CN114691680A - Maintenance case information query method and device, electronic equipment and storage medium - Google Patents
Maintenance case information query method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN114691680A CN114691680A CN202210217501.XA CN202210217501A CN114691680A CN 114691680 A CN114691680 A CN 114691680A CN 202210217501 A CN202210217501 A CN 202210217501A CN 114691680 A CN114691680 A CN 114691680A
- Authority
- CN
- China
- Prior art keywords
- information
- fault
- maintenance case
- feature
- maintenance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/22—Indexing; Data structures therefor; Storage structures
- G06F16/2228—Indexing structures
- G06F16/2237—Vectors, bitmaps or matrices
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2455—Query execution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/248—Presentation of query results
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
Abstract
The invention provides a maintenance case information query method, a maintenance case information query device, electronic equipment and a storage medium, wherein fault information to be queried is firstly acquired; then, extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector; further calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information; and finally, selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried according to the similarity. According to the method, a professional engineer is not required to perform field service, target maintenance case information can be determined in a query mode, corresponding maintenance information is found, and a user can maintain the fault equipment through the maintenance information, so that not only can the labor cost be greatly reduced, but also the time consumed by fault positioning and the enterprise cost can be reduced.
Description
Technical Field
The invention relates to the technical field of maintenance case query, in particular to a method and a device for querying maintenance case information, electronic equipment and a storage medium.
Background
With the rapid development of scientific technology and industrial internet application, the fault maintenance of equipment becomes the work key point of the after-sales departments of enterprises.
Currently, most companies often adopt professional engineers to carry out fault maintenance service on site, which not only increases labor cost, but also increases time consumption for fault location and enterprise cost.
Therefore, it is urgently needed to provide a maintenance case information query method.
Disclosure of Invention
The invention provides a maintenance case information query method, a maintenance case information query device, electronic equipment and a storage medium, which are used for overcoming the defects in the prior art.
The invention provides a maintenance case information query method, which comprises the following steps:
acquiring fault information to be inquired;
extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector;
calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information;
and selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarity.
According to the maintenance case information query method provided by the invention, the fault information to be queried comprises a plurality of fault information in different forms;
correspondingly, the extracting the features of the fault information to be queried to obtain a fault feature vector includes:
respectively extracting the characteristics of the multiple items of fault information to obtain characteristic vectors corresponding to the multiple items of fault information;
and fusing the characteristic vectors corresponding to the multiple items of fault information to obtain the fault characteristic vector.
According to the maintenance case information query method provided by the invention, the plurality of items of fault information comprise at least two items of fault text information, fault image information, fault video information and fault audio information.
According to the maintenance case information query method provided by the invention, the respectively extracting the characteristics of the plurality of items of fault information to obtain the characteristic vectors corresponding to the plurality of items of fault information comprises the following steps:
for any fault information in the multiple items of fault information, extracting the features of the any fault information based on a feature extraction model corresponding to the any fault information to obtain a feature vector corresponding to the any fault information;
the feature extraction model corresponding to any fault information is obtained based on sample fault information training corresponding to any fault information carrying a feature vector label.
According to the maintenance case information query method provided by the invention, the fusing the feature vectors corresponding to the plurality of items of fault information to obtain the fault feature vector comprises the following steps:
inputting the feature vectors corresponding to the multiple items of fault information into a multi-modal feature fusion model, and fusing the feature vectors corresponding to the multiple items of fault information by using the multi-modal feature fusion model to obtain and output the fault feature vectors;
the multi-modal feature fusion model is obtained by training based on feature vector samples of sample fault information corresponding to multiple items of fault information carrying fusion result labels.
According to the maintenance case information query method provided by the invention, the calculating of the similarity between the fault feature vector and the case fault feature vector corresponding to each maintenance case information comprises the following steps:
inputting the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information into a neural collaborative filtering model, and calculating and outputting the similarity between the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information by the neural collaborative filtering model;
the neural collaborative filtering model is obtained by training based on a first feature vector sample, a plurality of second feature vector samples and similarity labels between the first feature vector sample and the plurality of second feature vector samples.
According to the maintenance case information query method provided by the invention, based on each similarity, the step of selecting target maintenance case information from each maintenance case information as a query result corresponding to the fault information to be queried comprises the following steps:
sorting the maintenance case information according to the similarity to obtain a sorting result;
and determining the target maintenance case information as the query result based on the sequencing result.
The invention also provides a maintenance case information query device, which comprises:
the information acquisition module is used for acquiring the fault information to be inquired;
the characteristic extraction module is used for extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector;
the similarity calculation module is used for calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each piece of maintenance case information;
and the query result determining module is used for selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarity.
The invention also provides an electronic device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the program to realize the steps of the maintenance case information query method.
The present invention also provides a non-transitory computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the service case information query method as described in any of the above.
According to the maintenance case information query method, the maintenance case information query device, the electronic equipment and the storage medium, firstly, fault information to be queried is obtained; then, extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector; further calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information; and finally, selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried according to the similarity. According to the method, a professional engineer is not required to perform field service, target maintenance case information can be determined in a query mode, corresponding maintenance information is found, and a user can maintain the fault equipment through the maintenance information, so that not only can the labor cost be greatly reduced, but also the time consumed by fault positioning and the enterprise cost can be reduced.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for querying information about a maintenance case according to the present invention;
FIG. 2 is a schematic flow chart of the feature extraction and feature fusion steps provided by the present invention;
FIG. 3 is a schematic structural diagram of a maintenance case information query device provided in the present invention;
fig. 4 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the prior art, when the faulty equipment is maintained, a mode that a professional engineer carries out fault maintenance service on site is usually adopted, or a user searches for a maintenance case by a keyword retrieval method, so that maintenance information in the maintenance case information is obtained, and the user can maintain the faulty equipment by the maintenance information. The adoption of the field service mode of professional engineers greatly increases the labor cost, the time consumption of fault location and the enterprise cost. The maintenance cases are searched by adopting a method based on keyword search, so that the search is limited to the maintenance cases represented by characters, inevitably, the situation of unclear description can occur in the maintenance process, the accuracy of the search result can be reduced, and further, a user cannot obtain a satisfactory maintenance scheme, and the time consumption for fault location and the enterprise cost are increased in a variable manner. Therefore, the embodiment of the invention provides a maintenance case information query method.
Fig. 1 is a schematic flow chart of a method for querying information of a maintenance case provided in an embodiment of the present invention, as shown in fig. 1, the method includes:
s1, acquiring the fault information to be inquired;
s2, extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector;
s3, calculating the similarity between the fault feature vector and the case fault feature vector corresponding to each maintenance case information;
and S4, based on the similarity, selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried.
Specifically, in the maintenance case information query method provided in the embodiment of the present invention, the execution main body is a maintenance case information query device, the maintenance case information query device may be configured in a server, the server may be a local server or a cloud server, the local server may specifically be a computer, a tablet computer, a smart phone, and the like, and the embodiment of the present invention is not particularly limited thereto. It can be understood that the maintenance case information query device may be configured with a display interface, through which the fault information to be queried obtained in step S1 and the query result corresponding to the fault information to be queried obtained in step S4 may be displayed.
The maintenance case information related to the method, namely the historical maintenance case information, refers to the relevant information of the whole process that the same type of equipment of the historical fault equipment fails and completes maintenance. A fault occurrence and a corresponding maintenance action are completed, which is called a maintenance case. The service case information may include basic information, trouble information, and service information, which may be represented by a trouble service order.
First, step S1 is executed to obtain the fault information to be queried. An input box for inputting information by a user can be arranged on a display interface of the maintenance case information inquiry device, and the user can input the fault information to be inquired in the input box, so that the maintenance case information inquiry device can acquire and display the fault information to be inquired, and the user can further confirm whether the input fault information to be inquired is correct. Here, the user may be a serviceman, a customer service person, or a user of the faulty device.
In addition, the fault information to be inquired can be acquired through information acquisition equipment, the information acquisition equipment is in communication connection with the maintenance case information inquiry device, so that the fault information to be inquired acquired by the information acquisition equipment can be transmitted to the maintenance case information inquiry device, and a display interface of the maintenance case information inquiry device can display the acquired fault information to be inquired.
It is understood that the fault information to be queried may be fault information of a fault device that needs to determine a maintenance plan, and the fault device may be a device in any field, such as a work machine in an industrial field, a power distribution network in an electric field, an electric device, a property device in a property management field, and the like, and is not limited herein.
The to-be-queried fault information may include several fault information items in different forms, for example, the fault information item may be one fault information item of fault text information, fault image information, fault video information, and fault audio information, or a combination of at least two fault information items of text information, image information, video information, and audio information, which is not limited herein.
The text information refers to fault information described by a user in text, such as 'hydraulic pump is not controlled'; the image information refers to fault information represented by a picture, for example, a picture obtained by shooting a fault position of a fault device by a user; the video information refers to fault information represented by a video, for example, a video obtained by shooting a fault phenomenon of a fault device by a user, the length of the video can be adjusted according to needs, and the video can be a short video or a longer video, and is not specifically limited herein; the audio information refers to fault information described by a user through voice, and for example, the fault information can be obtained by dictating fault information of a fault device by the user and recording the fault information through a recording device.
And then, executing step S2, and performing feature extraction on the fault information to be queried to obtain a fault feature vector of the fault information to be queried. This process may be implemented by a feature extraction model. It can be understood that a plurality of items of fault information in different forms contained in the fault information to be queried need to be subjected to feature extraction in different ways, so that the plurality of items of fault information in different forms can be subjected to feature extraction by using the same feature extraction model, and at this time, the feature extraction model needs to have a function of extracting feature vectors of each item of fault information at the same time, and also needs to have a function of feature vector fusion.
For example, the feature extraction model may include a feature fusion layer and feature extraction layers corresponding to various forms of information, each feature extraction layer is connected to the feature fusion layer, each feature extraction layer is configured to extract a feature vector of fault information in a corresponding form, and the feature fusion layer is configured to fuse feature vectors input therein to obtain a fusion result. In the embodiment of the invention, the feature extraction model can be provided with feature extraction layers corresponding to four forms of fault information and respectively used for extracting feature vectors in fault text information, fault image information, fault video information and fault audio information.
If the fault information to be inquired only has fault information in a certain form, namely only one item of fault information, only the feature extraction layer corresponding to the item of fault information in the feature extraction model works, the input and the output of other feature extraction layers are both empty, and the input and the output of the feature fusion layer are both the output of the feature extraction layer corresponding to the item of fault information, namely the feature vector of the item of fault information. At this time, the feature vector of the fault information is the fault feature vector of the fault information to be queried.
If the fault information to be inquired comprises at least two forms of fault information, namely at least two items of fault information, the feature extraction layer corresponding to each item of fault information in the fault information to be inquired in the feature extraction model works, the input and the output of other feature extraction layers are both null, the input of the feature fusion layer is a feature vector of each item of fault information, and the output is a fusion result of the feature vectors of each item of fault information, namely the fault feature vector of the fault information to be inquired.
The feature extraction model may be obtained by training various fault information samples carrying fault feature vector labels in different forms, and is not specifically limited herein. It can be understood that each fault information sample is the same as the type of the fault device corresponding to the fault information to be queried.
Then, step S3 is executed to calculate the similarity between the fault feature vector of the fault information to be queried and the case fault feature vector corresponding to each maintenance case information. The maintenance case information refers to the information related to each fault and the corresponding maintenance scheme of the same type of equipment of the fault equipment in the life cycle. Each maintenance case corresponds to one piece of maintenance case information, and each piece of maintenance case information comprises basic information, fault information and maintenance information.
The basic information in each repair case information may include the fault repair person and his contact phone, the repair type, the area, the faulty equipment address, the faulty repair person and his contact phone, etc.
The fault information in each service case information may include related description information of the fault, for example, may include several items of fault information in different forms, such as fault text information, fault image information, fault video information, and fault audio information, which is not limited herein.
Preferably, in order to improve the effectiveness of the service case information query, the fault information in each service case information may include all forms of fault information, that is, at least may include fault text information, fault image information, fault video information, and fault audio information.
The maintenance information in each maintenance case information refers to a maintenance scheme, maintenance grade information and the like made for the fault information, and may further include evaluation information, score information and the like of the fault repair staff for the fault repair staff.
The case fault feature vector corresponding to each maintenance case information is a feature vector of the fault information in each maintenance case information, and a method for extracting features of the fault information to be inquired can be adopted in advance, the fault information in each maintenance case information can be obtained by extracting the features, or other methods can be adopted in advance, and the method is not particularly limited here.
And then, calculating the similarity between the fault feature vector of the fault information to be inquired and the case fault feature vector corresponding to each maintenance case information, wherein the similarity can be calculated in a cosine similarity calculation mode or a neural network model, and the similarity calculation mode is not particularly limited. The similarity may be in the form of a percentage or a percentile score.
The similarity is used for representing the similarity between the fault information to be inquired and the fault information in each maintenance case information, and the similarity is larger if the similarity is larger.
Finally, step S4 is executed, and target maintenance case information can be selected from each maintenance case information according to the similarity between the fault feature vector of the fault information to be queried and the case fault feature vector corresponding to each maintenance case information, and the target maintenance case information is used as a query result corresponding to the fault information to be queried. The target maintenance case information may be a plurality of maintenance case information with high similarity corresponding to each maintenance case information.
Furthermore, the fault information to be inquired can be subjected to fault positioning and maintenance according to the maintenance information in the target maintenance case information, and time consumption of fault positioning and enterprise cost can be greatly reduced.
The maintenance case information query method provided by the embodiment of the invention comprises the steps of firstly obtaining fault information to be queried; then, extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector; further calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information; and finally, selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried according to the similarity. According to the method, a professional engineer is not required to perform field service, target maintenance case information can be determined in a query mode, corresponding maintenance information is found, and a user can maintain the fault equipment through the maintenance information, so that not only can the labor cost be greatly reduced, but also the time consumed by fault positioning and the enterprise cost can be reduced.
On the basis of the above embodiment, in the maintenance case information query method provided in the embodiment of the present invention, the fault information to be queried includes a plurality of fault information in different forms;
correspondingly, the extracting the features of the fault information to be queried to obtain a fault feature vector includes:
respectively extracting the characteristics of the multiple items of fault information to obtain characteristic vectors corresponding to the multiple items of fault information;
and fusing the characteristic vectors corresponding to the multiple items of fault information to obtain the fault characteristic vector.
Specifically, in the embodiment of the present invention, the fault information to be queried may include multiple items of fault information in different forms, that is, at least two items of fault information in different forms. For example, at least two items of failure text information, failure image information, failure video information, and failure audio information.
Correspondingly, when the feature extraction is performed on the fault information to be queried to obtain the fault feature vector, the feature extraction may be performed on each item of fault information in different forms to obtain the feature vector corresponding to each item of fault information. Here, different feature extraction algorithms can be respectively used for extracting features of each item of fault information, and the specific algorithm can be selected according to needs, which is not specifically limited here.
Then, in order to fully utilize each item of fault information, the feature vectors corresponding to each item of fault information can be fused, so that the fault feature vector of the fault information to be inquired can be obtained, and the obtained fault feature vector can fully represent the fault information to be inquired.
During fusion, the feature vectors corresponding to each item of fault information may be spliced into a fault feature vector in a direct fusion manner, or the feature vectors corresponding to each item of fault information may be added or multiplied element by element to obtain a fault feature vector, where the fusion manner is not specifically limited.
In the embodiment of the invention, in the process of extracting the characteristics of the fault information to be inquired, when the fault information to be inquired contains a plurality of fault information, the characteristics are respectively extracted and the characteristics are fused, so that the obtained fault characteristic vector can represent the complete fault information to be inquired, the calculation of the subsequent similarity is convenient, and the accuracy of the recommendation result can be improved.
On the basis of the foregoing embodiment, the method for querying information of a maintenance case, provided in the embodiment of the present invention, for performing feature extraction on the multiple items of fault information to obtain feature vectors corresponding to the multiple items of fault information respectively includes:
for any fault information in the multiple items of fault information, extracting the features of the any fault information based on a feature extraction model corresponding to the any fault information to obtain a feature vector corresponding to the any fault information;
the feature extraction model corresponding to any fault information is obtained by training based on sample fault information corresponding to any fault information carrying a feature vector label.
Specifically, in the embodiment of the present invention, when feature extraction is performed on multiple items of fault information respectively to obtain feature vectors corresponding to multiple items of fault information, for any item of fault information in the multiple items of fault information, the any item of fault information may be each item of fault information in the multiple items of fault information, a feature extraction model corresponding to the any item of fault information may be introduced, and feature extraction is performed on the any item of fault information through the feature extraction model corresponding to the any item of fault information to obtain the feature vector corresponding to the any item of fault information.
It can be understood that the feature extraction model corresponding to any item of fault information may be constructed based on a neural network model, and is obtained by training sample fault information corresponding to any item of fault information carrying a feature vector label.
For example, if any one piece of fault information is fault text information, the feature extraction model corresponding to any one piece of fault information may be a text feature extraction model, and the text feature extraction model may be used to perform word segmentation, keyword extraction, entity word extraction, sentence vector extraction, and the like on the fault text information, and the obtained feature vector corresponding to any one piece of fault information may include a word vector and a sentence vector, and the word vector may include a keyword vector and an entity word vector, and the like.
Here, the text feature extraction model may be constructed based on a bert model, an LSTM model, an RCNN model, a BiLSTM model, an AlBert model, or a fastcnn model, etc. The text feature extraction model can be obtained through sample fault information training corresponding to the fault text information carrying the text feature vector label.
If any fault information is fault image information, the feature extraction model corresponding to any fault information may be an image feature extraction model, and the fault image information may be input to the image feature extraction model, so that an image feature vector corresponding to the fault image information may be obtained.
Here, the image feature extraction model may be constructed based on a bert model, an LSTM model, an RCNN model, a BiLSTM model, an AlBert model, or a fastcnn model, or the like. The image feature extraction model can be obtained through sample fault information training corresponding to fault image information carrying image feature vector labels.
If any fault information is fault video information, the feature extraction model corresponding to any fault information can be a video feature extraction model, and the fault video information can be input into the video feature extraction model to obtain a video feature vector corresponding to the fault video information.
Here, the video feature extraction model may be constructed based on a bert model, an LSTM model, an RCNN model, a BiLSTM model, an AlBert model, or a fastcnn model, etc. The video feature extraction model can be obtained through sample fault information training corresponding to fault video information carrying a video feature vector label.
If any fault information is fault audio information, the feature extraction model corresponding to any fault information may be an audio feature extraction model, and the fault audio information may be input to the audio feature extraction model, so that an audio feature vector corresponding to the fault audio information may be obtained.
Here, the audio feature extraction model may be constructed based on a bert model, an LSTM model, an RCNN model, a BiLSTM model, an AlBert model, or a fastcnn model, etc. The audio feature extraction model can be obtained through sample fault information training corresponding to fault audio information carrying an audio feature vector label.
In the embodiment of the invention, when the feature extraction is carried out, the corresponding feature extraction model is introduced for realizing, so that the feature extraction efficiency can be improved, and the retrieval efficiency can be further improved.
On the basis of the above embodiment, the method for querying information of a maintenance case, provided in the embodiment of the present invention, for fusing the feature vectors corresponding to the plurality of items of fault information to obtain the fault feature vector, includes:
inputting the feature vectors corresponding to the multiple items of fault information into a multi-modal feature fusion model, and fusing the feature vectors corresponding to the multiple items of fault information by using the multi-modal feature fusion model to obtain and output the fault feature vectors;
the multi-modal feature fusion model is obtained by training based on feature vector samples of sample fault information corresponding to multiple items of fault information carrying fusion result labels.
Specifically, in the embodiment of the present invention, when feature vectors corresponding to multiple items of fault information are fused to obtain a fault feature vector, the feature vectors corresponding to multiple items of fault information may be input to the multimodal feature fusion model, and the multimodal feature fusion model fuses the feature vectors corresponding to multiple items of fault information to obtain and output the fault feature vector.
The multi-modal feature fusion model may be a multi-modal decomposed Bilinear pool (MFB) model including a fully connected layers (FC)), an operation layer, and a pooling layer, and converts each input feature vector into a high-dimensional vector through the fully connected layers. And performing element-by-element operation on the high-dimensional vectors corresponding to the feature vectors through an operation layer to obtain an operation result. The operation may be a multiplication operation, an addition operation, or the like, and the SUM posing operation is performed on the operation result through the pooling layer to obtain the fault feature vector.
It can be understood that the multi-modal feature fusion model can be obtained by training feature vector samples of sample fault information corresponding to multiple items of fault information carrying fusion result labels.
In the embodiment of the invention, when the characteristics are fused, a multi-mode characteristic fusion model is introduced, so that the characteristic fusion efficiency can be improved, and the information retrieval efficiency can be further improved.
Fig. 2 is a schematic flow chart of the feature extraction and feature fusion steps, and as shown in fig. 2, the to-be-queried fault information includes four items of fault information including fault text information, fault image information, fault video information, and fault audio information, the fault text information performs text feature extraction, the fault image information performs image feature extraction, the fault video information performs video feature extraction, and the fault audio information performs audio feature extraction, so as to obtain text feature vectors, image feature vectors, video feature vectors, and audio feature vectors, respectively. And then, inputting each feature vector into a full connection layer, an operation layer and a pooling layer of the multi-modal feature fusion model, and finally obtaining a fault feature vector.
On the basis of the foregoing embodiment, the method for querying maintenance case information provided in the embodiment of the present invention for calculating the similarity between the fault feature vector and the case fault feature vector corresponding to each maintenance case information includes:
inputting the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information into a neural collaborative filtering model, and calculating and outputting the similarity between the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information by the neural collaborative filtering model;
the neural collaborative filtering model is obtained by training based on a first feature vector sample, a plurality of second feature vector samples and similarity labels between the first feature vector sample and the plurality of second feature vector samples.
Specifically, in the embodiment of the present invention, when the similarity between the fault feature vector and each case fault feature vector is calculated, a Neural Collaborative Filtering (NCF) model may be introduced, the fault feature vector and each case fault feature vector are used as the input of the NCF model, the similarity between the fault feature vector and each case fault feature vector is calculated through the NCF model, and the NCF model outputs each similarity in the form of score.
It is understood that the NCF model is trained by the first feature vector sample, the plurality of second feature vector samples, and the similarity label between the first feature vector sample and the plurality of second feature vector samples.
In the embodiment of the invention, the similarity among the characteristic vectors is calculated through the neural collaborative filtering model, so that the calculation result of the similarity is more accurate, the calculation efficiency is also improved, and the accuracy of the retrieval result and the retrieval efficiency are further improved.
On the basis of the above embodiment, the method for querying maintenance case information provided in the embodiment of the present invention, based on each similarity, selecting target maintenance case information from each maintenance case information as a query result corresponding to the fault information to be queried, includes:
sorting the maintenance case information according to the similarity to obtain a sorting result;
and determining the target maintenance case information as the query result based on the sequencing result.
Specifically, in the embodiment of the present invention, when the target maintenance case information is selected from the maintenance case information as the query result corresponding to the fault information to be queried according to the similarities, the maintenance case information may be ranked according to the similarities to obtain a ranking result. In the sorting, the similarity values may be arranged in an ascending order or in a descending order, which is not specifically limited herein.
And then, determining target maintenance case information as a query result according to the obtained sequencing result. The target maintenance case information may be maintenance case information corresponding to the first N similarities in the ascending sorting result, or may be maintenance case information corresponding to the last N similarities in the descending sorting result. Wherein N is more than or equal to 1.
In the embodiment of the invention, the query results are determined by sequencing the information of the maintenance cases according to the similarity, so that the query results can better meet the requirements of users, and the user experience is improved.
As shown in fig. 3, on the basis of the above embodiment, an embodiment of the present invention provides a maintenance case information query apparatus, including:
the information acquisition module 31 is configured to acquire fault information to be queried;
the feature extraction module 32 is configured to perform feature extraction on the fault information to be queried to obtain a fault feature vector;
the similarity calculation module 33 is configured to calculate similarities between the fault feature vectors and case fault feature vectors corresponding to the maintenance case information;
and the query result determining module 34 is configured to select target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarities.
On the basis of the above embodiment, in the maintenance case information query device provided in the embodiment of the present invention, the fault information to be queried includes a plurality of fault information in different forms;
accordingly, the feature extraction module is configured to:
respectively extracting the characteristics of the multiple items of fault information to obtain characteristic vectors corresponding to the multiple items of fault information;
and fusing the characteristic vectors corresponding to the multiple items of fault information to obtain the fault characteristic vector.
On the basis of the above embodiment, in the maintenance case information query device provided in the embodiment of the present invention, the plurality of items of fault information include at least two items of fault text information, fault image information, fault video information, and fault audio information.
On the basis of the foregoing embodiment, in the maintenance-case-information query apparatus provided in the embodiment of the present invention, the feature extraction module is specifically configured to:
for any fault information in the multiple items of fault information, extracting the features of the any fault information based on a feature extraction model corresponding to the any fault information to obtain a feature vector corresponding to the any fault information;
the feature extraction model corresponding to any fault information is obtained by training based on sample fault information corresponding to any fault information carrying a feature vector label.
On the basis of the foregoing embodiment, the feature extraction module of the maintenance case information query device provided in the embodiment of the present invention is further specifically configured to:
inputting the feature vectors corresponding to the multiple items of fault information into a multi-modal feature fusion model, and fusing the feature vectors corresponding to the multiple items of fault information by using the multi-modal feature fusion model to obtain and output the fault feature vectors;
the multi-modal feature fusion model is obtained by training based on feature vector samples of sample fault information corresponding to multiple items of fault information carrying fusion result labels.
On the basis of the foregoing embodiment, in the maintenance-case information query apparatus provided in the embodiment of the present invention, the similarity calculation module is configured to:
inputting the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information into a neural collaborative filtering model, and calculating and outputting the similarity between the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information by the neural collaborative filtering model;
the neural collaborative filtering model is obtained by training based on a first feature vector sample, a plurality of second feature vector samples and similarity labels between the first feature vector sample and the plurality of second feature vector samples.
On the basis of the foregoing embodiment, in the maintenance-case-information query apparatus provided in the embodiment of the present invention, the query-result determining module is configured to:
sorting the maintenance case information according to the similarity to obtain a sorting result;
and determining the target maintenance case information as the query result based on the sequencing result.
Specifically, the functions of the modules in the maintenance case information query device provided in the embodiment of the present invention correspond to the operation flows of the steps in the method embodiments one to one, and the implementation effects are also consistent.
Fig. 4 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 4: a processor (processor)410, a communication Interface 420, a memory (memory)430 and a communication bus 440, wherein the processor 410, the communication Interface 420 and the memory 430 are communicated with each other via the communication bus 440. The processor 410 may call logic instructions in the memory 430 to perform the service case information query method provided in the embodiments described above, the method comprising: acquiring fault information to be inquired; extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector; calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information; and selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarity.
In addition, the logic instructions in the memory 430 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention or a part thereof which substantially contributes to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, which when executed by a computer, enable the computer to perform the method for querying service case information provided in the above embodiments, the method comprising: acquiring fault information to be inquired; extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector; calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information; and selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarity.
In yet another aspect, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, the computer program being implemented by a processor to perform the method for service case information query provided in the above embodiments, the method including: acquiring fault information to be inquired; extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector; calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information; and selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarity.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment may be implemented by software plus a necessary general hardware platform, and may also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A maintenance case information query method is characterized by comprising the following steps:
acquiring fault information to be inquired;
extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector;
calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each maintenance case information;
and selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarity.
2. The maintenance case information query method according to claim 1, wherein the fault information to be queried includes a plurality of fault information in different forms;
correspondingly, the extracting the features of the fault information to be queried to obtain a fault feature vector includes:
respectively extracting the characteristics of the multiple items of fault information to obtain characteristic vectors corresponding to the multiple items of fault information;
and fusing the characteristic vectors corresponding to the multiple items of fault information to obtain the fault characteristic vector.
3. The service case information query method of claim 2, wherein the plurality of items of failure information include at least two items of failure text information, failure image information, failure video information, and failure audio information.
4. The method for querying information of a maintenance case as claimed in claim 2, wherein the respectively performing feature extraction on the plurality of items of fault information to obtain feature vectors corresponding to the plurality of items of fault information comprises:
for any fault information in the multiple items of fault information, extracting the features of the any fault information based on a feature extraction model corresponding to the any fault information to obtain a feature vector corresponding to the any fault information;
the feature extraction model corresponding to any fault information is obtained by training based on sample fault information corresponding to any fault information carrying a feature vector label.
5. The method for querying information of a maintenance case according to claim 2, wherein the fusing the feature vectors corresponding to the plurality of items of fault information to obtain the fault feature vector comprises:
inputting the feature vectors corresponding to the multiple items of fault information into a multi-modal feature fusion model, and fusing the feature vectors corresponding to the multiple items of fault information by using the multi-modal feature fusion model to obtain and output the fault feature vectors;
the multi-modal feature fusion model is obtained by training based on feature vector samples of sample fault information corresponding to multiple items of fault information carrying fusion result labels.
6. The method for querying information on maintenance cases as claimed in any one of claims 1 to 5, wherein the calculating the similarity between the fault feature vector and the case fault feature vector corresponding to each maintenance case information includes:
inputting the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information into a neural collaborative filtering model, and calculating and outputting the similarity between the fault feature vector and the case fault feature vector corresponding to each piece of maintenance case information by the neural collaborative filtering model;
the neural collaborative filtering model is obtained by training based on a first feature vector sample, a plurality of second feature vector samples and similarity labels between the first feature vector sample and the plurality of second feature vector samples.
7. The method for querying the maintenance case information according to any one of claims 1 to 5, wherein the selecting target maintenance case information from the maintenance case information as the query result corresponding to the fault information to be queried based on the similarities comprises:
sorting the maintenance case information according to the similarity to obtain a sorting result;
and determining the target maintenance case information as the query result based on the sequencing result.
8. A service case information inquiry apparatus, comprising:
the information acquisition module is used for acquiring the fault information to be inquired;
the characteristic extraction module is used for extracting the characteristics of the fault information to be inquired to obtain a fault characteristic vector;
the similarity calculation module is used for calculating the similarity between the fault characteristic vector and the case fault characteristic vector corresponding to each piece of maintenance case information;
and the query result determining module is used for selecting target maintenance case information from the maintenance case information as a query result corresponding to the fault information to be queried based on the similarity.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the method for service case information inquiry according to any one of claims 1 to 7 when executing the program.
10. A non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the service case information query method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210217501.XA CN114691680A (en) | 2022-03-07 | 2022-03-07 | Maintenance case information query method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210217501.XA CN114691680A (en) | 2022-03-07 | 2022-03-07 | Maintenance case information query method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114691680A true CN114691680A (en) | 2022-07-01 |
Family
ID=82137188
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210217501.XA Pending CN114691680A (en) | 2022-03-07 | 2022-03-07 | Maintenance case information query method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114691680A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114996536A (en) * | 2022-08-08 | 2022-09-02 | 深圳市信润富联数字科技有限公司 | Maintenance scheme query method, device, equipment and computer readable storage medium |
-
2022
- 2022-03-07 CN CN202210217501.XA patent/CN114691680A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114996536A (en) * | 2022-08-08 | 2022-09-02 | 深圳市信润富联数字科技有限公司 | Maintenance scheme query method, device, equipment and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110472090B (en) | Image retrieval method based on semantic tags, related device and storage medium | |
| CN102567475B (en) | User interface for interactive query reformulation | |
| JP6308708B1 (en) | Patent requirement conformity prediction device and patent requirement conformity prediction program | |
| CN109299245A (en) | The method and apparatus that knowledge point is recalled | |
| CN112765230B (en) | Payment big data analysis method and big data analysis system based on internet finance | |
| CN115114448B (en) | Intelligent multi-mode fusion power consumption inspection method, device, system, equipment and medium | |
| CN114691680A (en) | Maintenance case information query method and device, electronic equipment and storage medium | |
| CN108984777B (en) | Customer service method, apparatus and computer-readable storage medium | |
| CN114049174A (en) | Method and device for commodity recommendation, electronic equipment and storage medium | |
| CN112597292B (en) | Question reply recommendation method, device, computer equipment and storage medium | |
| CN110377803B (en) | Information processing method and device | |
| CN112163098A (en) | Knowledge graph creating method and device, storage medium and server | |
| CN111831286B (en) | User complaint processing method and device | |
| CN111126054A (en) | Method, device, storage medium and electronic equipment for determining similar texts | |
| CN114528851B (en) | Reply sentence determination method, reply sentence determination device, electronic equipment and storage medium | |
| CN114329016B (en) | Picture label generating method and text mapping method | |
| CN111782774A (en) | Question recommendation method and device | |
| CN113705201B (en) | Text-based event probability prediction evaluation algorithm, electronic device and storage medium | |
| CN111611353A (en) | Screening method and device, electronic equipment and computer readable storage medium | |
| CN108334522B (en) | Method for determining customs code, and method and system for determining type information | |
| CN115309865A (en) | Interactive retrieval method, device, equipment and storage medium based on double-tower model | |
| CN114281942A (en) | Question and answer processing method, related equipment and readable storage medium | |
| CN113344674A (en) | Product recommendation method, device, equipment and storage medium based on user purchasing power | |
| CN111476641A (en) | Method, system and storage medium for automatically placing order on mobile device by voice | |
| CN117033719B (en) | Method, device, equipment and medium for acquiring machine tool digital twin system information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |