Device risk prediction method and device, medium and product

Abstract

The embodiment of the invention provides a device risk prediction method and device, a medium and a product. The device risk prediction method comprises the steps of obtaining a group of characteristic values of each to-be-detected device in a to-be-detected device group on a group of characteristics in a prediction time period; obtaining a risk value range of the corresponding characteristic according to the characteristic values of all the to-be-detected devices on the same characteristic, and obtaining a deviation value relative to the corresponding risk value range according to the characteristic values of the to-be-detected devices; obtaining a preliminary risk value according to a comprehensive deviation result formed by the deviation value of each characteristic value of the to-be-detected device; inputting the initial risk value and a group of characteristic values into a target risk prediction model to obtain a target risk prediction value; and sorting the to-be-detected device group according to the target risk prediction value corresponding to the to-be-detected device so as to obtain a risk sorting result. According to the method, the initial risk value and the characteristic value form the composite feature vector input value to the trained target risk prediction model, so that a more accurate target risk prediction value is obtained.

Description

Technical Field

[0001] The present disclosure relates to the fields of artificial intelligence and big data technology, and in particular to device risk prediction methods, devices, media, and products. Background Art

[0002] In modern automotive manufacturing, the transmission, a core component of powertrain, directly determines the vehicle's shifting smoothness, power transmission efficiency, and driving safety. Industry practice shows that even minor quality defects in the transmission can lead to serious consequences such as abnormal wear and tear in the transmission system and power interruption. Therefore, quality control during the production process is extremely important.

[0003] In related technologies, the industry's commonly adopted quality control solution is a random sampling mechanism based on a product audit checklist. This involves randomly selecting samples from qualified products on the final inspection and calibration bench according to a preset cycle for quality audit. While this method can meet basic quality control needs, it has certain technical limitations. For example, with the improvement of modern manufacturing precision, the first-pass assembly pass rate in mainstream automotive factories has generally reached above 99.9%. In this case, the probability of catching potentially defective products through random sampling will decrease exponentially. For example, when the defect rate is 0.1%, there is still a 36.6% probability of missing a defective product when sampling 100 devices continuously. Summary of the Invention

[0004] In view of the above-mentioned shortcomings of the prior art, the purpose of the present disclosure is to provide a device risk prediction method, device, medium and product to solve the problems in the related art.

[0005] A first aspect of the present disclosure provides a device risk prediction method, comprising:

[0006] Obtaining a set of characteristic values ​​for each device to be detected in a group of devices to be detected within a prediction time period, wherein the group of devices to be detected is all devices to be detected within a continuous time window;

[0007] Obtaining a risk value range of the corresponding characteristic according to the characteristic value of each of the devices to be tested on the same characteristic, and obtaining a deviation relative to the corresponding risk value range according to the characteristic value of the devices to be tested;

[0008] Obtaining a preliminary risk value of the device to be detected based on at least a comprehensive deviation result formed by a deviation amount of each characteristic value of the device to be detected;

[0009] Inputting the preliminary risk value of the device to be inspected and a set of characteristic values ​​thereof into a target risk prediction model to obtain a target risk prediction value corresponding to the device to be inspected;

[0010] The group of devices to be detected is sorted according to the target risk prediction values ​​corresponding to the devices to be detected to obtain a risk sorting result.

[0011] In an embodiment of the first aspect, the target risk prediction model is the model with the best performance selected after comparing the prediction performance of the models corresponding to the trained random forest, linear regression, and gradient boosting tree.

[0012] In an embodiment of the first aspect, the prediction performance evaluation method includes mean square error evaluation and determination coefficient evaluation.

[0013] In an embodiment of the first aspect, the risk value range is a range centered on the corresponding characteristic value and bounded by a preset standard deviation multiple; or, the risk value range is an interval with a first preset value as an upper limit and a second preset value as a lower limit.

[0014] In an embodiment of the first aspect, the deviation is defined as a ratio of a difference between an average value of a characteristic of the group of devices to be detected and a characteristic value of the device to be detected corresponding to the characteristic to the average value.

[0015] In an embodiment of the first aspect, obtaining the preliminary risk value of the device to be detected based on at least a comprehensive deviation result formed by a deviation amount of each characteristic value of the device to be detected includes:

[0016] Determining the number of risk characteristics in the set of characteristics of the device to be tested; wherein a risk characteristic is defined as a characteristic whose characteristic value deviates from a risk value range within a preset range;

[0017] An initial risk value of the device to be inspected is obtained according to a proportion of the number of the risk characteristics in a group of the characteristics and a comprehensive result between comprehensive deviation results of a group of the characteristics.

[0018] In an embodiment of the first aspect, after obtaining the risk ranking result, the method further includes:

[0019] The device serial numbers of a preset number of devices to be inspected with the largest target risk values ​​in the risk ranking results are extracted and output.

[0020] A second aspect of the present disclosure provides a computer device, comprising:

[0021] processor and memory;

[0022] The memory stores program instructions;

[0023] The processor is configured to run the program instructions to execute any one of the above-mentioned device risk prediction methods.

[0024] A third aspect of the present disclosure provides a computer-readable storage medium, wherein program instructions are stored, and the program instructions are executed to perform any of the above-mentioned device risk prediction methods.

[0025] A fourth aspect of the present disclosure provides a computer program product, which includes: a method for executing any of the above-mentioned device risk prediction methods.

[0026] The beneficial effects of the present disclosure include: performing risk prediction on all devices to be inspected within a prediction time period, obtaining a target risk prediction value for each device to be inspected using a target risk prediction model, and then performing detailed inspections on devices to be inspected with excessively high risk values ​​based on the resulting risk ranking results. This effectively avoids the inefficiency and low probability of randomly selecting risky devices to be inspected. Furthermore, a composite feature vector is formed by combining the preliminary risk value and the eigenvalue and inputted into the target risk prediction model, making the target risk prediction value output by the model more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flow chart showing a risk prediction method in one embodiment of the present disclosure.

[0028] Figure 2 A schematic diagram showing a specific flow chart of step S3 in the risk prediction method in one embodiment of the present disclosure is shown.

[0029] Figure 3 A flow chart showing a risk prediction method in another embodiment of the present disclosure.

[0030] Figure 4 A schematic diagram showing the structure of a computer device in one embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0031] The following describes the embodiments of the present disclosure through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present disclosure from the information disclosed in this disclosure. The present disclosure can also be implemented or applied through different specific embodiments. The details of the present disclosure can also be modified or changed according to different viewpoints and application modules without departing from the spirit of the present disclosure. It should be noted that the embodiments and features in the embodiments of the present disclosure can be combined with each other unless there is a conflict.

[0032] The following is a detailed description of the embodiments of the present disclosure with reference to the accompanying drawings so that those skilled in the art can easily implement the present disclosure. The present disclosure can be embodied in many different forms and is not limited to the embodiments described herein.

[0033] Throughout the present disclosure, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present disclosure. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or a group of embodiments or examples. Furthermore, those skilled in the art may combine and integrate different embodiments or examples, and features of different embodiments or examples, as described in the present disclosure, without conflicting requirements.

[0034] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the context of this disclosure, "a group" means two or more, unless otherwise specifically defined.

[0035] In order to clearly describe the present disclosure, components not related to the description are omitted, and the same or similar components throughout the specification are denoted by the same reference numerals.

[0036] Throughout this specification, when a device is said to be "connected" to another device, this includes not only "direct connection" but also "indirect connection" with other elements interposed therebetween. Furthermore, when a device is said to "include" a certain component, unless otherwise stated, this does not exclude the inclusion of other components but rather implies that the device may include other components.

[0037] Although the terms first, second, etc. are used in this document to represent various elements in some examples, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, the first interface and the second interface, etc. are represented. Furthermore, as used in this document, the singular forms "one", "an", and "the" are intended to also include the plural forms, unless there is a contrary indication in the context. It should be further understood that the terms "comprise" and "include" indicate the presence of features, steps, operations, elements, modules, projects, types, and / or groups, but do not exclude the presence, occurrence, or addition of one or a group of other features, steps, operations, elements, modules, projects, types, and / or groups. The terms "or" and "and / or" used herein are interpreted as inclusive, or mean any one or any combination. Therefore, "A, B, or C" or "A, B, and / or C" means "any of the following: A; B; C; A and B; A and C; B and C; A, B, and C". Exceptions to this definition only occur when the combination of elements, functions, steps, or operations is inherently mutually exclusive in some way.

[0038] The technical terms used herein are intended only to refer to specific embodiments and are not intended to limit the present disclosure. The singular form used herein also includes the plural form unless the statement explicitly indicates otherwise. The term "comprising" as used in this specification is intended to specify specific features, regions, integers, steps, operations, elements, and / or components and does not exclude the presence or addition of other features, regions, integers, steps, operations, elements, and / or components.

[0039] Although not defined differently, all terms used herein, including technical and scientific terms, have the same meanings as those generally understood by those skilled in the art to which this disclosure belongs. Terms defined in commonly used dictionaries are additionally interpreted as having meanings consistent with relevant technical literature and the current message. Unless otherwise defined, they should not be overly interpreted as ideal or highly formalized meanings.

[0040] In the related art, the qualified products on the final inspection and calibration table of the production line need to be further inspected and audited, and the practice in the related art is to randomly select samples from the qualified products for audit at a prescribed frequency. However, when the internal qualification rate of the factory is high, the probability of selecting a product with potential risks during the product audit is extremely low, which is similar to the low probability situation in a random event. Therefore, there is a high possibility that risky products will flow into the market. Therefore, in one embodiment of the present disclosure, in order to solve the above problem, as shown in FIG. Figure 1 The embodiment shows a device risk prediction method, in which characteristic values ​​of all devices to be tested within a prediction time period are obtained, a preliminary risk value is first obtained based on the obtained characteristic values, and the preliminary risk value and the characteristic value form a composite feature vector input value into a trained target risk prediction model to obtain a more accurate target risk prediction value, thereby obtaining a ranked risk ranking result based on the target risk prediction values ​​of all devices to be tested, thereby intuitively displaying the level of product risk among the qualified products on the final inspection and calibration station of the production line for further processing.

[0041] exist Figure 1 In an embodiment, the device risk prediction method specifically includes:

[0042] Step S1: obtaining a set of characteristic values ​​of each device to be detected in the group of devices to be detected within a prediction time period on a set of characteristics.

[0043] The device group to be detected is all devices to be detected within a continuous time window.

[0044] Specifically, in some embodiments, the final inspection and calibration station further audits the devices within the qualified line. The characteristic values ​​are the content of the final inspection and calibration station's testing of the products within the qualified line, and therefore are the content of the product inspection without damaging the product itself. The characteristics of the device to be tested may include, but are not limited to, leakage testing, drag torque measurement, bearing end noise testing, and noise detection. In some embodiments, after completing the characteristic test of each device to be tested, the final inspection and calibration station generates corresponding characteristic value data. This data can be transmitted to a central database for storage and management. Therefore, in some embodiments, the characteristic values ​​of the devices to be tested within the group of devices to be tested can be directly obtained from the central database.

[0045] The preset time period can be used to limit the time range for data collection on the device. This time period can be a preset time period, such as all devices to be inspected on the final inspection and calibration station between 5:00 AM and 8:00 AM. In some preferred embodiments, the accuracy of risk prediction results can be improved by collecting characteristic values ​​of the devices to be inspected within the preset time period at fixed points. For example, if there are a large number of high-risk devices within a certain time period, the devices to be inspected within this time period can be regularly collected for risk prediction, thereby reducing the number of risky devices in the final output.

[0046] Step S2: obtaining a risk value range of a corresponding characteristic according to the characteristic value of each device to be detected on the same characteristic, and obtaining a deviation relative to the corresponding risk value range according to the characteristic value of the device to be detected.

[0047] In quality control, products that pass initial inspection are not necessarily completely risk-free. While some devices may meet basic acceptance criteria, the values ​​of certain characteristics may approach critical points (e.g., near the boundaries of the risk range). Therefore, it is necessary to calculate the deviation of the characteristics of the device under inspection from the risk range to quantify the risk value of the corresponding characteristic.

[0048] In Example 1, for a set of characteristics possessed by the devices to be tested, the average of the characteristic values ​​for the corresponding characteristic is calculated based on the set of characteristic values ​​for each device in the group. Then, with the average value of the corresponding characteristic as the center, boundaries are drawn above and below the average value using a preset standard deviation multiple. Within the boundaries is the risk value range; those within the risk value range are considered non-risk, while those outside the preset risk value range are considered risky. The preset standard deviation multiple can be adjusted based on actual conditions. In some embodiments, the preset range can be set based on the amount of excess, such as exceeding the preset standard deviation multiple by a preset number of standard deviations. Alternatively, the preset range can be determined based on fixed upper and lower acceptable limits (respectively outside the upper and lower limits of the risk value range), and exceeding the preset range can mean exceeding the range between the upper and lower acceptable limits. Characteristics outside the preset range are not considered in the risk prediction of the present embodiment. Because risk prediction in the present embodiment is based on the characteristics of products that have passed the initial inspection, products outside the preset range are directly considered to have failed the initial inspection.

[0049] In Example 2, risk value ranges can also be divided based on the risk profile of each characteristic in historical data. For example, if the corresponding characteristic A values ​​are all above a certain value when the device risk value is high, then this value is set as the upper limit of the risk range. The lower limit of the corresponding risk value range for characteristic A can also be determined using the above method, that is, if the corresponding characteristic A values ​​are all below a certain value when the device risk value is high. The method of obtaining risk value ranges based on historical data in Example 2 is universal and eliminates the need to re-define risk value ranges for each batch of devices to be tested when performing risk prediction.

[0050] In Example 3, a first preset value can be pre-set as the upper limit and a second preset value as the lower limit according to industry standards to obtain a risk value range. The advantage of setting the risk value range according to industry standards is its high adaptability, which can be applied to most devices to be tested. In contrast, the method for obtaining the risk value range in Example 1 is more targeted, and a corresponding risk value range can be generated for each group of devices to be tested, thereby making the resulting preliminary risk value more accurate. In actual use, the corresponding risk value range delineation method can be selected based on the actual situation of the device to be tested.

[0051] Optionally, the deviation of a device under test is defined as the ratio of the difference between the average value of a characteristic of the device under test and the characteristic value of the characteristic of the device under test, divided by the average value. The deviation quantifies the degree of difference between the characteristic value of a characteristic of the device under test and the average level of the entire device group.

[0052] The above content can be expressed as follows using formula 1: Deviation = ;

[0053] Among them, meanLine is the average value of the same characteristic of multiple devices under test in the device group under test. If there is a group of characteristics, a corresponding number of groups of average values ​​can be obtained; testbenchValue is the characteristic value of the characteristic of the device under test.

[0054] Among them, if the characteristic value of a characteristic of a device to be tested deviates significantly from the average value corresponding to the characteristic (whether it is too high or too low), it may indicate that the device to be tested has an abnormality or potential risk, which is manifested as the larger the absolute value of the deviation, the greater the degree of deviation from the average value.

[0055] Step S3: Obtaining a preliminary risk value of the device to be detected based on at least a comprehensive deviation result formed by the deviation amount of each characteristic value of the device to be detected.

[0056] In some embodiments, a preliminary risk value for the device under inspection can be directly derived from the sum of the deviations of a set of characteristics, without involving complex weight assignment or ratio adjustment. This approach is suitable for preliminary screening of devices that may pose risks. This approach is particularly useful in scenarios where high accuracy of the risk value is not required, and a rough risk grading is all that is needed.

[0057] In some embodiments, the risk value can be further refined based on the deviation amount combined with the risk value range to obtain a more accurate initial risk value, such as Figure 2 As shown in the embodiment, step S3 specifically includes:

[0058] Step S31: Determine the number of risk characteristics in the set of characteristics of the device to be detected; wherein a risk characteristic is defined as a characteristic value having a deviation within a preset range relative to a risk value range.

[0059] Step S32: obtaining an initial risk value of the device to be inspected according to a proportion of the number of the risk characteristics in a group of the characteristics and a comprehensive result between comprehensive deviation results of a group of the characteristics.

[0060] Specifically, if the characteristic value of a characteristic falls within the risk range, the characteristic is not considered a risk characteristic. If the characteristic value of a characteristic exceeds the risk range but remains within a preset range (for example, within a certain ratio of the upper and lower limits of the risk range), the characteristic is considered a risk characteristic. If the characteristic value of a characteristic completely exceeds the preset range, the device is considered to have failed the initial inspection and is not within the scope of testing for this application.

[0061] The above steps S31 and S32 can also be expressed as the following formula 2:

[0062] Wherein, Riskiness is the initial risk value, n is a weight in [0, 1] used to adjust the impact of the risky characteristic ratio and the comprehensive deviation result, nr. of risky characteristic is the number of risky characteristics in the device to be tested, total nr. of characteristic is the total number of characteristics in a group of devices to be tested, meanLine is the average value of any characteristic in the group of devices to be tested, and testbenchValue is the characteristic value of any characteristic in the device to be tested.

[0063] By combining the risk characteristic proportion and the comprehensive deviation results, the overall risk status of the device to be tested can be more accurately reflected.

[0064] Step S4: Inputting the preliminary risk value of the device to be inspected and the set of characteristic values ​​thereof into a target risk prediction model to obtain a target risk prediction value corresponding to the device to be inspected.

[0065] Specifically, in some embodiments, the preliminary risk value is a preliminary quantitative indicator of the overall risk level of the device to be tested. It integrates information on the deviation of a set of characteristic values ​​(and may also include the number of risk characteristics; a greater number indicates a greater risk), thereby reflecting the apparent risk status of the device to be tested. The target risk prediction model can extract more risk information based on the input preliminary risk value and the original characteristic values, including deeper potential risks, thereby improving prediction accuracy. Suppose a device has a high preliminary risk value (e.g., 0.8), but some characteristic values ​​are close to the normal range. In this case, the preliminary risk value can help the model identify potential overall risks and avoid underestimating risks due to the "normal" performance of individual characteristic values.

[0066] Optionally, by comparing multiple models, the model that best suits the current data characteristics can be found, thereby improving prediction accuracy. The target risk prediction model described above is the best performing model selected after comparing the prediction performance of the trained random forest, linear regression, and gradient boosting tree models. By comparing multiple models, a model that performs stably under different data conditions can be selected to enhance the robustness of the system. The training method for the target risk prediction model corresponds to the application method, which will not be elaborated on here.

[0067] Optionally, the prediction performance evaluation method includes mean square error evaluation and determination coefficient evaluation. Mean square error is a commonly used indicator to measure the difference between the predicted value and the true value. The smaller the mean square error, the smaller the difference between the predicted value and the true value, that is, the higher the prediction accuracy of the model. The determination coefficient, also known as the coefficient of determination, is used to measure the ability of the model to explain the variation of the data. The range of the determination coefficient is from negative infinity to 1, but usually its value is between 0 and 1. When the determination coefficient is close to 1, it means that the model can explain the variation of the data well; when the determination coefficient is close to 0, it means that the explanatory ability of the model is poor; if the determination coefficient is negative, it means that the performance of the model is worse than simply using the average value of the data.

[0068] In step S5, the group of devices to be detected may be sorted according to the target risk prediction values ​​corresponding to the devices to be detected to obtain a risk sorting result.

[0069] Specifically, in some embodiments, the devices to be tested can be sorted in descending order according to the size of the target risk prediction value. The advantage of this is that the devices to be tested with larger risk values ​​can be placed in front. In some embodiments, further testing of the devices to be tested that are ranked in front by a preset number of places is required to determine whether they can enter the market. In some embodiments, the results after sorting according to the target risk prediction value can also be graded, such as divided into three levels of high risk, medium risk and low risk, and a risk ranking result is formed. Among them, different operations can be performed on the devices to be tested of different levels. For example, the characteristics of the high-risk devices to be tested can be further tested for the characteristics with larger risk values; the medium-risk devices to be tested can be marked and subsequently tracked and investigated; the low-risk devices to be tested can be directly released into the market.

[0070] exist Figure 3 In this embodiment, after obtaining the risk ranking results in step S5, the process further includes extracting and outputting the device serial numbers of a preset number of devices to be inspected with the highest target risk values ​​from the risk ranking results. The preset number represents the number of devices to be inspected with the highest risk values ​​to be extracted. The output information for each device to be inspected includes: a device serial number (unique identifier) ​​and a predicted target risk value.

[0071] like Figure 4 FIG. 1 is a schematic diagram showing the structure of a computer device in one embodiment of the present disclosure.

[0072] The computer device 100 may be exemplified as a processing terminal in a cloud platform, such as a server, a desktop computer, a laptop computer, a tablet computer, a smart phone, or other terminals.

[0073] The computer device 100 includes a bus 101, a processor 102, and a memory 103. The processor 102 and the memory 103 can communicate with each other via the bus 101. The memory 103 can store program instructions. The processor 102 executes the program instructions in the memory 103 to implement the steps of the device risk prediction method in the previous embodiment, such as Figure 1 .

[0074] Bus 101 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. Buses can be classified into address buses, data buses, and control buses. For ease of illustration, although only one thick line is used in the figure, this does not mean that there is only one bus or only one type of bus.

[0075] In some embodiments, processor 102 may be implemented as a central processing unit (CPU), a microprocessor unit (MCU), a system on a chip (SoC), or a field programmable gate array (FPGA). Memory 103 may include volatile memory, such as random access memory (RAM), for temporarily storing data while running programs.

[0076] The memory 103 may also include a non-volatile memory (non-volatile memory) for data storage, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid-state disk (SSD).

[0077] In some embodiments, the computer device 100 may further include a communicator 104. The communicator 104 is used to communicate with the outside world. In a specific example, the communicator 104 may include one or a group of wired and / or wireless communication circuit modules. For example, the communicator 104 may include one or more of a wired network card, a USB module, a serial interface module, etc. The wireless communication protocols followed by the wireless communication module include, for example, near field communication (NFC) technology, infrared (IR) technology, Global System for Mobile communications (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Time-Division Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), Bluetooth (BT), Global Navigation Satellite System (GNSS), etc. One or more of the following.

[0078] An embodiment of the present disclosure may further provide a computer-readable storage medium storing program instructions, which implement the device risk prediction method in any of the previous embodiments when the program instructions are executed.

[0079] That is, the method steps in the above embodiments are implemented as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or are implemented as computer code originally stored in a remote recording medium or a non-transitory machine-readable medium downloaded via a network and to be stored in a local recording medium, so that the method represented herein can be stored in such software processing on a recording medium using a general-purpose computer, a dedicated processor or programmable or dedicated hardware (such as an ASIC or FPGA).

[0080] A computer program product may also be provided in an embodiment of the present disclosure, including program instructions for executing the device risk prediction method described in any of the above embodiments.

[0081] The above embodiments are merely illustrative of the principles and effects of this disclosure and are not intended to limit this disclosure. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of this disclosure. Therefore, any equivalent modifications or alterations made by a person skilled in the art without departing from the spirit and technical concepts disclosed herein shall be encompassed by the scope of protection of this disclosure.

Claims

1. A device risk prediction method, characterized in that: include: Obtaining a set of characteristic values ​​for each device to be inspected in the group of devices to be inspected within a prediction time period on a set of characteristics; Obtaining a risk value range of the corresponding characteristic according to the characteristic value of each of the devices to be tested on the same characteristic, and obtaining a deviation relative to the corresponding risk value range according to the characteristic value of the devices to be tested; Obtaining a preliminary risk value of the device to be detected based on at least a comprehensive deviation result formed by a deviation amount of each characteristic value of the device to be detected; Inputting the preliminary risk value of the device to be inspected and a set of characteristic values ​​thereof into a target risk prediction model to obtain a target risk prediction value corresponding to the device to be inspected; The group of devices to be detected is sorted according to the target risk prediction values ​​corresponding to the devices to be detected to obtain a risk sorting result.

2. The device risk prediction method according to claim 1, characterized in that: The target risk prediction model is the model with the best performance selected after comparing the prediction performance of the models corresponding to the trained random forest, linear regression, and gradient boosting tree.

3. The device risk prediction method according to claim 2, characterized in that: The prediction performance evaluation method includes mean square error evaluation and determination coefficient evaluation.

4. The device risk prediction method according to claim 1, characterized in that: The risk value range is a range centered on the corresponding characteristic value and bounded by a preset standard deviation multiple; or, the risk value range is an interval with a first preset value as an upper limit and a second preset value as a lower limit.

5. The device risk prediction method according to claim 1, characterized in that: The deviation is defined as the ratio of the difference between the average value of a characteristic of the device to be detected group and the characteristic value of the device to be detected corresponding to the characteristic to the average value.

6. The device risk prediction method according to claim 1, characterized in that: Obtaining the preliminary risk value of the device to be detected based on at least a comprehensive deviation result formed by the deviation amount of each characteristic value of the device to be detected includes: Determining the number of risk characteristics in the set of characteristics of the device to be tested; wherein the risk characteristic is defined as a characteristic whose characteristic value deviates from the risk value range within a preset range; An initial risk value of the device to be inspected is obtained according to a proportion of the number of the risk characteristics in a group of the characteristics and a comprehensive result between comprehensive deviation results of a group of the characteristics.

7. The device risk prediction method according to claim 1, characterized in that: After obtaining the risk ranking results, it also includes: The device serial numbers of a preset number of devices to be inspected with the largest target risk values ​​in the risk ranking results are extracted and output.

8. A computer device, characterized in that: include: processor and memory; The memory stores program instructions; The processor is configured to run the program instructions to execute the device risk prediction method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that Program instructions are stored, and the program instructions are executed to perform the device risk prediction method according to any one of claims 1 to 7.

10. A computer program product, characterized in that include: Used to execute the device risk prediction method according to any one of claims 1 to 7.

Images