Things. n1, n2, ..., nN can obtain the feature data of the corresponding non-target devices respectively, and n can obtain all the feature data of the target device. n performs keyless encryption on each feature data of the target device according to the hash function to obtain a hash table, and sends the hash function to n1, n2, ..., nN. It should be noted that the process of n1, n2, ..., nN obtaining the local model is similar, and the following is described based on the processing process of n1.

[0076] Step 202: Based on the local sensitive hashing algorithm, perform similarity analysis on the hash table and the hash value of the feature data of the detection point to determine replacement data of the feature data of the detection point from the feature data of the target device.

[0077] n1 receives the hash function sent by n, and performs keyless encryption on the feature data of the non-target device according to the hash function to obtain the hash value of the feature data of the detection point, and sends the hash value of the feature data of the detection point to n.

[0078] n receives the hash value of the feature data of the detection point, and based on the local sensitive hashing algorithm, performs similarity analysis on the hash table and the hash value of the feature data of the detection point to determine the replacement data of the feature data of the detection point from the various feature data of the target device, thereby establishing a data connection between the target device and the non-target device.

[0079] Step 203: Determine the distribution probability of the replacement data in the characteristic data of the target device as the distribution probability of the characteristic data of the detection point corresponding to the target device.

[0080] n determines the distribution probability of each feature data of the target device based on the kernel density estimation algorithm, and determines it as the distribution probability of the feature data of the detection point corresponding to the target device, and sends the distribution probability of the feature data of the detection point corresponding to the target device to n1.

[0081] Step 204: Determine the ratio of the distribution probability of the feature data of the detection point corresponding to the target device to the distribution probability corresponding to the non-target device as the weight of the feature data of the detection point.

[0082] n1 receives the distribution probability corresponding to the feature data of the detection point sent by n, and determines the distribution probability of the feature data of the detection point corresponding to the non-target device based on the kernel density estimation method, and determines the ratio of the distribution probability of the feature data of the detection point corresponding to the target device and the distribution probability corresponding to the non-target device as the weight of the feature data of the detection point.

[0083] Here, the distribution probability of the feature data of the detection point corresponding to the non-target device can be understood as the distribution probability of the feature data of the detection point corresponding to all the feature data of the non-target device at the detection point.

[0084] Step 205: Determine a local model of the detection point by using the weight of the feature data of the detection point and the device state label of the feature data of the detection point.

[0085] n1 performs model training based on the weight of the feature data of the detection point and the device status label to determine the local model of the target device. n2, ..., nN respectively determine the local model of the target device using a method similar to n1.

[0086] Step 206: Based on the joint learning algorithm, perform model fusion after model iteration on the local models of each detection point to establish a joint model.

[0087] n and n1, n2, ..., nN, use a joint learning algorithm to iterate the local models trained by n1, n2, ..., nN respectively, and then send them to n. n fuses the iterated local models trained by n1, n2, ..., nN respectively, so as to establish a joint model of the target device.

[0088] Step 207: Use the joint model to predict the device status of the target device.

[0089] nPredict the device status of the target device based on the joint model.

[0090] It can be seen from the above technical solution that the beneficial effect of this embodiment is: the establishment of a joint model of the target device is realized through the detection points of the target device and the non-target device in the Internet of Things. At the same time, the non-target device uses hash values ​​to replace feature data to share information with the target device. The detection points of the target device and the non-target device do not involve the sharing of feature data, thereby ensuring data security.

[0091] Based on the same concept as the device status information prediction method provided by the method embodiment of the present invention, please refer to Figure 3 The present invention implements a prediction device for device status information, which is applied to a joint detection point of a target device, comprising:

[0092] The data acquisition module 301 is used to acquire characteristic data of the target device;

[0093] A generating module 302, configured to generate a hash table of relevant data of the target device according to a hash function and feature data of the target device, wherein the relevant data includes feature data of detection points of non-target devices, and the feature data of the detection points are non-shared data;

[0094] A weight calculation module 303, used to calculate the weight of the feature data of the detection point according to the hash table;

[0095] A model building module 304, used to build a joint model by using the weights of the feature data of the detection points and the device status labels corresponding to the feature data of the detection points;

[0096] The prediction module 305 is used to use the joint model to predict the device state of the target device.

[0097] In one embodiment of the present invention, the weight determination module 303 includes: a hash value determination unit, a first probability determination unit, a second probability determination unit and a weight determination unit; wherein,

[0098] The hash value determination unit is used to determine the hash value of the feature data of the detection point;

[0099] The first probability determination unit is used to determine the distribution probability of the characteristic data of the detection point corresponding to the target device according to the hash table and the hash value of the characteristic data of the detection point;

[0100] The second probability determination unit is used to determine the distribution probability of the characteristic data of the detection point corresponding to the non-target device;

[0101] The weight determination unit is used to determine the ratio of the distribution probability of the feature data of the detection point corresponding to the target device and the distribution probability corresponding to the non-target device as the weight of the feature data of the detection point.

[0102] In one embodiment of the present invention, the first probability determination unit includes: a selection subunit and a probability determination subunit; wherein,

[0103] The selection subunit is used to determine replacement data of the feature data of the detection point from the feature data of the target device according to the hash table and the hash value of the feature data of the detection point;

[0104] The probability determination subunit is used to determine the distribution probability of the replacement data in the characteristic data of the target device as the distribution probability of the characteristic data of the detection point corresponding to the target device.

[0105] In one embodiment of the present invention, the selection subunit is used to perform similarity analysis on the hash value of the hash table and the feature data of the detection point based on a local sensitive hashing algorithm to determine replacement data of the feature data of the detection point from the feature data of the target device.

[0106] In one embodiment of the present invention, the hash value of the feature data of the detection point is determined according to the hash function.

[0107] In one embodiment of the present invention, the model building module 304 includes: a first model determination unit and a second model determination unit;

[0108] The first model determination unit is used to determine the local model of the detection point by using the weight of the feature data of the detection point and the device state label of the feature data of the detection point;

[0109] The second model determination unit is used to establish a joint model according to the local models of each detection point.

[0110] In one embodiment of the present invention, the second model determination unit is used to perform model fusion after model iteration on the local models of each of the detection points based on a joint learning algorithm to establish a joint model.

[0111] Figure 4It is a structural diagram of an electronic device provided by an embodiment of the present invention. At the hardware level, the electronic device includes a processor 401 and a memory 402 storing execution instructions, and optionally also includes an internal bus 403 and a network interface 404. Among them, the memory 402 may include a memory 4021, such as a high-speed random access memory (Random-Access Memory, RAM), and may also include a non-volatile memory 4022 (non-volatile memory), such as at least one disk storage, etc.; the processor 401, the network interface 404 and the memory 402 can be interconnected through an internal bus 403, and the internal bus 403 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc.; the internal bus 403 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus. Of course, the electronic device may also include hardware required for other services. When the processor 401 executes the execution instruction stored in the memory 402, the processor 401 executes the method in any embodiment of the present invention, and is at least used to execute the following steps: Figure 1 or Figure 2 The method shown.

[0112] In a possible implementation, the processor reads the corresponding execution instruction from the non-volatile memory into the memory and then runs it, and can also obtain the corresponding execution instruction from other devices to form a device state information prediction device at the logical level. The processor executes the execution instruction stored in the memory to implement a device state information prediction method provided in any embodiment of the present invention through the executed execution instruction.

[0113] The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or instructions in the form of software. The above processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0114] The embodiment of the present invention further provides a computer-readable storage medium, including an execution instruction. When a processor of an electronic device executes the execution instruction, the processor executes the method provided in any embodiment of the present invention. The electronic device may be specifically as follows: Figure 4 The electronic device shown; the execution instruction is a computer program corresponding to a method for predicting device state information.

[0115] Those skilled in the art will appreciate that the embodiments of the present invention may be provided as methods or computer program products. Therefore, the present invention may be implemented in the form of a complete hardware embodiment, a complete software embodiment, or a combination of software and hardware.

[0116] The multiple embodiments of the present invention are described in a progressive manner, and the same or similar parts between the multiple embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0117] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0118] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for predicting device status information, characterized in that: include: Obtain characteristic data of the target device; Generate a hash table of relevant data of the target device according to the hash function and the feature data of the target device, wherein the relevant data includes feature data of a detection point of a non-target device and feature data of a detection point of a target device, the target device and the non-target device are of the same device type, several features respectively corresponding to the feature data of the target device and the feature data of the non-target device are the same, and the feature data of the detection point is non-shared data; Calculating the weight of the feature data of the detection point according to the hash table; Establishing a joint model by using the weights of the feature data of the detection points and the device status labels corresponding to the feature data of the detection points; Using the joint model to predict the device state of the target device; The step of calculating the weight of the feature data of the detection point according to the hash table includes: Determine a hash value of the characteristic data of the detection point; Determine, according to the hash table and the hash value of the characteristic data of the detection point, the distribution probability of the characteristic data of the detection point corresponding to the target device; Determine the distribution probability of the characteristic data of the detection point corresponding to the non-target device; The ratio of the distribution probability of the feature data of the detection point corresponding to the target device to the distribution probability corresponding to the non-target device is determined as the weight of the feature data of the detection point.

2. The method according to claim 1, characterized in that Determining the distribution probability of the characteristic data of the detection point corresponding to the target device according to the hash table and the hash value of the characteristic data of the detection point includes: Determining replacement data for the feature data of the detection point from the feature data of the target device according to the hash table and the hash value of the feature data of the detection point; The distribution probability of the replacement data in the characteristic data of the target device is determined as the distribution probability of the characteristic data of the detection point corresponding to the target device.

3. The method according to claim 2, characterized in that The determining, according to the hash table and the hash value of the feature data of the detection point, replacement data of the feature data of the detection point from the feature data of the target device comprises: Based on the local sensitive hashing algorithm, a similarity analysis is performed on the hash table and the hash value of the feature data of the detection point to determine replacement data of the feature data of the detection point from the feature data of the target device.

4. The method according to claim 1, characterized in that The hash value of the feature data of the detection point is determined according to the hash function.

5. The method according to claim 1, characterized in that The establishing of a joint model by using the weight of the feature data of the detection point and the device status label corresponding to the feature data of the detection point includes: Determine a local model of the detection point by using a weight of the feature data of the detection point and a device state label of the feature data of the detection point; A joint model is established based on the local models of each detection point.

6. The method according to claim 5, characterized in that The establishing of a joint model according to the local models of each detection point comprises: Based on the joint learning algorithm, the local models of each detection point are iterated and then fused to establish a joint model.

7. A device for predicting device status information, characterized in that: include: A data acquisition module, used to acquire characteristic data of a target device; A generating module, configured to generate a hash table of relevant data of the target device according to a hash function and feature data of the target device, wherein the relevant data includes feature data of a detection point of a non-target device, the target device and the non-target device are of the same device type, several features respectively corresponding to the feature data of the target device and the feature data of the non-target device are the same, and the feature data of the detection point are non-shared data; A weight calculation module, used for calculating the weight of the feature data of the detection point according to the hash table; A model building module, used to build a joint model by using the weights of the feature data of the detection points and the device status labels corresponding to the feature data of the detection points; A prediction module, used for using the joint model to predict the device state of the target device; The weight calculation module is specifically used to: determine the hash value of the feature data of the detection point; determine the distribution probability of the feature data of the detection point corresponding to the target device based on the hash table and the hash value of the feature data of the detection point; determine the distribution probability of the feature data of the detection point corresponding to the non-target device; and determine the ratio of the distribution probability of the feature data of the detection point corresponding to the target device and the distribution probability corresponding to the non-target device as the weight of the feature data of the detection point.

8. A readable medium comprising an execution instruction. When a processor of an electronic device executes the execution instruction, the electronic device executes the method according to any one of claims 1 to 6.

9. An electronic device, comprising a processor and a memory storing execution instructions, wherein when the processor executes the execution instructions stored in the memory, the processor executes the method according to any one of claims 1 to 6.

Images