CN118171080A - Fault level determining method and related device - Google Patents

Abstract

The invention provides a fault grade determining method and a related device, which are used for quickly and accurately determining the fault grade of a work order to be processed by extracting a plurality of items of key information in the work order information after receiving the work order to be processed submitted by a user, respectively carrying out feature coding on the plurality of items of key information to obtain an observation sequence, inputting the observation sequence into a fault grade identification model, automatically analyzing the context connection of the plurality of items of key information and the relation between each item of key information and the fault grade by utilizing the fault grade identification model obtained by training a conditional random field model, thereby avoiding judgment errors caused by manually identifying the fault grade of the work order and improving the work order processing efficiency.

Description

Fault level determining method and related device

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a fault level determining method and a related device.

Background

With the rapid development of informatization, fault management in work order processing systems has become increasingly important. Because the failure level has direct influence on the priority of work order processing, the higher the failure level is, the higher the priority of work order processing is, so the quick and accurate identification of the failure level is particularly important.

However, conventional work order processing systems typically require manual identification and determination of work order failure levels, which are time consuming and difficult to respond to work order processing requirements accurately in a timely manner.

Disclosure of Invention

In view of the above, the invention provides a fault grade determining method and a related device, which can rapidly and accurately determine the fault grade of a work order to be processed, and avoid the judgment error caused by manually identifying the fault grade of the work order, thereby improving the work order processing efficiency.

In order to achieve the above purpose, the specific technical scheme provided by the invention is as follows:

in a first aspect, an embodiment of the present invention provides a fault level determining method, including:

Receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed;

Extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x2, & gt, xn };

Inputting the observation sequence into a fault level recognition model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } and conditional probabilities of each target sequence output by the fault level recognition model, wherein yi represents a fault level corresponding to xi, i is epsilon [1, n ], the fault level recognition model is obtained after training a conditional random field model by utilizing training samples comprising a sample observation sequence and a sample target sequence in advance, and a training target of the conditional random field model is a sample target sequence in the corresponding training sample, wherein the target sequence with the highest conditional probability output by the conditional random field model is the training target of the conditional random field model;

And determining the fault level of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault level identification model.

In some embodiments, the extracting the plurality of pieces of key information in the work order information and performing feature encoding on the plurality of pieces of key information respectively to obtain an observation sequence x= { x1, x 2..x n }, where the extracting includes:

Extracting a plurality of key information items in the work order information according to a pre-configured key information type, wherein the key information type at least comprises: work order submission time, fault equipment information and work order description content keywords;

And respectively converting a plurality of items of key information into numerical type feature vectors by using word embedding, wherein the feature vectors corresponding to the plurality of items of key information form the observation sequence.

In some embodiments, the training method of the fault level identification model includes:

obtaining a training sample;

training the conditional random field model by using a training sample, wherein the expression of the conditional random field model is as follows:

P(y|x)＝exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)]])/Z(x)；

Wherein P (y|x) represents the conditional probability of the target sequence y corresponding to the sample observation sequence x;

Z(x)＝sum(over y)[exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)])]；

λk and μk' are model parameters to be learned;

tk and sk are feature functions.

In some embodiments, further comprising:

setting an objective function log P (y|x);

In the process of training the conditional random field model by using the training sample, when the log P (y|x) is maximized or the log P (y|x) is minimized, the training is finished, and the fault level identification model is obtained.

In some embodiments, further comprising:

and adding regularization term into the objective function.

In some embodiments, the determining the fault level of the work order to be processed according to the plurality of target sequences output by the fault level identification model and the conditional probability of each target sequence includes:

determining the most probable target sequence by using a Viterbi algorithm according to a plurality of target sequences output by the fault level identification model and the conditional probability of each target sequence;

and determining the fault level of the work order to be processed according to the most probable target sequence.

In a second aspect, an embodiment of the present invention provides a fault level determining apparatus, including:

The work order receiving unit is used for receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed;

The feature extraction unit is used for extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x 2;

The model processing unit is used for inputting the observation sequence into a fault level recognition model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } output by the fault level recognition model and the conditional probability of each target sequence, yi represents the fault level corresponding to xi, i is epsilon [1, n ], the fault level recognition model is obtained after training a conditional random field model by utilizing training samples comprising sample observation sequences and sample target sequences in advance, and the target sequence with the highest conditional probability output by the conditional random field model is the sample target sequence in the corresponding training sample;

and the fault grade determining unit is used for determining the fault grade of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault grade identification model.

In some embodiments, the feature extraction unit is specifically configured to extract a plurality of pieces of key information in the work order information according to a pre-configured key information type, where the key information type at least includes: work order submission time, fault equipment information and work order description content keywords; and respectively converting a plurality of items of key information into numerical type feature vectors by using word embedding, wherein the feature vectors corresponding to the plurality of items of key information form the observation sequence.

In a third aspect, an embodiment of the present invention provides an electronic device, including a processor and a memory;

the memory is used for storing program codes and transmitting the program codes to the processor;

The processor is configured to perform a fault level determination method as described in any one of the implementations of the first aspect according to instructions in the program code.

In a fourth aspect, an embodiment of the present invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a fault level determination method as described in any one of the implementations of the first aspect.

Compared with the prior art, the invention has the following beneficial effects:

After receiving a work order to be processed submitted by a user, the method and the related device for determining the fault level disclosed by the invention are used for obtaining an observation sequence by extracting a plurality of items of key information in the work order information and respectively carrying out feature coding on the plurality of items of key information, inputting the observation sequence into a fault level recognition model, automatically analyzing the context connection of the plurality of items of key information and the relation between each item of key information and the fault level by utilizing the fault level recognition model obtained by training a conditional random field model, thereby rapidly and accurately determining the fault level of the work order to be processed, avoiding judgment errors caused by manually recognizing the fault level of the work order, and further improving the work order processing efficiency.

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flow chart of a fault level determining method according to an embodiment of the present invention;

Fig. 2 is a schematic structural diagram of a fault level determining apparatus according to an embodiment of the present invention;

Fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

The term "including" and variations thereof as used herein are intended to be open-ended, i.e., including, but not limited to. The term "based on" is based at least in part on. The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments. Related definitions of other terms will be given in the description below.

It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.

In order to more clearly clarify the technical scheme of the present application, the related concepts related to the present application are explained below.

CRF: conditional Random Fields, conditional random field model.

The invention provides a fault level determining method which can be realized through a computer program, such as a computer program installed in a work order system, wherein the work order system can be deployed in electronic equipment such as a server, a server cluster and the like.

Fig. 1 is a flow chart of a fault level determining method according to an embodiment of the present invention, which specifically includes the following steps:

s101: receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed;

Preprocessing the work order to be processed comprises data cleaning, repairing missing data, converting the data in the work order to be processed into a preset data format and the like, so that subsequent processing is facilitated.

The repair of the missing data specifically includes supplementing default values corresponding to the type of the missing data to the missing data.

The preset data format may be a text format.

S102: extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x2, & gt, xn };

specifically, firstly, extracting a plurality of key information items in work order information according to a pre-configured key information type, wherein the key information type at least comprises work order submitting time, fault equipment information and work order description content keywords, the work order submitting time represents fault occurrence time, the fault equipment information comprises information such as equipment type and the like, and the work order description content keywords comprise keywords such as fault type, fault code and instructions and the like.

And then converting the multiple key information into feature vectors, wherein the feature vectors corresponding to the multiple key information form an observation sequence, such as feature vectors which respectively convert the multiple key information into numerical values by using word embedding.

And converting the extracted multiple key information into an observation sequence suitable for model input through feature coding.

The work order to be processed is, for example, "the system suddenly crashes when executing instructions, an ERROR code ERROR404 is displayed", and the key information obtained through data preprocessing and key information extraction includes: "instruction", "crash", "ERROR code", "ERROR404", etc., and then converts this critical information into feature vectors.

The observation sequence x= { x1, x2,..xi in xn } represents a feature vector corresponding to the i-th item of key information, i e [1, n ].

S103: inputting the observation sequence into a fault level identification model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } and conditional probability of each target sequence, which are output by the fault level identification model;

The fault level recognition model is obtained by training the conditional random field model by utilizing a training sample containing a sample observation sequence and a sample target sequence in advance, wherein the training target of the conditional random field model is the sample target sequence in the corresponding training sample, and the target sequence with the highest conditional probability output by the conditional random field model is the training target of the conditional random field model.

Yi represents the failure level corresponding to xi, i e [1, n ], including, in failure level: three classes of low (L), medium (M), high (H) are exemplified, yi represents one of (L, M, H).

The conditional probability of the target sequence is denoted as P (y|x), and the conditional probability of yi in the target sequence is P (yi|xi).

The fault level recognition model will output all possible target sequences and the conditional probability for each target sequence.

For convenience of explanation, taking the above work order to be processed as an example, the observation sequence x= { x1, x2}, x1 represents a feature vector corresponding to the key information "ERROR 404", x2 represents a feature vector corresponding to the key information "crash", and an example of the conditional probability output by the failure level recognition model is as follows:

y1 represents that when the failure level corresponding to x1 is low, P (y1|x1) =0.1;

y1 represents that when the failure level corresponding to x1 is middle, P (y1|x1) =0.2;

y1 represents that when the failure level corresponding to x1 is high, P (y1|x1) =0.7;

y2 represents that when the fault level corresponding to x2 is low, P (y2|x2) =0.2;

y2 represents that when the fault level corresponding to x2 is middle, P (y2|x2) =0.3;

y2 indicates that when the fault level corresponding to x2 is high, P (y2|x2) =0.5.

S104: and determining the fault level of the work order to be processed according to the plurality of target sequences output by the fault level identification model and the conditional probability of each target sequence.

The method includes the steps of determining a target sequence with the largest probability according to a plurality of target sequences output by a fault level recognition model and the conditional probability of each target sequence, and determining the fault level of a work order to be processed according to the target sequence with the largest probability, for example, determining the fault level with the highest occurrence frequency in the target sequence with the largest probability as the fault level of the work order to be processed.

For example, since the number of feature vectors included in the observation sequence is large in practical application, the number of target sequences is usually exponential, the feasibility of direct calculation is weak, the most probable target sequence is effectively found by using a Viterbi (Viterbi) algorithm according to a plurality of target sequences output by the fault level recognition model and the conditional probability of each target sequence, and the fault level of the work order to be processed is determined according to the most probable target sequence.

The method of determining the most probable target sequence using the Viterbi algorithm is described below by way of a specific example.

Assume that the fault level, i.e., the observed state, includes: the feature vectors after feature coding only consider two dimensions of ERROR codes and behaviors, namely an observation sequence x= { x1, x2}, wherein x1 represents the feature vector corresponding to key information "ERROR404", and x2 represents the feature vector corresponding to key information "crash".

The assumed state transition probabilities are:

From low to low 0.6, from low to medium 0.2, from low to high 0.2;

from middle to low 0.3, from middle to middle 0.4, from middle to high 0.3;

from high to low 0.3, from high to medium 0.3, and from high to high 0.4.

The conditional probability of the output of the assumed fault level recognition model is:

y1 represents that when the failure level corresponding to x1 is low, P (y1|x1) =0.1;

y1 represents that when the failure level corresponding to x1 is middle, P (y1|x1) =0.2;

y1 represents that when the failure level corresponding to x1 is high, P (y1|x1) =0.7;

y2 represents that when the fault level corresponding to x2 is low, P (y2|x2) =0.2;

y2 represents that when the fault level corresponding to x2 is middle, P (y2|x2) =0.3;

y2 indicates that when the fault level corresponding to x2 is high, P (y2|x2) =0.5.

First, an initial state distribution probability is initialized, which is typically based on pre-training or a priori knowledge. The temporal assumption is equal probability, i.e. low, medium, high each 1/3. Thus, the initialization score under the first item of key information (ERROR 404) is calculated as follows:

V_1(L)＝1/3*0.1＝0.0333；

V_1(M)＝1/3*0.2＝0.0666；

V_1(H)＝1/3*0.7＝0.2333；

Next, for each state (L, M, H), the probability that this state is observed is calculated multiplied by the maximum probability of transitioning from the previous state to the current state. Taking as an example the calculation of the score for the transition of the second item of critical information (crash) from low to high level:

V_2(H)＝max[V_1(L)*0.2,V_1(M)*0.3,V_1(H)*0.4]*0.5；

V1 (L), V1 (M), V1 (H) was replaced with the value of the previous step:

V_2(H)＝max[0.0333*0.2,0.0666*0.3,0.2333*0.4]*0.5；

V_2(H)＝max[0.00666,0.01998,0.09332]*0.5＝0.09332*0.5＝0.04666。

Similarly, all V_2 (L) and V_2 (M) values are calculated.

Finally, the state with the highest score obtained under the last observation is selected as the last state, and the initial state is reversely tracked according to the stored path information, so that the most probable state sequence is obtained. In this example, if the final V_2 (H) value is maximum, it can be concluded that the most likely state sequence ends at high (H). Then trace back to find the optimal path to the high (H) state, assuming a transition from the M state, so that the sequence we get may be "medium- > high".

This simplified example illustrates the basic logic and computation of the Viterbi algorithm to find the most likely state sequence for a given observation sequence. In practical complex applications, there are more states and observations, and detailed parameter learning processes.

Further, the failure level of the work order to be processed determined in any of the above implementations is represented by a failure identifier, for example, L represents low, M represents high, and H represents high, so that in order to display the failure level of the final work order to be processed in a clear and readable form, the result represented by the failure identifier needs to be decoded back to a specific failure level, such as low, medium, and high.

According to the fault grade determining method disclosed by the embodiment, after receiving the to-be-processed work order submitted by the user, the fault grade of the to-be-processed work order is determined rapidly and accurately by extracting a plurality of pieces of key information in the work order information and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence, inputting the observation sequence into a fault grade identification model, automatically analyzing the context connection of the plurality of pieces of key information and the relation between each piece of key information and the fault grade by utilizing the fault grade identification model obtained by training a conditional random field model, and judging errors caused by manually identifying the fault grade of the work order are avoided, so that the work order processing efficiency is improved.

The following describes a training method of the failure level recognition model.

Obtaining a training sample;

training the conditional random field model by using a training sample, wherein the expression of the conditional random field model is as follows:

P(y|x)＝exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)]])/Z(x)；

Wherein P (y|x) represents the conditional probability of the target sequence y corresponding to the sample observation sequence x;

z (x) =sum (over y) [ exp (sum (over i) [ sum (over k) [ λk×tk (yi-1, yi, x, i) ]+sum (over k ') [ μk' ×sk (yi, x, i) ] ], represents a normalization factor that ensures that the sum of all possible target sequence probabilities is 1.

Each term within exp represents a feature function, i.e. tk and sk are feature functions, e.g. the feature function tk relates two adjacent target values yi-1 and yi to the whole observation sequence, whereas the feature function sk relates only the target value yi to the observation sequence.

Λk and μk' are model parameters to be learned, represent weights of feature functions, and define how much the corresponding feature function values play in decision making.

Taking the observation sequence x= { x1, x2}, x1 represents the feature vector corresponding to the key information "ERROR404", x2 represents the feature vector corresponding to the key information "crash", as an example, the feature function tk represents the association between "ERROR404" and a high failure level, i.e. when the ERROR code is "ERROR404", the work order failure level may be high. As another example, the association between the existence of a "crash" and a high failure level is indicated for the feature function sk, i.e., the work order failure level may be high when the term "crash" is included in the work order content.

Further, model parameters λk and μk' can be learned by optimization algorithms such as maximum likelihood estimation or gradient descent, for example, an objective function log P (y|x) is set, and in the process of training the conditional random field model by using the training sample, training is ended when log P (y|x) is maximized or log P (y|x) is minimized, so that a fault level identification model is obtained.

Further, to ensure the generalization performance of the model, a regularization term L1 or L2 may be added to the objective function, and for the L1 regularization term, λ_1Σ|w| is added to the objective function, where w represents a model parameter, and λ_1 is a regularized strength parameter. For the L2 regularization term, λ_2Σw2 is added to the objective function.

The regularization term is added to the objective function to prevent the model from overfitting training data and improve the generalization capability of the model on unseen data. It does this by penalizing excessive model parameter values, which can help keep model parameters small, thereby alleviating model complexity.

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

Although operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. In certain circumstances, multitasking and parallel processing may be advantageous.

It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.

Computer program code for carrying out operations of the present disclosure may be written in one or more programming languages, including, but not limited to, an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

Based on the method for determining a fault level disclosed in the foregoing embodiment, this embodiment correspondingly discloses a device for determining a fault level, please refer to fig. 2, which includes:

the work order receiving unit 201 is configured to receive a work order to be processed submitted by a user, and perform pretreatment on the work order to be processed to obtain work order information of the work order to be processed;

the feature extraction unit 202 is configured to extract a plurality of pieces of key information in the work order information, and perform feature encoding on the plurality of pieces of key information respectively to obtain an observation sequence x= { x1, x 2;

The model processing unit 203 is configured to input the observation sequence into a fault level recognition model, obtain a plurality of target sequences y= { y1, y2,.. the fault level recognition model is obtained by training a conditional random field model by utilizing a training sample comprising a sample observation sequence and a sample target sequence in advance, wherein the target sequence with the highest conditional probability output by the conditional random field model is the training target of the conditional random field model, and the training target is the sample target sequence in the corresponding training sample;

and the fault level determining unit 204 is configured to determine a fault level of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault level recognition model.

In some embodiments, the feature extraction unit 202 is specifically configured to extract a plurality of pieces of key information in the work order information according to a pre-configured key information type, where the key information type includes at least: work order submission time, fault equipment information and work order description content keywords; and respectively converting a plurality of items of key information into numerical type feature vectors by using word embedding, wherein the feature vectors corresponding to the plurality of items of key information form the observation sequence.

In some embodiments, further comprising:

The model training unit is used for acquiring training samples, training the conditional random field model by using the training samples, and the expression of the conditional random field model is as follows:

P(y|x)＝exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)]])/Z(x)；

Wherein P (y|x) represents the conditional probability of the target sequence y corresponding to the sample observation sequence x;

Z(x)＝sum(over y)[exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)])]；

λk and μk' are model parameters to be learned;

tk and sk are feature functions.

In some embodiments, the model training unit is further configured to set an objective function log P (y|x), and when the log P (y|x) is maximized or the log P (y|x) is minimized during training of the conditional random field model by using the training sample, the training is ended, so as to obtain the fault level identification model.

In some embodiments, the fault level determining unit is specifically configured to determine a most probable target sequence using a viterbi algorithm according to the plurality of target sequences and the conditional probability of each of the target sequences output by the fault level recognition model; and determining the fault level of the work order to be processed according to the most probable target sequence.

According to the fault grade determining device disclosed by the embodiment, after receiving a work order to be processed submitted by a user, a plurality of pieces of key information in the work order information are extracted, and feature codes are respectively carried out on the plurality of pieces of key information to obtain an observation sequence, the observation sequence is input into a fault grade identification model, the context connection of the plurality of pieces of key information and the relation between each piece of key information and the fault grade are automatically analyzed by using the fault grade identification model obtained through training of a conditional random field model, so that the fault grade of the work order to be processed is rapidly and accurately determined, judgment errors caused by manually identifying the fault grade of the work order are avoided, and the work order processing efficiency is improved.

The units involved in the embodiments of the present disclosure may be implemented by means of software, or may be implemented by means of hardware. Wherein the names of the units do not constitute a limitation of the units themselves in some cases.

The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.

Referring now to fig. 3, a schematic diagram of an electronic device suitable for use in implementing embodiments of the present disclosure is shown. The terminal devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and the like, and stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 3 is merely an example and should not be construed to limit the functionality and scope of use of the disclosed embodiments.

As shown in fig. 3, the electronic device may include a processing means (e.g., a central processor, a graphics processor, etc.) 301 that may perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 302 or a program loaded from a storage means 308 into a Random Access Memory (RAM) 303. In the RAM303, various programs and data required for the operation of the electronic device are also stored. The processing device 301, the ROM 302, and the RAM303 are connected to each other via a bus 304. An input/output (I/O) interface 305 is also connected to bus 304.

In general, the following devices may be connected to the I/O interface 305: input devices 306 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 307 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 308 including, for example, magnetic tape, hard disk, etc.; and communication means 309. The communication means 309 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. While fig. 3 shows an electronic device having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead.

The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed; extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x2, & gt, xn }; inputting the observation sequence into a fault level recognition model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } and conditional probabilities of each target sequence output by the fault level recognition model, wherein yi represents a fault level corresponding to xi, i is epsilon [1, n ], the fault level recognition model is obtained after training a conditional random field model by utilizing training samples comprising a sample observation sequence and a sample target sequence in advance, and a training target of the conditional random field model is a sample target sequence in the corresponding training sample, wherein the target sequence with the highest conditional probability output by the conditional random field model is the training target of the conditional random field model; and determining the fault level of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault level identification model.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

It should be noted that the computer readable medium described in the present disclosure may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present disclosure, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device.

In particular, according to embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a non-transitory computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via a communication device 309, or installed from a storage device 308, or installed from a ROM 302. The above-described functions defined in the methods of the embodiments of the present disclosure are performed when the computer program is executed by the processing means 301.

According to one or more embodiments of the present disclosure, there is provided a fault level determination method [ example 1 ], including:

Receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed;

Extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x2, & gt, xn };

Inputting the observation sequence into a fault level recognition model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } and conditional probabilities of each target sequence output by the fault level recognition model, wherein yi represents a fault level corresponding to xi, i is epsilon [1, n ], the fault level recognition model is obtained after training a conditional random field model by utilizing training samples comprising a sample observation sequence and a sample target sequence in advance, and a training target of the conditional random field model is a sample target sequence in the corresponding training sample, wherein the target sequence with the highest conditional probability output by the conditional random field model is the training target of the conditional random field model;

And determining the fault level of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault level identification model.

According to one or more embodiments of the present disclosure, there is provided a method of example 1 [ example 2 ], the extracting a plurality of pieces of key information in the work order information, and performing feature encoding on the plurality of pieces of key information respectively, to obtain an observation sequence x= { x1, x2,..:

Extracting a plurality of key information items in the work order information according to a pre-configured key information type, wherein the key information type at least comprises: work order submission time, fault equipment information and work order description content keywords;

And respectively converting a plurality of items of key information into numerical type feature vectors by using word embedding, wherein the feature vectors corresponding to the plurality of items of key information form the observation sequence.

According to one or more embodiments of the present disclosure, there is provided a method of example 1, the training method of the failure level recognition model, comprising:

obtaining a training sample;

training the conditional random field model by using a training sample, wherein the expression of the conditional random field model is as follows:

P(y|x)＝exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)]])/Z(x)；

Wherein P (y|x) represents the conditional probability of the target sequence y corresponding to the sample observation sequence x;

Z(x)＝sum(over y)[exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)])]；

λk and μk' are model parameters to be learned;

tk and sk are feature functions.

According to one or more embodiments of the present disclosure, there is provided a method of example 1 [ example 4], further comprising:

setting an objective function log P (y|x);

In the process of training the conditional random field model by using the training sample, when the log P (y|x) is maximized or the log P (y|x) is minimized, the training is finished, and the fault level identification model is obtained.

According to one or more embodiments of the present disclosure, there is provided a method of example 1 [ example 5 ], further comprising:

and adding regularization term into the objective function.

According to one or more embodiments of the present disclosure, there is provided a method of example 1, the determining a failure level of the work order to be processed according to a plurality of the target sequences output by the failure level recognition model and a conditional probability of each of the target sequences, including:

determining the most probable target sequence by using a Viterbi algorithm according to a plurality of target sequences output by the fault level identification model and the conditional probability of each target sequence;

and determining the fault level of the work order to be processed according to the most probable target sequence.

According to one or more embodiments of the present disclosure, there is provided a fault level determining apparatus [ example 7 ], including:

The work order receiving unit is used for receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed;

The feature extraction unit is used for extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x 2;

The model processing unit is used for inputting the observation sequence into a fault level recognition model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } output by the fault level recognition model and the conditional probability of each target sequence, yi represents the fault level corresponding to xi, i is epsilon [1, n ], the fault level recognition model is obtained after training a conditional random field model by utilizing training samples comprising sample observation sequences and sample target sequences in advance, and the target sequence with the highest conditional probability output by the conditional random field model is the sample target sequence in the corresponding training sample;

and the fault grade determining unit is used for determining the fault grade of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault grade identification model.

According to one or more embodiments of the present disclosure, there is provided the apparatus of example 7 [ example 8 ], wherein the feature extraction unit is specifically configured to extract a plurality of pieces of key information in the work order information according to a pre-configured key information type, where the key information type includes at least: work order submission time, fault equipment information and work order description content keywords; and respectively converting a plurality of items of key information into numerical type feature vectors by using word embedding, wherein the feature vectors corresponding to the plurality of items of key information form the observation sequence.

Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are example forms of implementing the claims.

While several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.

The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by persons skilled in the art that the scope of the disclosure referred to in this disclosure is not limited to the specific combinations of features described above, but also covers other embodiments which may be formed by any combination of features described above or equivalents thereof without departing from the spirit of the disclosure. Such as those described above, are mutually substituted with the technical features having similar functions disclosed in the present disclosure (but not limited thereto).

Claims (10)

1. A fault level determination method, comprising:

Receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed;

Extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x2, & gt, xn };

Inputting the observation sequence into a fault level recognition model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } and conditional probabilities of each target sequence output by the fault level recognition model, wherein yi represents a fault level corresponding to xi, i is epsilon [1, n ], the fault level recognition model is obtained after training a conditional random field model by utilizing training samples comprising a sample observation sequence and a sample target sequence in advance, and a training target of the conditional random field model is a sample target sequence in the corresponding training sample, wherein the target sequence with the highest conditional probability output by the conditional random field model is the training target of the conditional random field model;

And determining the fault level of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault level identification model.

2. The fault level determining method according to claim 1, wherein the extracting the plurality of pieces of key information in the work order information and performing feature encoding on the plurality of pieces of key information respectively to obtain an observation sequence x= { x1, x2, xn }, includes:

Extracting a plurality of key information items in the work order information according to a pre-configured key information type, wherein the key information type at least comprises: work order submission time, fault equipment information and work order description content keywords;

And respectively converting a plurality of items of key information into numerical type feature vectors by using word embedding, wherein the feature vectors corresponding to the plurality of items of key information form the observation sequence.

3. The failure level determination method according to claim 1, wherein the training method of the failure level recognition model includes:

obtaining a training sample;

training the conditional random field model by using a training sample, wherein the expression of the conditional random field model is as follows:

P(y|x)＝exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)]])/Z(x)；

Wherein P (y|x) represents the conditional probability of the target sequence y corresponding to the sample observation sequence x;

Z(x)＝sum(over y)[exp(sum(over i)[sum(over k)[λk*tk(yi-1,yi,x,i)]+sum(over k')[μk'*sk(yi,x,i)])]；

λk and μk' are model parameters to be learned;

tk and sk are feature functions.

4. The fault level determination method of claim 3, further comprising:

setting an objective function log P (y|x);

In the process of training the conditional random field model by using the training sample, when the log P (y|x) is maximized or the log P (y|x) is minimized, the training is finished, and the fault level identification model is obtained.

5. The fault level determination method of claim 4, further comprising:

and adding regularization term into the objective function.

6. The method according to claim 1, wherein determining the failure level of the work order to be processed based on the plurality of target sequences output by the failure level identification model and the conditional probability of each of the target sequences, comprises:

determining the most probable target sequence by using a Viterbi algorithm according to a plurality of target sequences output by the fault level identification model and the conditional probability of each target sequence;

and determining the fault level of the work order to be processed according to the most probable target sequence.

7. A failure level determination apparatus, comprising:

The work order receiving unit is used for receiving a work order to be processed submitted by a user, and preprocessing the work order to be processed to obtain work order information of the work order to be processed;

The feature extraction unit is used for extracting a plurality of pieces of key information in the work order information, and respectively carrying out feature coding on the plurality of pieces of key information to obtain an observation sequence x= { x1, x 2;

The model processing unit is used for inputting the observation sequence into a fault level recognition model to obtain a plurality of target sequences y= { y1, y2, & gt, yn } output by the fault level recognition model and the conditional probability of each target sequence, yi represents the fault level corresponding to xi, i is epsilon [1, n ], the fault level recognition model is obtained after training a conditional random field model by utilizing training samples comprising sample observation sequences and sample target sequences in advance, and the target sequence with the highest conditional probability output by the conditional random field model is the sample target sequence in the corresponding training sample;

and the fault grade determining unit is used for determining the fault grade of the work order to be processed according to the plurality of target sequences and the conditional probability of each target sequence output by the fault grade identification model.

8. The fault level determination apparatus according to claim 7, wherein the feature extraction unit is specifically configured to extract a plurality of pieces of key information in the work order information according to a pre-configured key information type, where the key information type includes at least: work order submission time, fault equipment information and work order description content keywords; and respectively converting a plurality of items of key information into numerical type feature vectors by using word embedding, wherein the feature vectors corresponding to the plurality of items of key information form the observation sequence.

9. An electronic device comprising a processor and a memory;

the memory is used for storing program codes and transmitting the program codes to the processor;

The processor is configured to perform a fault level determination method as claimed in any one of claims 1-6 in accordance with instructions in the program code.

10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements a fault level determination method according to any of claims 1-6.