CN112612891A - Training method of emergency disposal model, emergency disposal method and device - Google Patents

Training method of emergency disposal model, emergency disposal method and device Download PDF

Info

Publication number
CN112612891A
CN112612891A CN202011588726.3A CN202011588726A CN112612891A CN 112612891 A CN112612891 A CN 112612891A CN 202011588726 A CN202011588726 A CN 202011588726A CN 112612891 A CN112612891 A CN 112612891A
Authority
CN
China
Prior art keywords
alarm
phrase
alarm type
emergency disposal
model
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
Application number
CN202011588726.3A
Other languages
Chinese (zh)
Inventor
王哲
孙志斌
王灿
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Agricultural Bank of China
Original Assignee
Agricultural Bank of China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Agricultural Bank of China filed Critical Agricultural Bank of China
Priority to CN202011588726.3A priority Critical patent/CN112612891A/en
Publication of CN112612891A publication Critical patent/CN112612891A/en
Pending legal-status Critical Current

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
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G06F16/353—Clustering; Classification into predefined classes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
    • G06F11/079—Root cause analysis, i.e. error or fault diagnosis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
    • G06F11/0793—Remedial or corrective actions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G06F16/355—Class or cluster creation or modification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • 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
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Quality & Reliability (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Alarm Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The embodiment of the invention discloses a training method of an emergency disposal model, which comprises the following steps: the preset model is trained through the alarm information comprising various alarm types, the alarm grade corresponding to the alarm information of each alarm type and the training data set of the emergency disposal scheme, and the obtained trained model can be used for predicting the alarm grade and the emergency disposal scheme of the alarm type. Therefore, in practical application, the received alarm information can be predicted through the trained model, so that the alarm level and the emergency disposal scheme can be quickly obtained, the manual troubleshooting of operation and maintenance personnel is avoided, the problems of overlong disposal chain and low disposal efficiency are solved, and the fault treatment efficiency is improved.

Description

Training method of emergency disposal model, emergency disposal method and device
Technical Field
The invention relates to the field of emergency disposal, in particular to an emergency disposal method and device.
Background
Various problems occur in the operation process of the commercial banking system, and in order to guarantee the continuity of the business, emergency treatment needs to be carried out on the problems. However, commercial bank systems are numerous, and after receiving a system alarm, first-line emergency operators on duty need to notify relevant operation and maintenance personnel layer by layer to confirm due to unclear severity and disposal method of the alarm, and after receiving the alarm information, the system operation and maintenance personnel need to manually check according to alarm content to finally determine an emergency disposal scheme. In this case, the treatment chain is too long, which leads to low treatment efficiency.
Disclosure of Invention
The embodiment of the invention discloses a training method of an emergency disposal model, which solves the problems of overlong disposal chain and low disposal efficiency in the prior art, and predicts alarm information through the trained emergency disposal model, so that the alarm grade and the emergency disposal scheme of the alarm information are rapidly obtained, and the disposal efficiency is improved.
In a first aspect:
an emergency treatment model training method, comprising:
acquiring a training data set; the training data set comprises alarm information of a plurality of alarm types, alarm grades corresponding to the alarm information of each alarm type and an emergency disposal scheme;
performing word segmentation processing on the alarm information of each alarm type;
obtaining effective phrases corresponding to each alarm type from the word segmentation result, and carrying out vectorization processing on the effective phrases;
and taking the vector of the effective phrase corresponding to each alarm type as input, taking the alarm grade corresponding to each alarm type and the emergency disposal scheme as output, and training a preset model to obtain a trained emergency disposal model.
Optionally, the obtaining an effective phrase corresponding to each alarm type from the word segmentation result includes:
aiming at any phrase, calculating the frequency of the phrase in the alarm information to obtain the word frequency TF of the phrase, calculating the inverse text frequency index IDF of the phrase, and calculating the importance of the phrase through the word frequency TF of the phrase and the inverse text frequency index IDF;
aiming at any one alarm type, screening out effective phrases of the alarm type according to the importance degree of each phrase in the alarm type.
Optionally, the method further includes:
determining the data volume contained in the data set corresponding to each alarm type;
determining whether the data sets of each alarm type are balanced or not according to the data volume contained in the data set corresponding to each alarm type;
if the data sets of each alarm type are not balanced, determining the data set corresponding to the alarm type to be adjusted;
and adjusting the alarm information contained in the data set corresponding to the alarm type to be adjusted.
Optionally, the adjusting the alarm information included in the data set corresponding to the alarm type to be adjusted includes:
collecting alarm information of the alarm type to be adjusted;
and supplementing the acquired alarm information into a data set corresponding to the alarm type to be adjusted.
Optionally, the method further includes:
acquiring a test set; the test set comprises alarm information of various alarm types and alarms corresponding to the alarm types;
inputting the alarm information in the test set into a trained emergency disposal model, and outputting a prediction result containing an alarm grade and an emergency disposal scheme;
determining a real processing strategy corresponding to the test set and comprising an alarm level and an emergency treatment scheme;
evaluating the emergency treatment model by comparing the actual treatment strategy with the predicted result.
In a second aspect:
the embodiment of the invention discloses an emergency disposal method, which comprises the following steps:
after receiving the warning information, performing word segmentation processing on the warning information;
inputting the word segmentation result into a preset emergency disposal model to obtain an alarm grade and an emergency disposal strategy corresponding to the alarm information; the emergency treatment model is obtained by the training method of any one of the preceding claims 1 to 5.
In a third aspect:
the embodiment of the invention discloses a training device of an emergency disposal model, which comprises:
an acquisition unit for acquiring a training data set; the training data set comprises alarm information of a plurality of alarm types, alarm grades corresponding to the alarm information of each alarm type and an emergency disposal scheme;
the first word segmentation processing unit is used for carrying out word segmentation processing on the alarm information of each alarm type;
the phrase processing unit is used for acquiring effective phrases corresponding to each alarm type from the word segmentation result and carrying out vectorization processing on the effective phrases;
and the training unit is used for taking the vector of the effective phrase corresponding to each alarm type as input, taking the alarm grade corresponding to each alarm type and the emergency disposal scheme as output, and training a preset model to obtain a trained emergency disposal model.
Optionally, the phrase processing unit includes:
the importance degree calculation operator unit is used for calculating the frequency of the phrase in the alarm information aiming at any phrase to obtain the word frequency TF of the phrase, calculating the inverse text frequency index IDF of the phrase and calculating the importance degree of the phrase through the word frequency TF of the phrase and the inverse text frequency index IDF;
and the screening subunit is used for screening the effective phrases of the alarm type according to the importance of each phrase in the alarm type aiming at any one alarm type.
In a fourth aspect:
the embodiment of the invention discloses an emergency disposal device, which comprises:
the second word segmentation processing unit is used for performing word segmentation processing on the alarm information after the alarm information is received;
the prediction unit is used for inputting the word segmentation result into a preset emergency disposal model to obtain an alarm grade and an emergency disposal strategy corresponding to the alarm information; the emergency treatment model is obtained by the training method according to the first aspect.
In a fifth aspect:
the embodiment of the invention discloses an electronic device, which comprises:
a memory and a processor;
the memory is used for storing programs;
the processor is configured to execute the emergency treatment model training method of any one of claims 1-5 when executing the program stored in the memory.
The embodiment of the invention discloses a training method of an emergency disposal model, which comprises the following steps: the preset model is trained through the alarm information comprising various alarm types, the alarm grade corresponding to the alarm information of each alarm type and the training data set of the emergency disposal scheme, and the obtained trained model can be used for predicting the alarm grade and the emergency disposal scheme of the alarm type. Therefore, in practical application, the received alarm information can be predicted through the trained model, so that the alarm level and the emergency disposal scheme can be quickly obtained, the manual troubleshooting of operation and maintenance personnel is avoided, the problems of overlong disposal chain and low disposal efficiency are solved, and the fault treatment efficiency is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a schematic flow chart illustrating an emergency handling method according to an embodiment of the present invention;
fig. 2 is a flowchart illustrating a method for obtaining valid phrases according to an embodiment of the present invention;
fig. 3 is a schematic flow chart of an emergency handling method according to an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of a training apparatus for an emergency treatment model according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram illustrating an emergency disposal device according to an embodiment of the present invention;
fig. 6 shows a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. 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.
Referring to fig. 1, a schematic flow chart of an emergency disposal method provided in an embodiment of the present invention is shown, where the method includes:
s101: acquiring a training data set;
in this embodiment, the training data set includes alarm information of multiple alarm types, an alarm level corresponding to each type of alarm information, and an emergency disposal scheme.
The actual application process may include multiple alarm types, each of which may include multiple alarm modes, and each of which is represented by an alarm information mode.
In this embodiment, when the model is trained by using the alarm information of different alarm types, in order to ensure that the trained emergency treatment model does not generate a deviation for prediction of each alarm type, it is necessary to train a sample to ensure that the types of the alarm information of each alarm type are in balance, and the method further includes:
determining the data volume contained in the data set corresponding to each alarm type;
determining whether the data sets of each alarm type are balanced or not according to the data volume contained in the data set corresponding to each alarm type;
if the data sets of each alarm type are not balanced, determining the data set corresponding to the alarm type to be adjusted;
and performing data adjustment on the data set corresponding to the alarm type to be adjusted.
In this embodiment, whether the data sets of each alarm type are balanced may be determined by whether the data amounts contained in the data sets are balanced, for example, by the following method:
calculating the difference value of the data quantity in each two data sets;
if all the difference values are smaller than the preset threshold value, the data sets of each alarm type are balanced;
and if the difference value is larger than the preset threshold value, determining that the data sets of the alarm types are unbalanced.
Or whether the data sets of each alarm type are in a balanced state or not can be detected by a data exploration method.
In case of an imbalance between the data sets of each alarm type, data sets with a larger amount of data may be deleted, and collections with a smaller amount of data may be supplemented.
In practical applications, in order to improve the accuracy of model training, a data supplementing manner is usually adopted, and specifically, the method further includes:
collecting alarm information of the alarm type to be adjusted;
and supplementing the acquired alarm information into a data set corresponding to the alarm type to be adjusted.
In addition, in order to improve the quality of the training samples, the training samples may be preprocessed, for example, including: the abnormal or missing data is discarded.
S102: performing word segmentation processing on the alarm information of each alarm type;
in this embodiment, the word segmentation processing method includes many methods, and is not limited in this embodiment.
S103: obtaining effective phrases corresponding to each alarm type from word segmentation results, and carrying out vectorization processing on the effective phrases;
in this embodiment, the valid phrase may be expressed as a phrase that is more important for indicating the type of alarm.
After the word segmentation processing is performed on the alarm information, some words which have a high frequency of occurrence but have no meaning to indicate the alarm type, such as "what", and the like, are obtained, and therefore, these meaningless words need to be removed, and a word with a high degree of importance to indicate the alarm type needs to be found.
In this embodiment, the valid phrases corresponding to each alarm type may be obtained in multiple ways, which is not limited in this embodiment, and preferably, the following method may be adopted:
data sets for any one alarm type:
calculating the importance of each phrase in the data set;
sequencing all phrases in the data set according to the importance of each phrase;
and screening out the phrases with the importance degree ranking before a preset threshold value.
The importance of the phrases may be determined by calculating a TF-IDF (term frequency-inverse document frequency, chinese full name) of each phrase, and a specific implementation process will be described in detail hereinafter, which is not repeated in this embodiment.
After the valid phrases are obtained, the valid phrases can be vectorized for training the model.
The effective phrase vectorization method includes various methods, which are not limited in this embodiment, and preferably, a Word2vec model may be used to perform vectorization on the phrase.
S104: and taking the vector of the effective phrase corresponding to each alarm type as input, taking the alarm grade corresponding to each alarm type and the emergency disposal scheme as output, and training a preset model to obtain a trained emergency disposal model.
In this embodiment, the model used for training may be any one of machine learning models, or any one of convolutional neural network models, or a combination model (i.e., a combination model of multiple machine learning models, or a combination model of different kinds of models).
Preferably, an LSTM (Long Shprt-Term Memory, Chinese full name: Long short Term Memory network) model can be adopted.
In this embodiment, during training, the vector of the effective phrase corresponding to each alarm type is used as input, and the alarm level and the emergency disposal scheme corresponding to each alarm type are used as output. Firstly, inputting a vector of an effective phrase corresponding to each alarm type into a preset model to obtain a prediction result, carrying out error analysis on the prediction result and a real result (an alarm grade corresponding to the alarm type and an emergency disposal scheme), and adjusting parameters of the preset model through back propagation.
In this embodiment, the preset model is trained through a training data set including the alarm information of multiple alarm types, the alarm level corresponding to the alarm information of each alarm type, and the emergency disposal scheme, and the obtained trained model can be used for predicting the alarm level of the alarm type and the emergency disposal scheme. Therefore, in practical application, the received alarm information can be predicted through the trained model, so that the alarm level and the emergency disposal scheme can be quickly obtained, the manual troubleshooting of operation and maintenance personnel is avoided, the problems of overlong disposal chain and low disposal efficiency are solved, and the fault treatment efficiency is improved.
Further, in order to improve the quality of the emergency disposal model, the trained emergency disposal model can be evaluated, and if the evaluation effect is poor, the training is performed again.
Wherein, the evaluation of the emergency treatment model can be evaluated by an expert method.
In addition, the following method may be adopted:
acquiring a test set; the test set comprises alarm information of various alarm types and alarms corresponding to the alarm types;
inputting the alarm information in the test set into a trained emergency disposal model, and outputting a prediction result containing an alarm grade and an emergency disposal scheme;
determining a real processing strategy corresponding to the test set and comprising an alarm level and an emergency treatment scheme;
evaluating the emergency treatment model by comparing the actual treatment strategy with the predicted result.
Referring to fig. 2, a flowchart of a method for obtaining an effective phrase according to an embodiment of the present invention is shown, where in this embodiment, the method includes:
s201: aiming at any phrase, calculating the frequency of the phrase in the alarm information to obtain the word frequency TF of the phrase, calculating the inverse text frequency index IDF of the phrase, and calculating the importance of the phrase through the word frequency TF of the phrase and the inverse text frequency index IDF;
for example, the following steps are carried out: TF represents the frequency of a certain phrase appearing in the alarm information contained in the data set of a certain alarm type, and the inverse text frequency index IDF of the phrase can be calculated by the following formula 1):
Figure BDA0002868146610000081
wherein t represents vocabulary, D represents alarm data, D represents a corpus containing a plurality of alarm data, | D | represents total data, DF (t, D) represents alarm data frequency, representing the number of alarm data containing vocabulary t in a specific corpus environment.
The importance of each phrase in the alarm type can be calculated by the following formula 2):
2)TF-IDF(t,d,D)＝TF(t,d)·IDF(t,D)；
wherein TF represents the word frequency and IDF represents the inverse text index.
S202: aiming at any one alarm type, screening out effective phrases of the alarm type according to the importance degree of each phrase in the alarm type.
In this embodiment, the N words with the top rank may be selected according to the importance of the valid phrases. Wherein N is a positive integer greater than zero.
The number of N may be set according to actual conditions, and for example, N may be 50.
In this embodiment, by the above method, an effective phrase with a high importance degree for indicating the alarm type is screened out, so that when a preset model is trained through the effective phrase, an emergency disposal model with higher accuracy can be obtained.
Referring to fig. 3, a further schematic flow chart of an emergency disposal method provided in an embodiment of the present invention is shown, where in this embodiment, the method includes:
s301: after receiving the warning information, performing word segmentation processing on the warning information;
in this embodiment, a plurality of methods may be used to perform word segmentation processing on the warning information, and this embodiment is not limited.
S302: inputting the word segmentation result into a preset emergency disposal model to obtain an alarm grade and an emergency disposal strategy corresponding to the alarm information; the emergency treatment model is obtained by the training method described in any one of the above S101 to S104.
In the embodiment, the alarm grade and the emergency disposal strategy of the alarm information can be predicted through the trained emergency disposal model, so that the problems of overlong disposal chain and low disposal efficiency of operation and maintenance personnel are avoided, and the fault treatment efficiency is improved.
Referring to fig. 4, a schematic structural diagram of a training apparatus for an emergency treatment model according to an embodiment of the present invention is shown, including:
a first obtaining unit 401, configured to obtain a training data set; the training data set comprises alarm information of a plurality of alarm types, alarm grades corresponding to the alarm information of each alarm type and an emergency disposal scheme;
a first word segmentation processing unit 402, configured to perform word segmentation processing on the alarm information of each alarm type;
a phrase processing unit 403, configured to obtain an effective phrase corresponding to each alarm type from the word segmentation result, and perform vectorization processing on the effective phrase;
the training unit 404 is configured to train a preset model by using the vector of the effective phrase corresponding to each alarm type as an input and using the alarm level and the emergency disposal scheme corresponding to each alarm type as an output, so as to obtain a trained emergency disposal model.
Optionally, the phrase processing unit includes:
the importance degree calculation operator unit is used for calculating the frequency of the phrase in the alarm information aiming at any phrase to obtain the word frequency TF of the phrase, calculating the inverse text frequency index IDF of the phrase and calculating the importance degree of the phrase through the word frequency TF of the phrase and the inverse text frequency index IDF;
and the screening subunit is used for screening the effective phrases of the alarm type according to the importance of each phrase in the alarm type aiming at any one alarm type.
Optionally, the method further includes:
the first determining unit is used for determining the data volume contained in the data set corresponding to each alarm type;
the second determining unit is used for determining whether the data sets of each alarm type are balanced or not according to the data volume contained in the data set corresponding to each alarm type;
the third determining unit is used for determining the data set corresponding to the alarm type to be adjusted if the data sets of each alarm type are not balanced;
and the adjusting unit is used for adjusting the alarm information contained in the data set corresponding to the alarm type to be adjusted.
Optionally, the adjusting unit includes:
the acquisition subunit is used for acquiring the alarm information of the alarm type to be adjusted;
and the supplementing subunit is used for supplementing the acquired alarm information into the data set corresponding to the alarm type to be adjusted.
Optionally, the method further includes:
a second obtaining unit, configured to obtain a test set; the test set comprises alarm information of various alarm types and alarms corresponding to the alarm types;
the output unit is used for inputting the alarm information in the test set into a trained emergency disposal model and outputting a prediction result containing an alarm grade and an emergency disposal scheme;
a fourth determining unit, configured to determine a real processing policy including an alarm level and an emergency handling scheme corresponding to the test set;
and the evaluation unit is used for evaluating the emergency disposal model by comparing the real processing strategy with the prediction result.
According to the device of the embodiment, the preset model is trained through the alarm information comprising multiple alarm types, the alarm levels corresponding to the alarm information of each alarm type and the training data set of the emergency disposal scheme, and the obtained trained model can be used for predicting the alarm levels of the alarm types and the emergency disposal scheme. Therefore, in practical application, the received alarm information can be predicted through the trained model, so that the alarm level and the emergency disposal scheme can be quickly obtained, the manual troubleshooting of operation and maintenance personnel is avoided, the problems of overlong disposal chain and low disposal efficiency are solved, and the fault treatment efficiency is improved.
Referring to fig. 5, a schematic structural diagram of an emergency disposal device provided in an embodiment of the present invention is shown, where the emergency disposal device includes:
the second word segmentation processing unit 501 is configured to perform word segmentation processing on the alarm information after the alarm information is received;
the prediction unit 502 is configured to input the word segmentation result into a preset emergency disposal model, so as to obtain an alarm level and an emergency disposal policy corresponding to the alarm information; the emergency treatment model is obtained by the training method of any one of the preceding claims 1 to 5.
The device of this embodiment through the emergent processing model that trains, can predict the warning grade and the emergent processing strategy of reporting an emergency and asking for help or increased vigilance information to avoided fortune dimension personnel to investigate by hand, makeed the problem that the processing chain overlength, processing inefficiency, promoted fault handling efficiency.
Referring to fig. 6, a schematic structural diagram of an electronic device according to an embodiment of the present invention is shown, where in this embodiment, the electronic device includes:
a memory 601 and a processor 602;
the memory 601 is used for storing programs;
the processor 602 is configured to execute the emergency treatment model training method described above when executing the program stored in the memory 601:
acquiring a training data set; the training data set comprises alarm information of a plurality of alarm types, alarm grades corresponding to the alarm information of each alarm type and an emergency disposal scheme;
performing word segmentation processing on the alarm information of each alarm type;
obtaining effective phrases corresponding to each alarm type from the word segmentation result, and carrying out vectorization processing on the effective phrases;
and taking the vector of the effective phrase corresponding to each alarm type as input, taking the alarm grade corresponding to each alarm type and the emergency disposal scheme as output, and training a preset model to obtain a trained emergency disposal model.
Optionally, the obtaining an effective phrase corresponding to each alarm type from the word segmentation result includes:
aiming at any phrase, calculating the frequency of the phrase in the alarm information to obtain the word frequency TF of the phrase, calculating the inverse text frequency index IDF of the phrase, and calculating the importance of the phrase through the word frequency TF of the phrase and the inverse text frequency index IDF;
aiming at any one alarm type, screening out effective phrases of the alarm type according to the importance degree of each phrase in the alarm type.
Optionally, the method further includes:
determining the data volume contained in the data set corresponding to each alarm type;
determining whether the data sets of each alarm type are balanced or not according to the data volume contained in the data set corresponding to each alarm type;
if the data sets of each alarm type are not balanced, determining the data set corresponding to the alarm type to be adjusted;
and adjusting the alarm information contained in the data set corresponding to the alarm type to be adjusted.
Optionally, the adjusting the alarm information included in the data set corresponding to the alarm type to be adjusted includes:
collecting alarm information of the alarm type to be adjusted;
and supplementing the acquired alarm information into a data set corresponding to the alarm type to be adjusted.
Optionally, the method further includes:
acquiring a test set; the test set comprises alarm information of various alarm types and alarms corresponding to the alarm types;
inputting the alarm information in the test set into a trained emergency disposal model, and outputting a prediction result containing an alarm grade and an emergency disposal scheme;
determining a real processing strategy corresponding to the test set and comprising an alarm level and an emergency treatment scheme;
evaluating the emergency treatment model by comparing the actual treatment strategy with the predicted result.
It should be noted that, in the present specification, the embodiments are all described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments may be referred to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. An emergency treatment model training method, comprising:
acquiring a training data set; the training data set comprises alarm information of a plurality of alarm types, alarm grades corresponding to the alarm information of each alarm type and an emergency disposal scheme;
performing word segmentation processing on the alarm information of each alarm type;
obtaining effective phrases corresponding to each alarm type from the word segmentation result, and carrying out vectorization processing on the effective phrases;
and taking the vector of the effective phrase corresponding to each alarm type as input, taking the alarm grade corresponding to each alarm type and the emergency disposal scheme as output, and training a preset model to obtain a trained emergency disposal model.
2. The method according to claim 1, wherein the obtaining of the valid phrase corresponding to each alarm type from the word segmentation result comprises:
aiming at any phrase, calculating the frequency of the phrase in the alarm information to obtain the word frequency TF of the phrase, calculating the inverse text frequency index IDF of the phrase, and calculating the importance of the phrase through the word frequency TF of the phrase and the inverse text frequency index IDF;
aiming at any one alarm type, screening out effective phrases of the alarm type according to the importance degree of each phrase in the alarm type.
3. The method of claim 1, further comprising:
determining the data volume contained in the data set corresponding to each alarm type;
determining whether the data sets of each alarm type are balanced or not according to the data volume contained in the data set corresponding to each alarm type;
if the data sets of each alarm type are not balanced, determining the data set corresponding to the alarm type to be adjusted;
and adjusting the alarm information contained in the data set corresponding to the alarm type to be adjusted.
4. The method according to claim 3, wherein the adjusting the alarm information included in the data set corresponding to the alarm type to be adjusted includes:
collecting alarm information of the alarm type to be adjusted;
and supplementing the acquired alarm information into a data set corresponding to the alarm type to be adjusted.
5. The method of claim 1, further comprising:
acquiring a test set; the test set comprises alarm information of various alarm types and alarms corresponding to the alarm types;
inputting the alarm information in the test set into a trained emergency disposal model, and outputting a prediction result containing an alarm grade and an emergency disposal scheme;
determining a real processing strategy corresponding to the test set and comprising an alarm level and an emergency treatment scheme;
evaluating the emergency treatment model by comparing the actual treatment strategy with the predicted result.
6. An emergency disposal method, comprising:
after receiving the warning information, performing word segmentation processing on the warning information;
inputting the word segmentation result into a preset emergency disposal model to obtain an alarm grade and an emergency disposal strategy corresponding to the alarm information; the emergency treatment model is obtained by the training method of any one of the preceding claims 1 to 5.
7. An emergency treatment model training device, comprising:
a first acquisition unit for acquiring a training data set; the training data set comprises alarm information of a plurality of alarm types, alarm grades corresponding to the alarm information of each alarm type and an emergency disposal scheme;
the first word segmentation processing unit is used for carrying out word segmentation processing on the alarm information of each alarm type;
the phrase processing unit is used for acquiring effective phrases corresponding to each alarm type from the word segmentation result and carrying out vectorization processing on the effective phrases;
and the training unit is used for taking the vector of the effective phrase corresponding to each alarm type as input, taking the alarm grade corresponding to each alarm type and the emergency disposal scheme as output, and training a preset model to obtain a trained emergency disposal model.
8. The apparatus of claim 7, wherein the phrase processing unit comprises:
the importance degree calculation operator unit is used for calculating the frequency of the phrase in the alarm information aiming at any phrase to obtain the word frequency TF of the phrase, calculating the inverse text frequency index IDF of the phrase and calculating the importance degree of the phrase through the word frequency TF of the phrase and the inverse text frequency index IDF;
and the screening subunit is used for screening the effective phrases of the alarm type according to the importance of each phrase in the alarm type aiming at any one alarm type.
9. An emergency disposal device, comprising:
the second word segmentation processing unit is used for performing word segmentation processing on the alarm information after the alarm information is received;
the prediction unit is used for inputting the word segmentation result into a preset emergency disposal model to obtain an alarm grade and an emergency disposal strategy corresponding to the alarm information; the emergency treatment model is obtained by the training method of any one of the preceding claims 1 to 5.
10. An electronic device, comprising:
a memory and a processor;
the memory is used for storing programs;
the processor is configured to execute the emergency treatment model training method of any one of claims 1-5 when executing the program stored in the memory.
CN202011588726.3A 2020-12-29 2020-12-29 Training method of emergency disposal model, emergency disposal method and device Pending CN112612891A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011588726.3A CN112612891A (en) 2020-12-29 2020-12-29 Training method of emergency disposal model, emergency disposal method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011588726.3A CN112612891A (en) 2020-12-29 2020-12-29 Training method of emergency disposal model, emergency disposal method and device

Publications (1)

Publication Number Publication Date
CN112612891A true CN112612891A (en) 2021-04-06

Family

ID=75248677

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011588726.3A Pending CN112612891A (en) 2020-12-29 2020-12-29 Training method of emergency disposal model, emergency disposal method and device

Country Status (1)

Country Link
CN (1) CN112612891A (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111522704A (en) * 2020-03-04 2020-08-11 平安科技（深圳）有限公司 Alarm information processing method, device, computer device and storage medium
CN111666171A (en) * 2020-06-04 2020-09-15 中国工商银行股份有限公司 Fault identification method and device, electronic equipment and readable storage medium

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111522704A (en) * 2020-03-04 2020-08-11 平安科技（深圳）有限公司 Alarm information processing method, device, computer device and storage medium
CN111666171A (en) * 2020-06-04 2020-09-15 中国工商银行股份有限公司 Fault identification method and device, electronic equipment and readable storage medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
范昊 等: "Python语言基础与实践", 30 June 2020, 武汉大学出版社, pages: 489 - 491 *

Similar Documents

Publication Publication Date Title
US20200293924A1 (en) Gbdt model feature interpretation method and apparatus
CN111460250A (en) Image data cleaning method, image data cleaning device, image data cleaning medium, and electronic apparatus
CN111275338A (en) Method, device, equipment and storage medium for judging enterprise fraud behaviors
CN111177655B (en) Data processing method and device and electronic equipment
CN114444608B (en) Data set quality evaluation method and device, electronic equipment and storage medium
CN111209472A (en) Railway accident fault association and accident fault reason analysis method and system
CN115865483A (en) Abnormal behavior analysis method and device based on machine learning
CN113704389A (en) Data evaluation method and device, computer equipment and storage medium
CN114706986A (en) Multi-category emotion classification method and device and computer storage medium
CN109271495B (en) Question-answer recognition effect detection method, device, equipment and readable storage medium
CN107480703B (en) Transaction fault detection method and device
CN110704614B (en) Information processing method and device for predicting user group type in application
CN112612891A (en) Training method of emergency disposal model, emergency disposal method and device
CN116383048A (en) Software quality information processing method and device
CN116795978A (en) Complaint information processing method and device, electronic equipment and medium
CN116166967A (en) Data processing method, equipment and storage medium based on meta learning and residual error network
CN113962216A (en) Text processing method and device, electronic equipment and readable storage medium
CN110941709B (en) Information screening method and device, electronic equipment and readable storage medium
CN115018613A (en) Report analysis method, device, equipment, storage medium and product
CN113010339A (en) Method and device for automatically processing fault in online transaction test
CN111506877A (en) Testability verification method and device based on sequential network diagram inspection under Bayes framework
CN111754103A (en) Enterprise risk image method, device, computer equipment and readable storage medium
CN114662903A (en) Automated credit granting method, apparatus, device, medium, and program product
CN116955178A (en) Method and device for determining server, storage medium and electronic equipment
CN117764218A (en) Abnormal business behavior prediction method and system

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