Fault prediction method and device of automobile, electronic equipment and storage medium

Abstract

The embodiment of the invention provides an automobile fault prediction method, and the method comprises the steps: collecting the maintenance records of a plurality of automobiles, extracting a fault tag, carrying out the vectorization processing, and generating a fault preference row vector of each automobile. And for the to-be-predicted automobile, extracting the target fault tag of the to-be-predicted automobile, comparing the target fault tag with a preset threshold, and screening out a first type of automobiles with the number of the target fault tags exceeding the threshold. Performing vectorization processing on the target fault labels of the automobiles to generate a target fault row vector, and calculating the similarity between the target fault row vector and the fault preference row vector; and according to a similarity calculation result, screening out a first type of fault preference vehicles whose similarity with the first type of vehicles satisfies a preset value, and screening out a fault tag satisfying the preset value from the vehicles as a first type of fault prediction result to be recommended to the first type of vehicles.

Description

Technical Field

[0001] The present invention relates to the field of automobile technology, and in particular to a method for predicting automobile faults, a device for predicting automobile faults, an electronic device, and a computer-readable storage medium. Background Art

[0002] Cars have become an indispensable tool for people to travel. At the same time, with the advancement of technology and the improvement of living standards, users' expectations for cars are no longer limited to its speed and convenience, but pay more attention to its safety and riding comfort during driving. However, cars will inevitably encounter various faults during long-term use. These faults are like hidden dangers lurking in the dark. Once they break out, they may bring unpredictable risks to drivers and passengers. Therefore, it is particularly important to predict possible faults of cars. Through regular inspection and maintenance, as well as the use of advanced diagnostic technology, the faults can be nipped in the bud before they appear.

[0003] For automobile failure prediction, the most common method is annual vehicle inspection. Since a car is a complex mechanical system, its operating status is affected by many factors, including but not limited to driving habits, maintenance, and aging of parts. Therefore, although the annual inspection system improves vehicle safety to a certain extent, it cannot completely prevent all possible failures. Summary of the invention

[0004] The purpose of the present invention is to provide a method for predicting automobile failures, by analyzing a large number of automobile maintenance history records to identify similarities between automobiles related to specific failure modes. The present invention takes into account patterns and trends in time series, and the analysis based on historical data can improve the accuracy of future failure predictions.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] Collecting maintenance records of several vehicles, and extracting a fault label corresponding to each vehicle from the maintenance records;

[0007] Vectorizing the fault labels to generate a fault preference row vector corresponding to each of the vehicles;

[0008] Obtaining a target maintenance record of the vehicle to be predicted, and extracting a target fault label of the vehicle to be predicted from the target maintenance record;

[0009] Comparing the number of target fault labels with a preset label threshold to obtain a first type of automobile whose number of target fault labels is greater than the preset threshold;

[0010] Vectorizing the target fault labels of the first type of vehicles to generate target fault row vectors corresponding to the first type of vehicles;

[0011] Calculating the similarity between the target fault row vector and the fault preference row vector to obtain a first-type similarity calculation result, and selecting a first-type fault preference vehicle whose similarity with the first-type vehicle meets a preset value from a plurality of the vehicles according to the first-type similarity calculation result;

[0012] A fault label that meets a preset value is selected from the fault preference row vector corresponding to the first type of fault preference vehicle as a first type of fault prediction result, and the first type of fault prediction result is recommended to the first type of vehicle.

[0013] Optionally, the method further comprises:

[0014] Collecting file information of a number of the cars, and extracting a file information tag corresponding to each of the cars from the file information;

[0015] The file information labels are vectorized to generate a file information feature vector corresponding to each of the cars.

[0016] Optionally, the vectorizing the fault label includes:

[0017] Creating a row vector for each of the cars, wherein the row vector includes a plurality of element values, and the element values ​​are used to represent the occurrence frequency of the fault label;

[0018] According to a preset fault label sequence, the element value corresponding to each fault label is filled into the row vector to obtain a fault preference row vector corresponding to each of the vehicles.

[0019] Optionally, the selecting a fault label satisfying a preset value from the fault preference row vector corresponding to the first type of fault preference vehicle as the first type of fault prediction result includes:

[0020] The element value in the fault preference row vector corresponding to the first type of fault preference vehicle is compared with a preset element reference value, and the fault label with the element value greater than or equal to the element reference value is taken as the first type of fault prediction result.

[0021] Optionally, comparing the number of the target fault labels with a preset label threshold includes:

[0022] If the number of the target fault labels is greater than the preset label threshold, the vehicle to be predicted is regarded as a first-category vehicle;

[0023] If the number of the target fault labels is less than or equal to the preset label threshold, the vehicle to be predicted is regarded as a second type of vehicle.

[0024] Optionally, when the to-be-predicted automobile is the second type of automobile, the method further comprises:

[0025] Acquire target file information of the second type of automobile, and extract target file information tags of the second type of automobile from the target file information;

[0026] Performing vectorization processing on the target file information label to generate a target file information feature vector corresponding to the second type of automobile;

[0027] Calculating a feature difference value of each of the vehicles based on the feature vector of the archive information;

[0028] Calculating a target feature difference of the second type of automobile based on the target profile information feature vector;

[0029] Calculating the similarity between the feature difference value and the target feature difference value to obtain a second type of similarity calculation result;

[0030] Sorting the second-category similarity calculation results in descending order, so as to select the second-category fault preference car whose similarity with the second-category car meets a preset value from the plurality of cars;

[0031] A fault label that meets a preset value is selected from the fault preference row vector corresponding to the second-category fault preference vehicle as a second-category fault prediction result, and the second-category fault prediction result is recommended to the second-category vehicle.

[0032] Optionally, the calculating the similarity between the target fault row vector and the fault preference row vector to obtain a first type of similarity calculation result includes:

[0033] The cosine similarity between the target fault row vector and the fault preference row vector is calculated, wherein the cosine similarity is used to represent the degree of fault preference between the first type of vehicle and a plurality of the first type of vehicles, wherein the cosine similarity is proportional to the degree of fault preference.

[0034] The present invention discloses a vehicle fault prediction device, the device comprising:

[0035] A fault label extraction module is used to collect maintenance records of several vehicles and extract a fault label corresponding to each vehicle from the maintenance records;

[0036] A fault label vectorization module, used for performing vectorization processing on the fault label to generate a fault preference row vector corresponding to each of the vehicles;

[0037] A target fault label extraction module is used to obtain a target maintenance record of the vehicle to be predicted, and extract a target fault label of the vehicle to be predicted from the target maintenance record;

[0038] A first comparison module, used for comparing the number of the target fault labels with a preset label threshold, so as to obtain a first type of automobile whose number of the target fault labels is greater than the preset threshold;

[0039] a target fault label vectorization module, used for performing vectorization processing on the target fault label of the first type of vehicle to generate a target fault row vector corresponding to the first type of vehicle;

[0040] A first-type similarity calculation module, configured to calculate the similarity between the target fault row vector and the fault preference row vector to obtain a first-type similarity calculation result, and to select a first-type fault preference vehicle whose similarity with the first-type vehicle meets a preset value from a plurality of the vehicles according to the first-type similarity calculation result;

[0041] The first type of fault prediction result generating module is used to select the fault labels meeting the preset values ​​from the fault preference row vectors corresponding to the first type of fault preference vehicles as the first type of fault prediction results, and recommend the first type of fault prediction results to the first type of vehicles.

[0042] An embodiment of the present invention further discloses an electronic device, comprising at least one processor and a memory communicatively connected to the at least one processor;

[0043] The memory is used to store computer programs;

[0044] The processor is used to implement the method described in the embodiment of the present invention when executing the program stored in the memory.

[0045] The embodiment of the present invention further discloses a computer-readable storage medium on which a computer program is stored. When executed by one or more processors, the processors are enabled to execute the method described in the embodiment of the present invention.

[0046] Beneficial effects of the present invention:

[0047] The embodiment of the present invention collects maintenance records of several vehicles, extracts the fault label corresponding to each vehicle from the maintenance records, and vectorizes the fault label to generate a fault preference row vector corresponding to each vehicle; obtains a target maintenance record of the vehicle to be predicted, extracts the target fault label of the vehicle to be predicted from the target maintenance record, compares the number of target fault labels with a preset label threshold, and obtains a first category of vehicles whose number of target fault labels is greater than the preset threshold; vectorizes the target fault label of the first category of vehicles to generate a target fault row vector corresponding to the first category of vehicles; calculates the similarity between the target fault row vector and the fault preference row vector to obtain a first category similarity calculation result, and selects a first category of fault preference vehicles whose similarity with the first category of vehicles meets a preset value from several vehicles according to the first category similarity calculation result; selects a fault label that meets a preset value from the fault preference row vector corresponding to the first category of fault preference vehicles as a first category of fault prediction result, and recommends the first category of fault prediction result to the first category of vehicles. The present invention combines the fault prediction technology of vehicle maintenance records and collaborative filtering algorithms, which can improve the accuracy of fault prediction and the personalized service level on the one hand, and optimize resource allocation, enhance user experience, and promote data-based decision making on the other hand. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flowchart of a method for predicting automobile faults provided in an embodiment of the present invention;

[0049] Figure 2 A schematic diagram of maintenance record collection of a method for predicting automobile faults provided in an embodiment of the present invention;

[0050] Figure 3 A schematic diagram of archive information collection of a method for predicting automobile faults provided in an embodiment of the present invention;

[0051] Figure 4 A schematic diagram of the structure of a vehicle fault prediction device provided in an embodiment of the present invention;

[0052] Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0054] Reference Figure 1 , shows a flowchart of a method for predicting a vehicle fault provided in an embodiment of the present invention, which specifically includes the following steps:

[0055] Step S11, collecting maintenance records of several vehicles, and extracting the fault label corresponding to each vehicle from the maintenance records.

[0056] In some feasible embodiments of the present invention, the maintenance records of the automobile are collected from the automobile maintenance service center, the manufacturer's database or the third-party data provider. These records usually contain detailed information such as vehicle identification information, maintenance date, maintenance items, replaced parts, fault description, etc.

[0057] Analyze each maintenance record, extract the fault-related information from each maintenance record, and form a fault label, such as Figure 2 As shown. These fault labels may include specific fault codes, fault types, affected systems or components, etc. For example, if a car is repaired because of engine overheating, the corresponding fault label may be "engine overheating" or "cooling system failure". Furthermore, the extracted fault labels can be classified for subsequent analysis and processing. The classification can be based on the faulty system (such as electrical system, engine system, chassis system, etc.) or the severity of the fault (such as minor, moderate, severe).

[0058] The embodiment of the present invention can also sort out the extracted fault labels to ensure the accuracy and consistency of the labels. The sorting process includes removing duplicate labels, correcting erroneous labels, and unifying label naming.

[0059] Through the above steps, a detailed fault label library containing several cars can be established. These labels will be used for subsequent fault prediction and analysis to help identify the types of faults that may occur in the car, so as to carry out prevention and maintenance in advance.

[0060] Step S12, vectorizing the fault labels to generate a fault preference row vector corresponding to each vehicle.

[0061] In one embodiment, each fault label needs to be converted into a numerical value or vector representation.

[0062] For each car, the vector representations of all its fault labels are combined to form a row vector. This row vector represents the fault preference shown by the car in its historical maintenance records. For example, if a car has three labels: "engine failure", "brake system failure" and "electrical system failure", then its fault preference row vector will contain the vector representations of these three labels. In order to ensure that the fault preference row vectors between different cars are comparable, these vectors can be standardized, including normalization, scaling or using other statistical methods to eliminate differences in dimensions and numerical ranges.

[0063] In some cases, different fault labels may have different importances. Therefore, each element in the fault preference row vector can be weighted according to the frequency, severity, or impact of the fault label on vehicle performance.

[0064] In a feasible specific embodiment of the present invention, a row vector is created for each car, and the row vector contains a number of element values, and the element values ​​are used to represent the frequency of occurrence of the fault label. According to the pre-set fault label order, the element value corresponding to each fault label is filled into the row vector to obtain the fault preference row vector corresponding to each car. If a certain fault label appears in the fault record of the car, the corresponding element value is increased.

[0065] For example, there are the following fault label sets: engine fault, brake system fault, electrical system fault, transmission system fault, suspension system fault, and the preset fault label order is [engine fault, brake system fault, electrical system fault, transmission system fault, suspension system fault].

[0066] If the fault record of car A includes 3 engine faults and 1 brake system fault, then the fault record of car A is:

[0067] Engine failure: 3 times

[0068] Braking system failure: 1 time

[0069] Therefore, according to the preset fault label order of [engine fault, brake system fault, electrical system fault, transmission system fault, suspension system fault], the fault preference row vector of car A can be expressed as:

[0070] {Fault preference row vector of car A} = [3, 1, 0, 0, 0]

[0071] This vector indicates that car A has a higher preference for engine failure and brake system failure, but no record for other failure types.

[0072] Through the above process, the fault label of each car is converted into a numerical fault preference row vector, which provides a basis for subsequent similarity calculation, fault prediction and recommendation.

[0073] Step S13, obtaining a target maintenance record of the vehicle to be predicted, and extracting a target fault label of the vehicle to be predicted from the target maintenance record.

[0074] First, the latest maintenance records of the car to be predicted need to be obtained. These records can be obtained from the car repair service center, the manufacturer's database, or the maintenance history provided by the car owner. The records should include detailed information such as the vehicle identification information, maintenance date, maintenance items, replaced parts, fault description, etc. of the car to be predicted.

[0075] In one embodiment, the target maintenance records are analyzed in detail to extract key information related to the fault and form target fault labels. These labels may include specific fault codes, fault types, affected systems or components, etc. For example, if the record shows that the predicted car was repaired due to brake failure, the extracted target fault label may be "brake system failure". After extracting the target fault labels, these labels can be screened and processed to ensure that they are consistent with the fault labels extracted from several cars during the data collection phase, including removing irrelevant labels, correcting erroneous labels, and unifying label naming.

[0076] Step S14, comparing the number of target fault labels with a preset label threshold to obtain a first category of vehicles whose number of target fault labels is greater than the preset threshold.

[0077] In a feasible specific embodiment of the present invention, if the number of target fault labels is greater than a preset label threshold, the car to be predicted is regarded as a first-category car; if the number of target fault labels is less than or equal to the preset label threshold, the car to be predicted is regarded as a second-category car.

[0078] First, a preset label threshold is set in advance. The preset label threshold is used to evaluate whether the car to be predicted belongs to the first category or the second category. The number of target fault labels of the car to be predicted is counted, and the number of target fault labels of the car to be predicted is compared with the preset label threshold. The number of target fault labels obtained by counting is compared with the preset label threshold. If the number of target fault labels is greater than the preset label threshold, the car to be predicted is classified as the first category of cars, which are generally considered to have a higher risk of failure or a complex failure mode. If the number of target fault labels is less than or equal to the preset label threshold, the car to be predicted is classified as the second category of cars, which are generally considered to have a lower risk of failure or a simpler failure mode. According to the classification results, different fault prediction strategies can be adopted for the car to be predicted. For example, the label threshold is set in advance to 2, that is, the car with the number of target fault labels greater than 2 is the first category of cars. If there are 2 engine failures and 2 brake system failures in the fault record of the car Q to be predicted, then the number of target fault labels of the car Q to be predicted is 4. At this time, the number of target fault labels of the car Q to be predicted is greater than the preset label threshold, and it can be determined that the car Q to be predicted is the first category of cars.

[0079] Through the above process, fault prediction resources can be managed and optimized more effectively, ensuring that high-risk vehicles are monitored and predicted in a focused manner, thereby improving the efficiency and accuracy of the entire fault prediction system.

[0080] Step S15, vectorizing the target fault labels of the first category of vehicles to generate target fault row vectors corresponding to the first category of vehicles.

[0081] In one embodiment, step S15 includes:

[0082] Convert each target fault label of the first class of cars into a numeric or vector representation, and then combine the vector representations of all target fault labels to form a row vector.

[0083] In a feasible specific embodiment of the present invention, a row vector of the first type of automobile is created, and the row vector contains a plurality of element values, and the element values ​​are used to represent the frequency of occurrence of the target fault label obtained by the first type of automobile. According to the pre-set fault label sequence, the element value corresponding to each target fault label is filled into the row vector to obtain the target fault preference row vector corresponding to the first type of automobile. If a certain target fault label appears in the fault record of the first type of automobile, the corresponding element value is increased.

[0084] For example, if there are 2 engine failures and 2 brake system failures in the failure record of the first type of car Q, then the failure record of the first type of car Q is:

[0085] Engine failure: 2 times

[0086] Braking system failure: 2 times

[0087] Therefore, according to the preset fault label order of [engine fault, brake system fault, electrical system fault, transmission system fault, suspension system fault], the target fault preference row vector of the first type of car Q can be expressed as:

[0088] {target fault preference row vector of car Q} = [2, 2, 0, 0, 0]

[0089] Step S16, calculating the similarity between the target fault row vector and the fault preference row vector to obtain a first-category similarity calculation result, and selecting a first-category fault preference vehicle whose similarity with the first-category vehicle meets a preset value from a plurality of vehicles according to the first-category similarity calculation result.

[0090] In a feasible embodiment of the present invention, the cosine similarity between the target fault row vector and the fault preference row vector can be calculated, and the cosine similarity is used to represent the degree of fault preference between the first type of car and several of the cars, wherein the cosine similarity is proportional to the degree of fault preference.

[0091] Taking the above car A and the first type of car Q as an example, the cosine similarity algorithm is used to calculate the similarity between the two. The specific calculation formula is as follows:

[0092]

[0093] Among them, sim represents the first-class similarity calculation result between car A and the first-class car Q, n represents the number of fault labels of car A, i is a variable, representing the index of summation, i=1 represents the starting point of summation, that is, the first element value in the row vector, a i represents the element value corresponding to the fault label of car A, q i Represents the element value corresponding to the target fault label of the first category of car Q.

[0094] In other feasible embodiments, when measuring the fault preference similarity between vehicles, the present invention may also use other similarity calculation formulas, such as the Jaccard similarity coefficient, the Pearson correlation coefficient, etc., and the present invention is not limited to this.

[0095] In the embodiment of the present invention, a similarity value needs to be preset as an evaluation standard. After calculating the first type of similarity calculation result, the first type of similarity calculation result is compared with the preset similarity value, wherein the automobile corresponding to the first type of similarity calculation result greater than or equal to the similarity value is a first type of fault preference vehicle similar to the first type of automobile. The first type of fault preference vehicle refers to a vehicle with a high probability of having the same fault as the automobile to be predicted.

[0096] For example, the preset similarity value is 0.85, where

[0097] The fault preference row vector of car A is: {fault preference row vector of car A} = [3, 1, 0, 0, 0]

[0098] The fault preference row vector of car B is: {the fault preference row vector of car A} = [3, 2, 0, 0, 0]

[0099] The fault preference row vector of car C is: {the fault preference row vector of car A} = [1, 0, 0, 1, 0]

[0100] After similarity calculation, the similarity calculation result between car A and the first type of car Q is about 0.89 (retain two decimal places), the similarity calculation result between car B and the first type of car Q is about 0.94 (retain two decimal places), and the similarity calculation result between car C and the first type of car Q is 0.5. Therefore, it can be determined that car A and car B are first-type fault preference vehicles.

[0101] Step S17, selecting a fault label that meets a preset value from the fault preference row vector corresponding to the first type of fault preference vehicle as a first type of fault prediction result, and recommending the first type of fault prediction result to the first type of vehicle.

[0102] After determining the first type of fault preference vehicle, the most preferred fault among the first type of fault preference vehicles is recommended to the first type of car. In an embodiment of the present invention, specifically, the element value in the fault preference row vector corresponding to the first type of fault preference vehicle is compared with a preset element benchmark value, and the fault label with an element value greater than or equal to the element benchmark value is used as the first type of fault prediction result.

[0103] Taking the aforementioned cars A and B as examples, the preset element reference value is 2. In car A, the element value of engine failure is 3, and the element value of brake system failure is 1. Therefore, the engine failure greater than the element reference value is used as the first type of failure prediction result of car A relative to the first type of car Q. In car B, the element value of engine failure is 3, and the element value of brake system failure is 2. Therefore, the engine failure and brake system failure greater than the element reference value are used as the first type of failure prediction result of car B relative to the first type of car Q.

[0104] The embodiment of the present invention can also sort the cars according to the first type of similarity calculation results to obtain the car most similar to the first type of car, and then select the fault label with the highest element value from all fault labels of the most similar car as the first type of fault prediction result and recommend it to the first type of car.

[0105] Taking the aforementioned cars A and B as examples, the similarity calculation result between car A and the first type of car Q is about 0.89 (retain two decimal places), and the similarity calculation result between car B and the first type of car Q is about 0.94 (retain two decimal places). Therefore, car B is the first type of fault preference vehicle with the highest similarity to the first type of car Q. In car B, the element value of engine fault is 3, and the element value of brake system fault is 2. The element values ​​corresponding to engine fault and brake system fault are both greater than the element reference value. The engine fault with the highest element value is selected as the first type of fault prediction result and recommended to the first type of car.

[0106] The embodiment of the present invention can also collect the archive information of several cars, such as Figure 3 As shown, the archival information labels corresponding to each car are extracted from the archival information, such as brand, model, engine model, annual inspection time, purchase time and other labels, and the archival information labels are vectorized to generate the archival information feature vector corresponding to each car.

[0107] The labels in the archival information (such as brand, model, engine model, etc.) are converted into numerical form. And the numerical labels are standardized to eliminate the influence of different dimensions and numerical ranges. Common methods include Min-Max Normalization and Z-score Normalization. Each label is combined into a feature vector. For example, if the archival information includes five labels: brand, model, engine model, annual inspection time and purchase time, then the generated feature vector will contain the eigenvalues ​​of these five labels. In an embodiment of the present invention, it is necessary to ensure that the dimensions of the feature vectors are consistent. If the eigenvalue dimensions of some labels are different, they can be aligned by filling zero values ​​or other methods. In some cases, the importance of different archival information labels may be different. Therefore, each archival information label can be weighted according to the importance of the archival information label.

[0108] Specifically, when the vehicle to be predicted is a second type of vehicle, the method further includes:

[0109] The target file information of the second category of automobiles is obtained, and the target file information labels of the second category of automobiles, such as brand, model, engine model, annual inspection time, purchase time and other labels, are extracted from the target file information, and the target file information labels are vectorized to generate the target file information feature vector corresponding to the second category of automobiles.

[0110] The labels in the target archive information are converted into numerical form, and the numerical labels are standardized, and each label is combined into a target feature vector. For example, if the target archive information includes five labels, namely, brand, model, engine model, annual inspection time and purchase time, the target feature vector will contain the eigenvalues ​​of these five labels. In an embodiment of the present invention, it is necessary to ensure that the dimension of the target feature vector is consistent. If the eigenvalue dimensions of some labels are different, alignment can be performed by filling zero values ​​or other methods. In some cases, the importance of different target archive information labels may be different. Therefore, each target archive information label can be weighted according to the importance of the target archive information label.

[0111] The characteristic difference of each car is calculated based on the characteristic vector of the archive information. The specific calculation formula is as follows:

[0112]

[0113] Among them, u refers to car u, v refers to the various file information of car u, k refers to the total number of file information items, i refers to the starting item of the file information, γ i It refers to the weight of the i-th file information of car u, and attr(u, v) refers to the feature difference of car u.

[0114] The target characteristic difference of the second type of car is calculated based on the characteristic vector of the target file information. The specific calculation formula is the same as the formula for calculating the characteristic difference of each car.

[0115] The similarity between the feature difference value and the target feature difference value is calculated to obtain a second type of similarity calculation result. The specific similarity calculation formula is as follows:

[0116]

[0117] Among them, sim refers to the second type of similarity calculation result.

[0118] The second-category similarity calculation results are sorted in descending order to select second-category fault preference vehicles whose similarity with the second-category vehicles meets a preset value from a number of vehicles. In an embodiment of the present invention, a similarity value needs to be pre-set as an evaluation criterion, and after the second-category similarity calculation results are calculated, the second-category similarity calculation results are compared with the preset similarity value, wherein the vehicles corresponding to the second-category similarity calculation results that are greater than or equal to the similarity value are second-category fault preference vehicles similar to the second-category vehicles. Second-category fault preference vehicles refer to vehicles that are more likely to have the same fault as the vehicle to be predicted.

[0119] A fault label that meets a preset value is selected from the fault preference row vector corresponding to the second-category fault preference vehicle as a second-category fault prediction result, and the second-category fault prediction result is recommended to the second-category vehicle. After determining the second-category fault preference vehicle, the most preferred fault in the second-category fault preference vehicle is recommended to the second-category vehicle. In an embodiment of the present invention, this is specifically performed by comparing the element value in the fault preference row vector corresponding to the second-category fault preference vehicle with a preset element reference value, and taking the fault label whose element value is greater than or equal to the element reference value as the second-category fault prediction result.

[0120] In the embodiment of the present invention, even for low-risk vehicles with a small number of fault tags, valuable fault prediction results can be obtained by calculating the similarity of archive information. By utilizing the similarity in historical data, a more personalized fault prediction service is provided for low-risk vehicles, which helps to improve the reliability and safety of vehicles.

[0121] In another feasible embodiment of the present invention, when predicting automobile failure, it is also possible to perform classification prediction according to the vehicle model and vehicle series, or to perform classification prediction according to the destination of the vehicle.

[0122] An automobile fault prediction method provided in an embodiment of the present invention specifically comprises the following steps:

[0123] First, the maintenance records of several cars are collected, and the fault labels corresponding to each car are extracted from the maintenance records. The fault labels are vectorized to generate the fault preference row vector corresponding to each car. Then, the target maintenance records of the car to be predicted are obtained, and the target fault labels of the car to be predicted are extracted from the target maintenance records. The number of target fault labels is compared with the preset label threshold. If the number of target fault labels is greater than the preset label threshold, the car to be predicted is regarded as the first type of car. If the number of target fault labels is less than or equal to the preset label threshold, the car to be predicted is regarded as the second type of car.

[0124] When the car to be predicted is a first-class car, the target fault label of the first-class car is vectorized to generate a target fault row vector corresponding to the first-class car, and the similarity between the target fault row vector and the fault preference row vector is calculated to obtain a first-class similarity calculation result. According to the first-class similarity calculation result, a first-class fault preference vehicle whose similarity with the first-class car meets a preset value is screened out from a number of cars.

[0125] When the car to be predicted is a second-category car, firstly, the archive information of several cars is collected, and the archive information label corresponding to each car is extracted from the archive information, and the archive information label is vectorized to generate the archive information feature vector corresponding to each car. Then, the target archive information of the second-category car is obtained, and the target archive information label of the second-category car is extracted from the target archive information, and the target archive information label is vectorized to generate the target archive information feature vector corresponding to the second-category car, and the feature difference of each car is calculated based on the archive information feature vector, and the target feature difference of the second-category car is calculated based on the target archive information feature vector, and the similarity between the feature difference and the target feature difference is calculated to obtain the second-category similarity calculation result, and the second-category similarity calculation result is sorted in descending order to screen out the second-category fault preference car whose similarity with the second-category car meets the preset value from several cars.

[0126] Finally, the fault labels that meet the preset values ​​are selected from the fault preference row vectors corresponding to the first type of fault preference vehicles (second type of fault preference vehicles) as the first type of fault prediction results (second type of fault prediction results), and the first type of fault prediction results (second type of fault prediction results) are recommended to the first type of vehicles (second type of vehicles).

[0127] like Figure 4 As shown, the present invention discloses a vehicle fault prediction device, the device comprising:

[0128] The fault label extraction module 401 is used to collect maintenance records of several vehicles and extract the fault label corresponding to each vehicle from the maintenance records.

[0129] The fault label vectorization module 402 is used to perform vectorization processing on the fault labels to generate a fault preference row vector corresponding to each vehicle.

[0130] The target fault label extraction module 403 is used to obtain the target maintenance record of the vehicle to be predicted and extract the target fault label of the vehicle to be predicted from the target maintenance record.

[0131] The first comparison module 404 is used to compare the number of target fault labels with a preset label threshold to obtain a first type of automobile whose number of target fault labels is greater than the preset threshold.

[0132] The target fault label vectorization module 405 is used to perform vectorization processing on the target fault label of the first type of automobile to generate a target fault row vector corresponding to the first type of automobile.

[0133] The first type similarity calculation module 406 is used to calculate the similarity between the target fault row vector and the fault preference row vector to obtain a first type similarity calculation result, and select a first type fault preference vehicle whose similarity with the first type of vehicle meets a preset value from a plurality of vehicles according to the first type similarity calculation result.

[0134] The first type fault prediction result generating module 407 is used to select fault labels that meet preset values ​​from the fault preference row vectors corresponding to the first type fault preference vehicles as the first type fault prediction results, and recommend the first type fault prediction results to the first type vehicles.

[0135] In one embodiment, the apparatus further comprises:

[0136] The file information label extraction module is used to collect the file information of several cars and extract the file information label corresponding to each car from the file information.

[0137] The file information label vectorization module is used to vectorize the file information labels to generate the file information feature vector corresponding to each car.

[0138] In one embodiment, the fault label vectorization module 402 is specifically used to create a row vector for each car, where the row vector contains a number of element values, and the element values ​​are used to represent the frequency of occurrence of the fault label; according to a pre-set fault label order, the element value corresponding to each fault label is filled into the row vector to obtain a fault preference row vector corresponding to each car.

[0139] In one embodiment, the fault label vectorization module 402 is further specifically used to compare the element value in the fault preference row vector corresponding to the first type of fault preference vehicle with a preset element benchmark value, and take the fault label with an element value greater than or equal to the element benchmark value as the first type of fault prediction result.

[0140] In one embodiment, the first comparison module 404 is specifically used to treat the vehicle to be predicted as a first-category vehicle if the number of target fault labels is greater than a preset label threshold; and to treat the vehicle to be predicted as a second-category vehicle if the number of target fault labels is less than or equal to the preset label threshold.

[0141] In one embodiment, when the vehicle to be predicted is a second type of vehicle, the first comparison module 404 may further include:

[0142] The target file information label extraction module is used to obtain the target file information of the second category of vehicles and extract the target file information label of the second category of vehicles from the target file information.

[0143] The target file information label vectorization module is used to perform vectorization processing on the target file information label to generate a target file information feature vector corresponding to the second type of automobile.

[0144] The feature difference calculation module is used to calculate the feature difference of each car based on the file information feature vector; and calculate the target feature difference of the second type of car based on the target file information feature vector.

[0145] The second similarity calculation module is used to calculate the similarity between the feature difference value and the target feature difference value to obtain the second similarity calculation result.

[0146] The second-category fault preference car screening module is used to sort the second-category similarity calculation results in descending order, so as to screen out the second-category fault preference car whose similarity with the second-category car meets a preset value from a plurality of cars.

[0147] The second type fault prediction result generating module is used to select the fault labels meeting the preset values ​​from the fault preference row vectors corresponding to the second type fault preference vehicles as the second type fault prediction results, and recommend the second type fault prediction results to the second type vehicles.

[0148] In one embodiment, the first type similarity calculation module 406 is specifically used to calculate the cosine similarity between the target fault row vector and the fault preference row vector, and the cosine similarity is used to represent the degree of fault preference between the first type of car and several cars, wherein the cosine similarity is proportional to the degree of fault preference.

[0149] The embodiment of the present invention collects maintenance records of several vehicles, extracts the fault label corresponding to each vehicle from the maintenance records, and vectorizes the fault label to generate a fault preference row vector corresponding to each vehicle; obtains a target maintenance record of the vehicle to be predicted, extracts the target fault label of the vehicle to be predicted from the target maintenance record, compares the number of target fault labels with a preset label threshold, and obtains a first category of vehicles whose number of target fault labels is greater than the preset threshold; vectorizes the target fault label of the first category of vehicles to generate a target fault row vector corresponding to the first category of vehicles; calculates the similarity between the target fault row vector and the fault preference row vector to obtain a first category similarity calculation result, and selects a first category of fault preference vehicles whose similarity with the first category of vehicles meets a preset value from several vehicles according to the first category similarity calculation result; selects a fault label that meets a preset value from the fault preference row vector corresponding to the first category of fault preference vehicles as a first category of fault prediction result, and recommends the first category of fault prediction result to the first category of vehicles. The present invention combines the fault prediction technology of vehicle maintenance records and collaborative filtering algorithms, which can improve the accuracy of fault prediction and the personalized service level on the one hand, and optimize resource allocation, enhance user experience, and promote data-based decision making on the other hand.

[0150] The embodiment of the present invention further provides an electronic device, such as Figure 5 As shown, it includes a processor 501, a device interface 502, a memory 503 and a bus 504;

[0151] The memory 503 is used to store computer programs.

[0152] The processor 501 is used to execute the program stored in the memory 503, and implements the following steps:

[0153] Collect maintenance records of several cars and extract the fault label corresponding to each car from the maintenance records.

[0154] The fault labels are vectorized to generate a row vector of fault preferences corresponding to each car.

[0155] Obtain the target maintenance record of the vehicle to be predicted, and extract the target fault label of the vehicle to be predicted from the target maintenance record.

[0156] The number of target fault labels is compared with a preset label threshold to obtain a first category of cars whose number of target fault labels is greater than the preset threshold.

[0157] The target fault labels of the first category of vehicles are vectorized to generate target fault row vectors corresponding to the first category of vehicles.

[0158] The similarity between the target fault row vector and the fault preference row vector is calculated to obtain a first-category similarity calculation result, and a first-category fault preference vehicle whose similarity with the first-category vehicle meets a preset value is screened out from a number of vehicles according to the first-category similarity calculation result.

[0159] A fault label that meets a preset value is selected from the fault preference row vector corresponding to the first-category fault preference vehicle as a first-category fault prediction result, and the first-category fault prediction result is recommended to the first-category vehicle.

[0160] The bus mentioned in the above terminal can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0161] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0162] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can 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 gate or transistor logic devices, discrete hardware components.

[0163] The present invention also provides a storage medium. When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute the automobile fault prediction method of the aforementioned embodiment.

[0164] As 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.

[0165] The algorithm and display provided herein are not inherently related to any particular computer, virtual device or other equipment. According to the above description, it is obvious that the structure required for constructing this type of device is. In addition, the present invention is not directed to any specific programming language. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description of the above specific language is to disclose the best mode of the present invention.

[0166] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0167] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: that the claimed invention requires more features than those expressly recited in each claim. Rather, as reflected in the claims below, inventive aspects lie in less than all of the features of the individual embodiments disclosed above. Therefore, the claims that follow the detailed description are hereby expressly incorporated into the detailed description, with each claim itself serving as a separate embodiment of the present invention.

[0168] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0169] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in the sorting device according to the present invention. The present invention may also be implemented as a device or apparatus program for executing part or all of the methods described herein. Such a program for implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0170] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.

[0171] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, apparatuses and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0172] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0173] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0174] It should be noted that the various data-related processes in the embodiments of the present invention are all carried out in compliance with the corresponding data protection laws and policies of the country where the device is located and with the authorization given by the corresponding device owner.

Claims

1. A method for predicting automobile failure, characterized in that: The method comprises: Collecting maintenance records of several vehicles, and extracting a fault label corresponding to each vehicle from the maintenance records; Vectorizing the fault labels to generate a fault preference row vector corresponding to each of the vehicles; Obtaining a target maintenance record of the vehicle to be predicted, and extracting a target fault label of the vehicle to be predicted from the target maintenance record; Comparing the number of target fault labels with a preset label threshold to obtain a first type of automobile whose number of target fault labels is greater than the preset threshold; Vectorizing the target fault labels of the first type of vehicles to generate target fault row vectors corresponding to the first type of vehicles; Calculating the similarity between the target fault row vector and the fault preference row vector to obtain a first-type similarity calculation result, and selecting a first-type fault preference vehicle whose similarity with the first-type vehicle meets a preset value from a plurality of the vehicles according to the first-type similarity calculation result; A fault label that meets a preset value is selected from the fault preference row vector corresponding to the first type of fault preference vehicle as a first type of fault prediction result, and the first type of fault prediction result is recommended to the first type of vehicle.

2. The method according to claim 1, characterized in that The method further comprises: Collect the archival information of several cars, and extract the archival information label corresponding to each car from the archival information; vectorize the archival information label to generate the archival information feature vector corresponding to each car.

3. The method according to claim 1, characterized in that The vectorization processing of the fault label includes: creating a row vector for each of the vehicles, the row vector including a plurality of element values, the element values ​​being used to represent the occurrence frequency of the fault label; According to a preset fault label sequence, the element value corresponding to each fault label is filled into the row vector to obtain a fault preference row vector corresponding to each of the vehicles.

4. The method according to claim 3, characterized in that The step of selecting a fault label that satisfies a preset value from the fault preference row vector corresponding to the first type of fault preference vehicle as a first type of fault prediction result includes: The element value in the fault preference row vector corresponding to the first type of fault preference vehicle is compared with a preset element reference value, and the fault label with the element value greater than or equal to the element reference value is taken as the first type of fault prediction result.

5. The method according to claim 1, characterized in that The comparing the number of the target fault labels with a preset label threshold includes: If the number of the target fault labels is greater than the preset label threshold, the vehicle to be predicted is regarded as a first-category vehicle; If the number of the target fault labels is less than or equal to the preset label threshold, the vehicle to be predicted is regarded as a second type of vehicle.

6. The method according to claim 5, characterized in that In the case where the to-be-predicted automobile is the second-category automobile, the method further includes: Acquire target file information of the second type of automobile, and extract target file information tags of the second type of automobile from the target file information; Performing vectorization processing on the target file information label to generate a target file information feature vector corresponding to the second type of automobile; Calculating a feature difference value of each of the vehicles based on the feature vector of the archive information; Calculating a target feature difference of the second type of automobile based on the target profile information feature vector; Calculating the similarity between the feature difference value and the target feature difference value to obtain a second type of similarity calculation result; Sorting the second-category similarity calculation results in descending order, so as to select the second-category fault preference car whose similarity with the second-category car meets a preset value from the plurality of cars; A fault label that meets a preset value is selected from the fault preference row vector corresponding to the second-category fault preference vehicle as a second-category fault prediction result, and the second-category fault prediction result is recommended to the second-category vehicle.

7. The method according to claim 1, characterized in that The calculating the similarity between the target fault row vector and the fault preference row vector to obtain a first type of similarity calculation result includes: The cosine similarity between the target fault row vector and the fault preference row vector is calculated, wherein the cosine similarity is used to represent the degree of fault preference between the first type of vehicle and a plurality of the first type of vehicles, wherein the cosine similarity is proportional to the degree of fault preference.

8. A vehicle fault prediction device, characterized in that: The device comprises: A fault label extraction module is used to collect maintenance records of several vehicles and extract a fault label corresponding to each vehicle from the maintenance records; A fault label vectorization module, used for performing vectorization processing on the fault label to generate a fault preference row vector corresponding to each of the vehicles; A target fault label extraction module is used to obtain a target maintenance record of the vehicle to be predicted, and extract a target fault label of the vehicle to be predicted from the target maintenance record; A first comparison module, used for comparing the number of the target fault labels with a preset label threshold, so as to obtain a first type of automobile whose number of the target fault labels is greater than the preset threshold; a target fault label vectorization module, used for performing vectorization processing on the target fault label of the first type of vehicle to generate a target fault row vector corresponding to the first type of vehicle; A first-type similarity calculation module, configured to calculate the similarity between the target fault row vector and the fault preference row vector to obtain a first-type similarity calculation result, and to select a first-type fault preference vehicle whose similarity with the first-type vehicle meets a preset value from a plurality of the vehicles according to the first-type similarity calculation result; The first type of fault prediction result generating module is used to select the fault labels meeting the preset values ​​from the fault preference row vectors corresponding to the first type of fault preference vehicles as the first type of fault prediction results, and recommend the first type of fault prediction results to the first type of vehicles.

9. An electronic device, characterized in that: include: at least one processor; a memory communicatively coupled to the at least one processor; The memory has a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images