Method, device, equipment, storage medium and program product for mining operation and maintenance knowledge

Abstract

The present disclosure provides a method for mining operation and maintenance knowledge, which can be applied to the field of artificial intelligence technology. The method for mining operation and maintenance knowledge includes: obtaining text data of at least one work order; performing named entity recognition on the text data through a first neural network model to determine at least one entity; performing entity alignment on the determined entity; classifying the text data through a second neural network model to determine the work order type of the text data; generating operation and maintenance knowledge based on the work order type and the entity that has undergone entity alignment. The present disclosure also provides an operation and maintenance knowledge mining device, equipment, storage medium and program product.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and specifically to a method, device, equipment, storage medium and program product for mining operation and maintenance knowledge. Background Art

[0002] In recent years, with the improvement of enterprise informatization, the scale of IT systems has become larger and larger, and the technologies involved have become more and more abundant. When a failure occurs in the online operation of the IT system, problem location, root cause analysis and fault handling are often done manually, and the timeliness and effectiveness of the handling are heavily dependent on the individual skills and experience of the operation and maintenance personnel.

[0003] At present, operations knowledge can be used to assist staff in quickly handling faults. Operations knowledge is a special data set used for IT operations management. It is a collection of domain knowledge required to solve operations problems. However, operations knowledge often requires manual mining, which is costly. Therefore, how to use existing resources to automatically mine operations knowledge has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of the above problems, the present disclosure provides a method, apparatus, device, storage medium and program product for mining operation and maintenance knowledge.

[0005] According to a first aspect of the present disclosure, a method for mining operation and maintenance knowledge is provided, comprising:

[0006] Get text data of at least one work order;

[0007] Performing named entity recognition on the text data by using a first neural network model to determine at least one entity;

[0008] Performing entity alignment on the determined entities;

[0009] Classifying the text data by a neural network model to determine the work order type of the text data;

[0010] Operation and maintenance knowledge is generated according to the work order type and the entity that has undergone entity alignment.

[0011] According to an embodiment of the present disclosure, performing named entity recognition on the text data set by using a first neural network model to determine at least one entity includes:

[0012] Preprocessing the text data to obtain a target word set;

[0013] Extracting at least one target word from the target word set to obtain a sequence to be recognized;

[0014] According to the preset entity type, an entity label is assigned to each word in the sequence to be identified;

[0015] Named entity recognition is performed on the sequence to be recognized according to the entity tag of each word in the sequence to be recognized and the entity tag of at least one word adjacent to each word to determine at least one entity.

[0016] According to an embodiment of the present disclosure, generating operation and maintenance knowledge according to the work order type and the entity after entity alignment includes: for each entity,

[0017] A piece of information is formed by the entity, the entity type of the entity and the work order type, serving as the operation and maintenance knowledge.

[0018] According to an embodiment of the present disclosure, the target word set includes a word vector of at least one word, and the dimension of the word vector is set between 150 and 350.

[0019] According to an embodiment of the present disclosure, the first neural network model includes: a multi-layer bidirectional long short-term memory network layer, a Relu activation function and a conditional random field network layer.

[0020] According to an embodiment of the present disclosure, the number of layers of the bidirectional long short-term memory network layer is set between 10 and 35 layers.

[0021] According to an embodiment of the present disclosure, a regularization operation is performed on features output by at least one layer of the bidirectional long short-term memory network layer;

[0022] The weight of the regularization operation is set between 0.3 and 0.9.

[0023] According to an embodiment of the present disclosure, the classifying the text data by using a second neural network model to determine the work order type of the text data includes:

[0024] Obtaining a sentence vector of the text data;

[0025] Determine the work order type of the text data by using a multinomial logistic regression algorithm according to the acquired sentence vector; and / or,

[0026] Obtaining a document vector for the text data;

[0027] According to the acquired document vector, the work order type of the text data is determined by a multinomial logistic regression algorithm.

[0028] According to an embodiment of the present disclosure, performing entity alignment on the determined entity includes:

[0029] According to a preset mapping rule table, entities belonging to the same real object among the determined entities are unified into one entity.

[0030] A second aspect of the present disclosure provides a device for mining operation and maintenance knowledge, including:

[0031] An acquisition module, used to acquire text data of at least one work order;

[0032] A first processing module, configured to perform named entity recognition on the text data through a first neural network model to determine at least one entity;

[0033] A second processing module, used for performing entity alignment on the determined entities;

[0034] A third processing module, used for classifying the text data through a second neural network model to determine the work order type of the text data;

[0035] A generation module is used to generate operation and maintenance knowledge according to the work order type and the entity that has undergone entity alignment.

[0036] The third aspect of the present disclosure provides an electronic device, comprising: one or more processors; a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above-mentioned operation and maintenance knowledge mining method.

[0037] The fourth aspect of the present disclosure further provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the above-mentioned operation and maintenance knowledge mining method.

[0038] The fifth aspect of the present disclosure also provides a computer program product, including a computer program, which implements the above-mentioned operation and maintenance knowledge mining method when executed by a processor.

[0039] One or more of the above embodiments have the following advantages or beneficial effects:

[0040] The present disclosure can utilize the work orders generated during the operation and maintenance process to mine operation and maintenance knowledge, which is low-cost and high-efficiency compared to the traditional technical solution of manually mining operation and maintenance knowledge. In addition, the present disclosure can mine operation and maintenance knowledge from unstructured text data by classifying text data and identifying entities. At the same time, after performing named entity recognition, the identified entities are standardized through entity alignment, thereby eliminating ambiguity and achieving higher data readability. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0042] Figure 1 The application scenario diagram schematically shows the method, device, equipment, storage medium and program product for mining operation and maintenance knowledge according to the embodiment of the present disclosure;

[0043] Figure 2 A flowchart of a method for mining operation and maintenance knowledge according to an embodiment of the present disclosure is schematically shown;

[0044] Figure 3 A flowchart for performing entity recognition according to an embodiment of the present disclosure is schematically shown;

[0045] Figure 4 A flowchart of entity alignment according to an embodiment of the present disclosure is schematically shown;

[0046] Figure 5 One of the flowcharts for determining the classification of a work order according to an embodiment of the present disclosure is schematically shown:

[0047] Figure 6 The second flowchart of determining the classification of a work order according to an embodiment of the present disclosure is schematically shown;

[0048] Figure 7 A structural block diagram of an operation and maintenance knowledge mining device according to an embodiment of the present disclosure is schematically shown;

[0049] Figure 8 A block diagram of an electronic device suitable for implementing a method for mining operation and maintenance knowledge according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0050] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0051] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0052] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0053] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0054] The log collection method, device, equipment, storage medium and program product disclosed in the present invention can be used in the information transmission business in the financial field, and can also be used in any field other than the financial field. The application field of the log collection method, device, equipment, storage medium and program product disclosed in the present invention is not limited.

[0055] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision, disclosure and application of user personal information involved comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

[0056] An embodiment of the present disclosure provides a method for mining operation and maintenance knowledge, which includes: obtaining a text data set, the text data set including text data of at least one work order; performing named entity recognition on the text data through a first neural network model to determine at least one entity; performing entity alignment on the determined entities; classifying the text data through a second neural network model to determine the work order type of the text data; and generating operation and maintenance knowledge based on the work order type and the entity that has undergone entity alignment.

[0057] The operation and maintenance knowledge mining method of the embodiment of the present disclosure can utilize the work orders generated during the operation and maintenance process to mine the operation and maintenance knowledge, which is low-cost and high-efficiency compared to the scheme of manually mining the operation and maintenance knowledge in the traditional technical scheme. In addition, the present disclosure can mine the operation and maintenance knowledge from the unstructured text data by classifying the text data and identifying entities. At the same time, after the named entity recognition, the identified entities are standardized by entity alignment, thereby eliminating ambiguity and achieving higher data readability.

[0058] Figure 1 The application scenario diagram of the operation and maintenance knowledge mining method, device, equipment, storage medium and program product according to the embodiment of the present disclosure is schematically shown.

[0059] like Figure 1 As shown, the application scenario 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0060] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0061] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0062] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0063] It should be noted that the mining method of operation and maintenance knowledge provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the mining device of operation and maintenance knowledge provided in the embodiment of the present disclosure can generally be set in the server 105. The mining method of operation and maintenance knowledge provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the mining device of operation and maintenance knowledge provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0064] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0065] The following will be based on Figure 1 The scene described by Figure 2 to Figure 6 The method for mining operation and maintenance knowledge of the disclosed embodiment is described in detail.

[0066] Figure 2 The flowchart of the method for mining operation and maintenance knowledge according to an embodiment of the present disclosure is schematically shown. Figure 2 As shown, the operation and maintenance knowledge mining method of this embodiment includes steps S210 to S250. It should be noted that although Figure 2 The steps in the figure are shown in sequence as indicated by the arrows, but the steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of the steps, and they can be executed in other orders. Moreover, at least some of the steps in the figure may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0067] In step S210, text data of at least one work order is obtained.

[0068] Issue Ticket refers to the carrier of online process flow due to failures, requests, consultations, etc. during the operation and maintenance of enterprise IT systems.

[0069] Text data can be unstructured data. For example, text data can refer to data obtained by extracting part or all of the text content in a work order. For example, the summary in the work order can be extracted as text data. Generally speaking, the summary in the work order records the key information of the work order, such as the fault description or the consultation content, etc. Therefore, using the summary in the work order as text data can ensure the validity of the text data on the one hand, and on the other hand, since it is not necessary to collect the entire content of the entire work order, it can reduce the workload of data collection.

[0070] Optionally, the text data of multiple work orders can be mined for operation and maintenance knowledge in sequence according to a preset collection strategy, or the text data of multiple work orders can be mined for operation and maintenance knowledge in batches at the same time, which can be determined according to actual needs. For the sake of clarity, the following takes the mining of operation and maintenance knowledge of the text data of a work order as an example to illustrate the present disclosure.

[0071] In step S220, named entity recognition is performed on the text data using a first neural network model to determine at least one entity.

[0072] In the embodiments of the present disclosure, named entity recognition (NER) refers to identifying entities with specific meanings in text data, such as device names, fault contents, etc.

[0073] Optionally, specific entity types can be configured according to the content involved in the operation and maintenance process. When performing named entity recognition, the content corresponding to the entity type is extracted from the text data to determine the entity. For example, entity types may include: event, failure, examination, object, operation, etc. Taking "failure" as an example, the corresponding entity may include device name, failure content, etc.

[0074] Optionally, the first neural network model may include a BiLSTM-CRF model. The BiLSTM-CRF model can accurately determine the named entity of the target to be identified in combination with the context of the target to be identified. The specific content will be described in detail below and will not be repeated here.

[0075] In step S230, entity alignment is performed on the determined entities.

[0076] In the disclosed embodiment, entity alignment may be: unifying entities corresponding to the same real object into one description, thereby standardizing the identified entities and eliminating ambiguity. For example, assuming that entity A1 and entity A2 are obtained after entity recognition, entity A1 and entity A2 both represent the same real object A, for example, entity A1 is "computer", entity A2 is "computer" in English, that is, entity A2 is "Computer", entity A1 and entity A2 correspond to the same real object, both are "computer", therefore, entity A1 and entity A2 may be aligned to unify entity A1 and entity A2 into "computer".

[0077] In step S240, the text data is classified by a second neural network model to determine the work order type of the text data.

[0078] Optionally, in the embodiments of the present disclosure, some work order types may be pre-configured according to actual needs in the operation and maintenance process. For example, in the IT operation and maintenance field, several work order types including "problem", "change" and "consultation" may be configured according to the purpose of initiating the work order. According to the business field, several work order types including "application", "computing", "storage", "network", "security" and "environment" may be configured.

[0079] Optionally, in the embodiment of the present disclosure, the existing work order type can also be updated according to the current text data during the text classification process. For example, for a certain text data, after classification, it cannot be classified into the existing work order type. At this time, the text data can be classified into the "unknown" work order type, and the work order type of the text data can be determined by computer or manual means in the subsequent process, and the determined work order type is used as a new work order type for updating.

[0080] Optionally, the second neural network model may include a FastText model. By using the FastText model, the training speed can be accelerated while ensuring classification accuracy close to that of a deep neural network, thereby improving work efficiency.

[0081] In step S250, operation and maintenance knowledge is generated according to the work order type and the entity that has undergone entity alignment.

[0082] In the disclosed embodiment, the work order type and the entity after entity alignment can constitute a piece of information as operation and maintenance knowledge. Optionally, the entity type mentioned above can also be used as a label to constitute a piece of information together with the work order type and the entity after entity alignment as operation and maintenance knowledge.

[0083] For example, for a summary message in a work order: "Linux available memory is low, current value 9.98, threshold 0-10", through the above step S220, the entity "Linux available memory is low" with entity type as event (Event) can be extracted, and after entity alignment through step S230, "Linux insufficient memory" is obtained. Through step S240, the above content is classified into "application" in the business field. At this time, through step S250, "event", "application", and "Linux insufficient memory" can form a piece of information as operation and maintenance knowledge.

[0084] The operation and maintenance knowledge mining method of the embodiment of the present disclosure can utilize the work orders generated during the operation and maintenance process to mine the operation and maintenance knowledge, which is low-cost and high-efficiency compared to the scheme of manually mining the operation and maintenance knowledge in the traditional technical scheme. In addition, the present disclosure can mine the operation and maintenance knowledge from the unstructured text data by classifying the text data and identifying entities. At the same time, after the named entity recognition, the identified entities are standardized by entity alignment, thereby eliminating ambiguity and achieving higher data readability.

[0085] Combine the following Figures 3 to 6 The method for mining operation and maintenance knowledge of the embodiment of the present disclosure is further explained.

[0086] Figure 3Schematically shown is a flowchart for entity recognition according to an embodiment of the present disclosure, as Figure 3 shown. In some specific embodiments, step S220 includes steps S221 to S224.

[0087] In step S221, the text data is preprocessed to obtain a target word set.

[0088] In an embodiment of the present disclosure, the text data of the work order can be obtained from the background database of the work order system, and the preprocessing of the text may include data cleaning. Optionally, the data cleaning may include: removing useless symbols such as extra spaces, tokenizing the text, and removing stop words, for example, "de" (of) and "ya" (ah), etc.

[0089] Exemplarily, for a summary information in the work order described above: "The available memory of Linux is low", after tokenizing and removing stop words through step S211, etc., we get: "Linux", "The available memory is low".

[0090] In step S222, at least one target word is extracted from the target word set to obtain a sequence to be recognized.

[0091] In step S223, according to the entity type, an entity label is assigned to each word in the sequence to be recognized;

[0092] In step S224, named entity recognition is performed on the sequence to be recognized according to the entity labels of each word in the sequence to be recognized and the entity labels of at least one adjacent word to each word to determine at least one entity.

[0093] In some specific embodiments, the first neural network model includes a BiLSTM-CRF model. Optionally, the target word set includes word vectors of at least one word, and the first neural network model includes a word vector embedding layer. Through the word vector embedding layer, the word vector of each word in the target word set can be obtained. For example, the word vector embedding layer may include a Word2vec model. Through the Word2vec model, according to each word in the target word set and at least one adjacent word to the word, the word vector of the word can be determined.

[0094] Optionally, in an embodiment of the present disclosure, the existing text data of the work order can be used as a sample to train the Word2vec algorithm through the Skip-Gram word representation method. Through Skip-Gram, the word vector representation of nearby words can be predicted based on the input word. After training the Word2vec algorithm through Skip-Gram, the word vector of each word in the sample can be obtained. Among them, each word corresponds to a word vector, and the similarity between words can be determined by calculating the cosine similarity between the word vectors of the two words.

[0095] In some specific embodiments, the dimension of the word vector is set between 150 and 350, for example, the dimension of the word vector is set to 200; or, the dimension of the word vector is set to 300. Optionally, in the embodiment of the present disclosure, the dimension of the word vector is set to 300.

[0096] In some specific embodiments, the first neural network model includes a multi-layer bidirectional long short-term memory network layer, a Relu activation function, and a conditional random field network layer.

[0097] In the disclosed embodiment, part or all of the content can be extracted from the target word set to form a sequence to be recognized. For example, the target word set includes words in the text data of a work order. Therefore, all the content can be extracted from the target word set to form a sequence to be recognized. For example, "Linux", "available memory", "current value", "greater than", "threshold" and "low" are extracted from the target word set to form a sequence to be recognized. When processing a sentence, the bidirectional long short-term memory network layer will consider every word in the entire sentence, rather than just the most recent word or words. For example, when processing the sentence "Linux available memory is low", "available memory" has "Linux" in front and "low" in the back. When recognizing "available memory", the bidirectional long short-term memory network layer considers the front "Linux" and the back "low". By combining the context, it can more accurately determine that "available memory" is a performance parameter of a computer device. If only one-directional LSTM network is used, the information about the order of characters and words in the sentence will be lost. For example, "the current value is greater than the threshold" and "the threshold is greater than the current value" will not be distinguished. If a bidirectional LSTM network is used, the "current value is greater than the threshold" is processed by the forward LSTM network, and the "threshold is greater than the current value" is processed by the reverse LSTM network. The results of the two LSTM network processing are combined to obtain the order relationship of each character and word in the sentence, thereby distinguishing "the current value is greater than the threshold" and "the threshold is greater than the current value".

[0098] The input of the bidirectional LSTM network layer can be the output of the above-mentioned word embedding layer. For example, the output of the word embedding layer is a 70*300 matrix, where 70 refers to 70 words and 300 refers to the dimension of the word embedding. Then the output of the bidirectional LSTM network layer is also a matrix with a size of [2*70, HIDDENUNIT], where 70 refers to 70 words, the forward LSTM network plus the reverse LSTM network is 140, and HIDDENUNIT is a preset length, such as 300. Optionally, each element of the matrix is ​​a floating point number.

[0099] The output of the bidirectional long short-term memory network passes through the Relu activation function and serves as the input of the conditional random field network layer. The Relu activation function is a piecewise linear function. If the input of the Relu activation function is positive, the input will be output directly. If the input of the Relu activation function is negative, the Relu activation function will output "0".

[0100] The conditional random field network layer finally determines the entity label of each word based on the received input.

[0101] Optionally, in the disclosed embodiment, the entity label of each word can be annotated using the BIOES annotation method according to the preset entity type. Among them, B represents belonging to the target entity type and is the first word; I represents belonging to the target entity type but not the first word; O represents not belonging to the target entity type; E represents belonging to the target entity type and is the last word; S represents an entity composed of a single word. Entity types may include: Event, Failure, Examination, Object, Operation, etc.

[0102] For example, taking the entity type of "event" as an example, for the sequence to be identified consisting of "Linux", "available memory", "current value", "greater than", "threshold" and "low", after assigning entity labels, we obtain "Linux, B-event", "available memory, I-event", "current value, O", "greater than, O", "threshold, O", "low, E-event", and then determine the entity "Linux available memory low" with the entity type of "event".

[0103] In some specific embodiments, the number of layers of the bidirectional long short-term memory network layer is set between 10 and 35 layers, for example, the number of layers of the bidirectional long short-term memory network layer is set to 12 layers; or, the number of layers of the bidirectional long short-term memory network layer is set to 32 layers. In this way, on the one hand, in the embodiments of the present disclosure, the number of layers of the bidirectional long short-term memory network layer will not be too many, thereby saving computing power; on the other hand, the actual needs of the operation and maintenance knowledge scenario can be met through the 12-layer (or 32-layer) bidirectional long short-term memory network layer, with good accuracy of entity recognition.

[0104] It should be noted that in the embodiments of the present disclosure, when training the first neural network model, it is not limited to providing text data of one work order at a time. For example, in some specific embodiments, text data of multiple work orders can be extracted at the same time to form a text data set, and the text data set is used as a training sample of the first neural network model to train the first neural network model.

[0105] Optionally, in the embodiment of the present disclosure, 70% of the training samples may be selected as a training set and 30% as a test set.

[0106] In some specific embodiments, the learning parameter of the first neural network model can be set between 0.01 and 0.0001. For example, the learning parameter can be set to 0.001. The specific value can be determined according to actual needs and is not limited here.

[0107] In some specific embodiments, a regularization operation is performed on the features output by at least one bidirectional long short-term memory network layer. The weight of the regularization operation is set between 0.3 and 0.9. For example, the weight of the regularization operation is set to 0.6, or the weight of the regularization operation is set to 0.8. Through the regularization operation, overfitting in the first neural network model can be reduced.

[0108] Optionally, during the training of the first neural network model, a dropout operation (Dropout) may be performed on the features output by each bidirectional long short-term memory network layer, wherein the dropout ratio may be set between 20% and 60%, for example, the dropout ratio is set to 40%. The dropout operation may reduce overfitting of the first neural network model during the training process.

[0109] Figure 4 A flow chart of entity alignment according to an embodiment of the present disclosure is schematically shown. Figure 4 As shown, in some specific embodiments, step S230 includes step S231.

[0110] In step S231, entities belonging to the same real object among the determined entities are unified into one entity according to a preset mapping rule table.

[0111] Currently, in the context of IT operations and maintenance, different named entities often correspond to the same real-world object, especially when it comes to the representation of some entities, such as the Chinese name and English abbreviation / English name of an IT system.

[0112] In the disclosed embodiment, the mapping rule table can be obtained by configuring a computer based on big data or by a staff member based on business experience. In the mapping rule table, an object in the real world can be associated with multiple entities. For example, "computer", "Computer" and "PC" can all be associated with "computer". When performing entity alignment, "computer", "Computer" and "PC" are unified as "computer". For another example, for the "Linux available memory is low" identified by entity in the previous text, it can be unified as "Linux insufficient memory". Through entity alignment, the description of the same object in the real world can be unified, thereby eliminating the ambiguity caused by different descriptions and improving readability.

[0113] Figure 5 One of the flowcharts for determining the classification of a work order according to an embodiment of the present disclosure is schematically shown. Figure 6 The second flowchart of determining the work order classification according to the embodiment of the present disclosure is schematically shown. Figure 5 and Figure 6 As shown, in some specific embodiments, step S240 includes step S241 and step S242, and / or, step S243 and step S244.

[0114] Step S241, obtaining the sentence vector of the text data.

[0115] Step S242, determining the work order type of the text data through a multinomial logistic regression algorithm based on the acquired sentence vector. And / or,

[0116] Step S243, obtaining the document vector of the text data.

[0117] Step S244, determining the work order type of the text data through a multinomial logistic regression algorithm based on the acquired document vector.

[0118] In the disclosed embodiment, the second neural network model includes the FastText model, which is a model open sourced by Facebook for effectively learning word representation and text classification. FastText supports supervised (classification) and unsupervised (embedding) representation of words and sentences.

[0119] Optionally, in the embodiments of the present disclosure, a sentence vector or a document vector may be obtained by averaging words or n-grams.

[0120] Taking n-gram as an example, for example, "insufficient system memory" is segmented into "system", "within the system", "memory", "cannot store", and "insufficient" through the 2-gram method, and then the segmented words are vectorized through word embedding and other methods. Finally, the word vectors are weighted averaged and other steps are performed to output the classification label of the text data, that is, to determine the type of work order for the text data.

[0121] Optionally, in the embodiments of the present disclosure, a hierarchical softmax may be used to speed up the classification process.

[0122] Optionally, you can configure the ticket type according to the purpose of initiating the ticket, such as problem, change, consultation, etc.

[0123] Optionally, the work order type can be configured by business area, for example, application, computing, storage, network, security, environment, etc.

[0124] For example, the text data "Linux available memory is low" mentioned in the previous article can be classified into the work order type "Application".

[0125] In summary, after performing entity recognition and entity alignment on the text data, we get "Linux insufficient memory" and the entity type is "event". After classifying the text data, we get that the text data belongs to the type of "application".

[0126] Finally, in step S250, "application" can be used as the work order type label, "event" can be used as the entity type label, and "Linux insufficient memory" can be used as the content to form a piece of information as operation and maintenance knowledge.

[0127] By using the operation and maintenance knowledge mining method of the embodiment of the present invention, the accompanying data in the traditional operation and maintenance process of enterprise IT, that is, the work order, can be used to mine the operation and maintenance knowledge from the work order text, thereby achieving the purpose of reducing costs and increasing efficiency. In addition, by using the BiLSTM-CRF model to perform named entity recognition, then eliminating ambiguity through entity alignment, and then using FastText to classify the text data, the intrusion into the existing IT system can be minimized, and valuable data can be extracted from unstructured text in the most cost-effective way to form operation and maintenance knowledge and enhance the enterprise IT operation and maintenance management level.

[0128] Based on the above operation and maintenance knowledge mining method, the present disclosure also provides an operation and maintenance knowledge mining device. Figure 7 The device is described in detail.

[0129] Figure 7 The structural block diagram of the operation and maintenance knowledge mining device according to the embodiment of the present disclosure is schematically shown. Figure 7As shown, the operation and maintenance knowledge mining device 700 of this embodiment includes an acquisition module 710 , a first processing module 720 , a second processing module 730 , a third processing module 740 and a generation module 750 .

[0130] The acquisition module 710 is used to acquire text data of at least one work order. In one embodiment, the acquisition module 710 can be used to execute the step S210 described above, which will not be described in detail here.

[0131] The first processing module 720 is used to perform named entity recognition on the text data through the first neural network model to determine at least one entity. In one embodiment, the first processing module 720 can be used to execute step S220 described above, which will not be repeated here.

[0132] The second processing module 730 is used to perform entity alignment on the determined entities. In one embodiment, the second processing module 730 can be used to execute the step S230 described above, which will not be described in detail here.

[0133] The third processing module 740 is used to classify the text data through the second neural network model to determine the work order type of the text data. In one embodiment, the third processing module 740 can be used to execute the step S240 described above, which will not be repeated here.

[0134] The generation module 750 is used to generate operation and maintenance knowledge according to the work order type and the entity that has undergone entity alignment. In one embodiment, the generation module 750 can be used to execute the step S250 described above, which will not be described in detail here.

[0135] The operation and maintenance knowledge mining device of the embodiment of the present disclosure can mine the operation and maintenance knowledge using the work orders generated during the operation and maintenance process, which has the characteristics of low cost and high efficiency compared to the scheme of manually mining the operation and maintenance knowledge in the traditional technical scheme. In addition, the present disclosure can mine the operation and maintenance knowledge from the unstructured text data by classifying the text data and identifying the entities. At the same time, after the named entity recognition, the identified entities are standardized by entity alignment, thereby eliminating ambiguity and achieving higher data readability.

[0136] According to an embodiment of the present disclosure, any multiple modules among the acquisition module 710, the first processing module 720, the second processing module 730, the third processing module 740 and the generation module 750 can be combined in one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the acquisition module 710, the first processing module 720, the second processing module 730, the third processing module 740 and the generation module 750 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation modes of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of the acquisition module 710, the first processing module 720, the second processing module 730, the third processing module 740 and the generation module 750 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0137] The operation and maintenance knowledge mining device of the embodiment of the present invention can utilize the accompanying data in the traditional operation and maintenance process of enterprise IT, that is, the work order, to mine the operation and maintenance knowledge from the work order text, thereby achieving the purpose of reducing costs and increasing efficiency. In addition, by using the BiLSTM-CRF model to perform named entity recognition, then eliminating ambiguity through entity alignment, and then using FastText to classify the text data, the intrusion into the existing IT system can be minimized, and valuable data can be extracted from unstructured text in the most cost-effective way to form operation and maintenance knowledge and enhance the enterprise IT operation and maintenance management level.

[0138] Figure 8 A block diagram of an electronic device suitable for implementing a method for mining operation and maintenance knowledge according to an embodiment of the present disclosure is schematically shown. Figure 8As shown, the electronic device 800 according to an embodiment of the present disclosure includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage part 808 to a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include an onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0139] In RAM 803, various programs and data required for the operation of electronic device 800 are stored. Processor 801, ROM 802 and RAM 803 are connected to each other via bus 804. Processor 801 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 802 and / or RAM 803. It should be noted that the program can also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in one or more memories.

[0140] According to an embodiment of the present disclosure, the electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to the bus 804. The electronic device 800 may further include one or more of the following components connected to the I / O interface 805: an input portion 806 including a keyboard, a mouse, etc. An output portion 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc. A storage portion 808 including a hard disk, etc. And a communication portion 809 including a network interface card such as a LAN card, a modem, etc. The communication portion 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed, so that the computer program read therefrom is installed into the storage portion 808 as needed.

[0141] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment. It may also exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method for mining operation and maintenance knowledge according to the embodiment of the present disclosure is implemented.

[0142] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present disclosure, a computer-readable storage medium may include the ROM 802 and / or RAM 803 described above and / or one or more memories other than ROM 802 and RAM 803.

[0143] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program includes a program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the operation and maintenance knowledge mining method provided by the embodiment of the present disclosure.

[0144] The above functions defined in the system / device of the embodiment of the present disclosure are executed when the computer program is executed by the processor 801. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0145] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 809, and / or installed from a removable medium 811. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0146] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the processor 801, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0147] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0148] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0149] It will be appreciated by those skilled in the art that the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations and / or combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways without departing from the spirit and teachings of the present disclosure. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0150] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A method for mining operation and maintenance knowledge, characterized in that: include: Get text data of at least one work order; Performing named entity recognition on the text data by using a first neural network model to determine at least one entity; Performing entity alignment on the determined entities; Classifying the text data by a second neural network model to determine the work order type of the text data; Generate operation and maintenance knowledge according to the work order type and the entity that has undergone entity alignment; The step of performing named entity recognition on the text dataset by using a first neural network model to determine at least one entity includes: Preprocessing the text data to obtain a target word set; Extracting at least one target word from the target word set to obtain a sequence to be recognized; According to the preset entity type, an entity label is assigned to each word in the sequence to be identified; Performing named entity recognition on the sequence to be recognized according to the entity tag of each word in the sequence to be recognized and the entity tag of at least one word adjacent to each word to determine at least one entity; The first neural network model includes: a multi-layer bidirectional long short-term memory network layer, a Relu activation function and a conditional random field network layer; the number of layers of the bidirectional long short-term memory network layer is set between 10 and 35; Performing a regularization operation on the features output by at least one layer of the bidirectional long short-term memory network layer; wherein the weight of the regularization operation is set between 0.3 and 0.9; The classifying the text data by using a second neural network model to determine the work order type of the text data includes: Obtaining a sentence vector of the text data; Determine the work order type of the text data by using a multinomial logistic regression algorithm according to the acquired sentence vector; and / or, Obtaining a document vector for the text data; According to the acquired document vector, the work order type of the text data is determined by a multinomial logistic regression algorithm.

2. The excavation method according to claim 1, characterized in that: The generating operation and maintenance knowledge according to the work order type and the entity after entity alignment includes: for each entity, A piece of information is formed by the entity, the entity type of the entity and the work order type, serving as the operation and maintenance knowledge.

3. The excavation method according to claim 1, characterized in that: in, The target word set includes a word vector of at least one word, and the dimension of the word vector is set between 150 and 350.

4. The excavation method according to claim 1, characterized in that: The performing entity alignment on the determined entities includes: According to a preset mapping rule table, entities belonging to the same real object among the determined entities are unified into one entity.

5. A device for mining operation and maintenance knowledge, characterized in that: include: An acquisition module, used to acquire text data of at least one work order; A first processing module, configured to perform named entity recognition on the text data through a first neural network model to determine at least one entity; A second processing module, used for performing entity alignment on the determined entities; A third processing module, used for classifying the text data through a second neural network model to determine the work order type of the text data; A generation module, used for generating operation and maintenance knowledge according to the work order type and the entity after entity alignment; The first processing module is also used to pre-process the text data to obtain a target word set; Extracting at least one target word from the target word set to obtain a sequence to be recognized; assigning an entity label to each word in the sequence to be recognized according to a preset entity type; Named entity recognition is performed on the sequence to be recognized according to the entity tag of each word in the sequence to be recognized and the entity tag of at least one word adjacent to each word to determine at least one entity tag. The first neural network model includes: a multi-layer bidirectional long short-term memory network layer, a Relu activation function and a conditional random field network layer; the number of layers of the bidirectional long short-term memory network layer is set between 10 and 35 layers; a regularization operation is performed on the features output by at least one layer of the bidirectional long short-term memory network layer; wherein the weight of the regularization operation is set between 0.3 and 0.9; The third processing module is also used to obtain a sentence vector of the text data; determine the work order type of the text data through a multinomial logistic regression algorithm based on the obtained sentence vector; and / or obtain a document vector of the text data; determine the work order type of the text data through a multinomial logistic regression algorithm based on the obtained document vector.

6. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors execute the method for mining operation and maintenance knowledge according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that: Executable instructions are stored thereon, and when the instructions are executed by a processor, the processor executes the method for mining operation and maintenance knowledge according to any one of claims 1 to 4.

8. A computer program product, characterized in that The method comprises a computer program, which, when executed by a processor, implements the method for mining operation and maintenance knowledge according to any one of claims 1 to 4.

Images