Work order topic analysis method, device, electronic device and storage medium

Abstract

The present invention discloses a work order topic analysis method, device, electronic device and storage medium, which are used to accurately classify work order topics. The method of the present invention includes: extracting effective information from the work order text data to be analyzed through a representation learning algorithm; extracting characteristic key information from the effective information; performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation; using the characteristic word segmentation and a pre-trained topic discovery model to extract topic data from the work order text data to be analyzed; performing correlation analysis on the topic data to obtain related topic data; clustering the related topic data to obtain topic clusters; performing semantic matching on the topic clusters to obtain the work order topics corresponding to each topic cluster. Thereby, the accuracy of work order topic classification is improved, which is convenient for subsequent problem warning, automatic dispatching and hotline analysis of similar topics.

Description

Technical Field

[0001] The present invention relates to the technical field of topic analysis, and in particular to a topic analysis method, device, electronic equipment and storage medium. Background Art

[0002] In recent years, statistical natural language processing technology based on large corpora has developed rapidly, enabling machines to "understand" human language to a certain extent. However, as data accumulates, a large amount of short text data is not used reasonably, resulting in a large amount of effective information being wasted. Therefore, it is necessary to extract and analyze the accumulated data so as to make reasonable use of the data.

[0003] There are two main types of text classification models. One is topic classification, which helps relevant departments identify sensitive topics and analyze hot issues. The other is sentiment analysis, which analyzes the content of the appeal and understands the emotional changes of customers, so as to adjust marketing strategies in time and improve the quality of the content and customer satisfaction.

[0004] The number of work orders received by the 12345 hotline is large, the types are wide, and there are many difficult issues. In addition, a large amount of social and public opinion information data has been accumulated during years of operation. How to conduct topic analysis on the large amount of collected social and public opinion information data is a technical problem that technical personnel in this field urgently need to solve.

[0005] In the prior art, the classification of work order data is mainly achieved by performing text mining on the collected data, extracting keywords, and then matching work orders according to the keywords. However, the above method relies on the accuracy of keyword extraction, so it is easy to produce classification anomalies. Summary of the invention

[0006] The present invention provides a work order topic analysis method, device, electronic device and storage medium for accurately classifying work order topics, thereby improving the accuracy of work order topic classification and facilitating subsequent problem warning, automatic dispatching and hotline analysis for similar topics.

[0007] The present invention provides a work order topic analysis method, comprising:

[0008] Extract effective information from the work order text data to be analyzed through representation learning algorithms;

[0009] Extracting characteristic key information from the effective information;

[0010] Performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation;

[0011] Using the feature segmentation and pre-trained topic discovery model, extracting topic data from the work order text data to be analyzed;

[0012] Performing correlation analysis on the topic data to obtain related topic data;

[0013] Clustering the related topic data to obtain topic clusters;

[0014] Semantic matching is performed on the topic clusters to obtain work order topics corresponding to each topic cluster.

[0015] Optionally, the step of extracting valid information from the work order text data to be analyzed by using a representation learning algorithm includes:

[0016] Determine redundant information in the work order text data to be analyzed by a representation learning algorithm;

[0017] The redundant information is eliminated to obtain valid information in the work order text data to be analyzed.

[0018] Optionally, the step of extracting topic data from the work order text data to be analyzed by using the feature segmentation and the pre-trained topic discovery model includes:

[0019] Encode the feature word segmentation using a semantic representation algorithm to obtain a text vector;

[0020] Mapping the text vector to a low-dimensional space using a word embedding algorithm to obtain a low-dimensional text vector;

[0021] The low-dimensional text vector is input into a pre-trained topic discovery model to output topic data.

[0022] Optionally, the step of performing correlation analysis on the topic data to obtain correlated topic data includes:

[0023] Extracting the associated data between the topic data from the work order text data to be analyzed; the associated data includes the complaining subject, the responsible subject, the disposal measures, and the time and space location;

[0024] Generate a knowledge graph of the topic data and the associated data to obtain associated topic data.

[0025] Optionally, it also includes:

[0026] Obtain multi-dimensional information of each work order topic, including ranking, number of requests, month-on-month comparison, and ranking change;

[0027] Hot topics are filtered from the work order topics according to the ranking, the number of requests, the month-on-month change and the ranking change.

[0028] The present invention also provides a work order topic analysis device, comprising:

[0029] The effective information extraction module is used to extract effective information from the work order text data to be analyzed through the representation learning algorithm;

[0030] A feature key information extraction module, used to extract feature key information from the valid information;

[0031] A word segmentation module is used to perform word segmentation processing on the characteristic key information to obtain characteristic word segmentations;

[0032] A topic data extraction module, used to extract topic data from the work order text data to be analyzed by using the feature segmentation and the pre-trained topic discovery model;

[0033] A correlation analysis module, used to perform correlation analysis on the topic data to obtain related topic data;

[0034] A clustering module, used for clustering the related topic data to obtain topic clusters;

[0035] The semantic matching module is used to perform semantic matching on the topic clusters to obtain the work order topics corresponding to each topic cluster.

[0036] Optionally, the effective information extraction module includes:

[0037] A redundant information determination submodule, used to determine the redundant information in the work order text data to be analyzed by using a representation learning algorithm;

[0038] The effective information extraction submodule is used to remove the redundant information and obtain the effective information in the work order text data to be analyzed.

[0039] Optionally, the topic data extraction module includes:

[0040] An encoding submodule, used for encoding the feature segmentation words using a semantic representation algorithm to obtain a text vector;

[0041] A low-dimensional text vector acquisition submodule is used to map the text vector to a low-dimensional space using a word embedding algorithm to obtain a low-dimensional text vector;

[0042] The topic data output submodule is used to input the low-dimensional text vector into the pre-trained topic discovery model and output the topic data.

[0043] The present invention also provides an electronic device, the device comprising a processor and a memory;

[0044] The memory is used to store program code and transmit the program code to the processor;

[0045] The processor is used to execute the work order topic analysis method as described in any one of the above items according to the instructions in the program code.

[0046] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store program code, and the program code is used to execute the work order topic analysis method as described in any one of the above items.

[0047] It can be seen from the above technical scheme that the present invention has the following advantages: the present invention discloses a method for analyzing work order topics, including: extracting effective information from the work order text data to be analyzed through a representation learning algorithm; extracting characteristic key information from the effective information; performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation; extracting topic data from the work order text data to be analyzed using characteristic word segmentation and a pre-trained topic discovery model; performing association analysis on the topic data to obtain associated topic data; clustering the associated topic data to obtain topic clusters; performing semantic matching on the clusters to obtain the work order topics corresponding to each topic cluster. This improves the accuracy of work order topic classification, and facilitates subsequent problem warning, automatic dispatching, and hotline analysis of similar topics. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0049] Figure 1 A flowchart of a work order topic analysis method provided by an embodiment of the present invention;

[0050] Figure 2 A flowchart of a work order topic analysis method provided by another embodiment of the present invention;

[0051] Figure 3 A schematic diagram of a knowledge graph established based on a work order provided in an embodiment of the present invention;

[0052] Figure 4 A schematic diagram of work order association provided by an embodiment of the present invention;

[0053] Figure 5 A schematic diagram of the structure of a work order topic analysis device provided by an embodiment of the present invention;

[0054] Figure 6 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Implementation

[0055] The embodiment of the present invention provides a work order topic analysis method, device, electronic device and storage medium for accurately classifying work order topics, thereby improving the accuracy of work order topic classification and facilitating subsequent problem warning, automatic dispatching and hotline analysis for similar topics.

[0056] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] See also Figure 1 , Figure 1 A flowchart of a work order topic analysis method provided by an embodiment of the present invention.

[0058] The present invention provides a work order topic analysis method, which may specifically include the following steps:

[0059] Step 101, extracting valid information from the work order text data to be analyzed through a representation learning algorithm;

[0060] Representation learning is a collection of techniques for learning a feature: converting raw data into a form that can be effectively exploited by machine learning. It avoids the trouble of manual feature extraction and allows computers to use features while also learning how to extract features.

[0061] A work order is a work document consisting of one or more operations; the content of the operation can be the basis for the superior department to issue tasks and the subordinate department to accept tasks.

[0062] In an embodiment of the present invention, for a massive amount of work order text data to be analyzed, effective information can be extracted therefrom through a representation learning algorithm.

[0063] Among them, effective information can be the content, frequency, relationship of the topic, the complaining subject, the responsible subject, the handling measures, the time and space location, etc.

[0064] Step 102, extracting key feature information from valid information;

[0065] After extracting valid information from the work order text data to be analyzed, feature extraction can be performed to obtain feature key information.

[0066] Step 103, performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation;

[0067] Word segmentation is the process of recombining continuous word sequences into word sequences according to certain specifications. Chinese word segmentation technology is the basis of natural language processing (NLP), which is mainly divided into word segmentation methods based on string matching, word segmentation methods based on understanding, and word segmentation methods based on statistics. The embodiments of the present invention do not impose specific restrictions on the specific algorithms used in word segmentation.

[0068] In a specific implementation, word segmentation processing can be used to parse the feature key information into information that is easy to use for topic mining and reasoning; therefore, after obtaining the feature key information, word segmentation processing can be performed on the feature key information to obtain feature word segmentations.

[0069] Step 104, using feature segmentation and a pre-trained topic discovery model to extract topic data from the work order text data to be analyzed;

[0070] After obtaining the feature segmentation, the feature segmentation and the pre-trained topic discovery model can be used to extract the topic data in the work order text data to be analyzed.

[0071] In the specific implementation, the topic discovery model can be trained through the following steps:

[0072] 1) Obtain historical work orders and mark the topics in the historical work orders;

[0073] 2) Pre-process the content description data of historical work orders marked with topics to remove redundant information in historical work orders, such as removing meaningless information such as "department", "address", and "call feedback" in the work order content description;

[0074] 3) Perform word segmentation on the historical work orders with redundant information removed, and divide the work order content description into words;

[0075] 4) Use the segmented words as training samples, topics as target outputs, and use deep learning algorithms to train a topic discovery model.

[0076] Step 105, performing correlation analysis on the topic data to obtain correlated topic data;

[0077] After the topic data in the feature segmentation is extracted through the topic discovery model, the obtained topic data can be subjected to a correlation analysis to obtain the related topic data.

[0078] In a specific implementation, the topic data can be analyzed for relevance through a knowledge graph. The relevance analysis can be the relevance between topics, the relevance between topics and information related to the topics, etc. The object of the relevance analysis can be selected according to the actual situation, and the embodiment of the present invention does not impose any specific restrictions on this.

[0079] Step 106, clustering the related topic data to obtain topic clusters;

[0080] Clustering: The process of dividing a collection of physical or abstract objects into multiple classes consisting of similar objects is called clustering. The clusters generated by clustering are a set of data objects that are similar to the objects in the same cluster and different from the objects in other clusters.

[0081] Through clustering, similar topics can be grouped together to obtain corresponding topic clusters. Each topic cluster can represent a topic classification. This makes it easier to summarize the topics in the work order text data to be analyzed.

[0082] Step 107: semantically match the topic clusters to obtain the work order topics corresponding to each topic cluster.

[0083] After completing the clustering of the work order text data to be analyzed, semantic matching can be performed on each topic cluster according to the corresponding semantic words of each pre-set topic classification, so as to obtain the work order topic corresponding to each topic cluster.

[0084] The present invention discloses a method for analyzing work order topics, including: extracting effective information from work order text data to be analyzed by a representation learning algorithm; extracting characteristic key information from the effective information; performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation; extracting topic data from the work order text data to be analyzed by using characteristic word segmentation and a pre-trained topic discovery model; performing association analysis on the topic data to obtain associated topic data; clustering the associated topic data to obtain topic clusters; performing semantic matching on the clusters to obtain work order topics corresponding to each topic cluster. This improves the accuracy of work order topic classification, and facilitates subsequent problem warning, automatic dispatching, and hotline analysis of similar topics.

[0085] See also Figure 2 , Figure 2 This is a flowchart of a work order topic analysis method provided by another embodiment of the present invention. Specifically, the following steps may be included:

[0086] Step 201, determining redundant information in the work order text data to be analyzed by a representation learning algorithm;

[0087] Step 202, removing redundant information to obtain valid information in the work order text data to be analyzed;

[0088] In an embodiment of the present invention, redundant information in a large amount of work order text data to be analyzed can be removed through a representation learning algorithm, for example: meaningless information such as "department", "address", "call feedback" in the work order content description can be removed to obtain valid information related to the work order topic.

[0089] Step 203, extracting key feature information from the valid information;

[0090] Step 204, performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation;

[0091] Steps 203-204 are the same as steps 102-103. For details, please refer to the description of steps 102-103, which will not be repeated here.

[0092] Step 205, using feature segmentation and a pre-trained topic discovery model to extract topic data from the work order text data to be analyzed;

[0093] After obtaining the feature segmentation, the feature segmentation and the pre-trained topic discovery model can be used to extract the topic data in the work order text data to be analyzed.

[0094] In one example, step 205 may include the following sub-steps:

[0095] S51, use the semantic representation algorithm to encode the feature word segmentation to obtain the text vector;

[0096] S52, using a word embedding algorithm to map the text vector to a low-dimensional space to obtain a low-dimensional text vector;

[0097] S53, input the low-dimensional text vector into the pre-trained topic discovery model and output the topic data.

[0098] Word embedding is a general term for language models and representation learning techniques in natural language processing (NLP). Conceptually, it refers to embedding a high-dimensional space with the number of all words into a continuous vector space with a much lower dimension, where each word or phrase is mapped to a vector in the real number field.

[0099] In the specific implementation, the feature segmentation can be encoded using a meaning representation algorithm to obtain a text vector, and then the text vector can be mapped to a low-dimensional space using a word embedding algorithm to obtain a low-dimensional text vector that is easy to process, and then the low-dimensional text vector can be input into a pre-trained topic discovery model to obtain the topic data carried in the feature segmentation.

[0100] Step 206, performing correlation analysis on the topic data to obtain correlated topic data;

[0101] After the topic data in the feature segmentation is extracted through the topic discovery model, the obtained topic data can be subjected to a correlation analysis to obtain the related topic data.

[0102] In a specific implementation, the topic data can be analyzed for relevance through a knowledge graph. The relevance analysis can be the relevance between topics, the relevance between topics and information related to the topics, etc. The object of the relevance analysis can be selected according to the actual situation, and the embodiment of the present invention does not impose any specific restrictions on this.

[0103] In one example, the step of performing correlation analysis on topic data to obtain correlated topic data may include the following sub-steps:

[0104] S61, extracting the associated data between the topic data from the work order text data to be analyzed; the associated data includes the complaining subject, the responsible subject, the disposal measures, and the time and space location;

[0105] S62, generating a knowledge graph of topic data and associated data to obtain associated topic data.

[0106] In the specific implementation, topic correlation analysis can form a knowledge graph by correlating the topics involved in the work order event (including content, frequency, relationship), the complaining subject, the responsible subject, the handling measures, the time and space location, etc.

[0107] Among them, the entity types of the knowledge graph may include topics, responsible parties, appeal parties, handling departments, locations, etc.; the association relationship types may include topics and topics, topics and responsible parties, topics and appeal parties, topics and handling departments, topics and involved locations, etc.

[0108] See also Figure 3 , Figure 3 Schematic diagram of a knowledge graph based on a work order provided in an embodiment of the present invention. Figure 3 As shown, hot topics (such as safety hazards, nuisance, fire hazards, etc.) can be associated with complaining groups (such as students, migrant workers, low-income people), involved locations (such as Taihe Town, Dayuan Village), disposal measures (such as centralized rectification, policy notification, joint investigation and handling), disposal departments (such as the Municipal Supervision Bureau), and involved units (such as relevant enterprises, institutions, etc.), thereby establishing an associated information cluster of related topics, making it easier for staff to have a comprehensive understanding of the causes, processes, involved parties, disposal measures, disposal results, etc. related to the topics.

[0109] Each entity has a mapping relationship with the corresponding work order. Therefore, when analyzing a topic, the corresponding work order can be obtained based on the mapping relationship between the entity and the work order, so as to conduct in-depth analysis on the work order.

[0110] See also Figure 4 , Figure 4 A work order association schematic diagram provided for an embodiment of the present invention.

[0111] like Figure 4 As shown, Figure 4 The entities involved in the work order include: handling department, measure C1, appeal subject S1, location (customized regional granularity), topic H1, topic H3, responsible subject Z1, third-level item category, and second-level item category; their associated relationships may include: undertaken, occurred in, included, taken, reflected, occurred, involved, covered, proposed, belonged to, and handled attribution, etc.

[0112] according to Figure 4 The work order association diagram shown can directly understand the handling department, request subject, responsible subject, etc. of the corresponding topic of the work order.

[0113] Step 207, clustering the related topic data to obtain topic clusters;

[0114] Through clustering, similar topics can be grouped together to obtain corresponding topic clusters. Each topic cluster can represent a topic classification. This makes it easier to summarize the topics in the work order text data to be analyzed.

[0115] Step 208: semantically match the topic clusters to obtain the work order topics corresponding to each topic cluster.

[0116] After completing the clustering of the work order text data to be analyzed, semantic matching can be performed on each topic cluster according to the corresponding semantic words of each pre-set topic classification, so as to obtain the work order topic corresponding to each topic cluster.

[0117] Furthermore, in the embodiment of the present invention, after the work order topics are extracted, hot topics may be screened, and the specific steps include:

[0118] S1, obtain multi-dimensional information of each work order topic, including ranking, number of requests, month-on-month, and ranking change;

[0119] S2, screens hot topics from work order topics based on ranking, number of requests, month-on-month and ranking changes.

[0120] Topic ranking: The ranking of the number of topics of appeal within the currently selected time and area, which can be ranked from large to small.

[0121] Number of requests: the cumulative number of work orders for the request topic within the currently selected time and area;

[0122] Month-on-month: The growth rate of the number of citizen demands in the current area and on this topic in the current selected time period compared with the previous period.

[0123] Ranking change: The ranking increase of the current selected time period compared with the previous period in the current regional scope and the citizen demands corresponding to this topic.

[0124] In the specific implementation, hot topics can be filtered from the work order topics based on the real-time ranking, number of requests, month-on-month, ranking changes, etc.

[0125] In addition, after filtering out hot topics, you can conduct a multi-dimensional analysis of the hot topics to form analysis results of different dimensions, so as to facilitate timely grasp of hot topics, topic details and topic work order trends related to people's livelihood demands. You can also perform conditional screening on the time, region and matter classification of the topic to achieve the effect of quickly responding to user-specified needs. At the same time, for a single topic, you can realize the linkage between the regional overview section and the topic location section to analyze the recent development trend and high-incidence areas of the demand topic to meet the further needs of users.

[0126] Furthermore, in an embodiment of the present invention, it is possible to monitor in real time whether each step in the work order topic analysis process has an abnormality; when an abnormality occurs, the abnormality type and abnormal content are determined according to the step where the abnormality occurs, and abnormal alarm information is generated in real time according to the abnormality type and abnormal content, and the abnormal alarm information is sent to the operation and maintenance personnel for early warning in the form of SMS, email, etc. It can be understood that the judgment of the specific abnormal step can be that each step corresponds to a data interface, and if the data interface fails to successfully transmit data (input or output), it is considered that the step corresponding to the data interface has an abnormality.

[0127] The present invention discloses a method for analyzing work order topics, including: extracting effective information from work order text data to be analyzed by a representation learning algorithm; extracting characteristic key information from the effective information; performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation; extracting topic data from the work order text data to be analyzed by using characteristic word segmentation and a pre-trained topic discovery model; performing association analysis on the topic data to obtain associated topic data; clustering the associated topic data to obtain topic clusters; performing semantic matching on the clusters to obtain work order topics corresponding to each topic cluster. This improves the accuracy of work order topic classification, and facilitates subsequent problem warning, automatic dispatching, and hotline analysis of similar topics.

[0128] See also Figure 5 , Figure 5 A schematic diagram of the structure of a work order topic analysis device provided by an embodiment of the present invention.

[0129] An embodiment of the present invention provides a work order topic analysis device, comprising:

[0130] An effective information extraction module 501 is used to extract effective information from the work order text data to be analyzed through a representation learning algorithm;

[0131] A feature key information extraction module 502 is used to extract feature key information from valid information;

[0132] The word segmentation module 503 is used to perform word segmentation processing on the characteristic key information to obtain characteristic word segmentations;

[0133] The topic data extraction module 504 is used to extract the topic data from the work order text data to be analyzed by using feature segmentation and a pre-trained topic discovery model;

[0134] A correlation analysis module 505 is used to perform correlation analysis on the topic data to obtain related topic data;

[0135] A clustering module 506 is used to cluster the related topic data to obtain topic clusters;

[0136] The semantic matching module 507 is used to perform semantic matching on the topic clusters to obtain the work order topics corresponding to each topic cluster.

[0137] In the embodiment of the present invention, the effective information extraction module 501 includes:

[0138] A redundant information determination submodule is used to determine redundant information in the work order text data to be analyzed through a representation learning algorithm;

[0139] The effective information extraction submodule is used to remove redundant information and obtain the effective information in the work order text data to be analyzed.

[0140] In the embodiment of the present invention, the topic data extraction module 504 includes:

[0141] The encoding submodule is used to encode the feature segmentation using a semantic representation algorithm to obtain a text vector;

[0142] The low-dimensional text vector acquisition submodule is used to map the text vector to a low-dimensional space using a word embedding algorithm to obtain a low-dimensional text vector;

[0143] The topic data output submodule is used to input the low-dimensional text vector into the pre-trained topic discovery model and output the topic data.

[0144] In the embodiment of the present invention, the correlation analysis module 505 includes:

[0145] The associated data extraction submodule is used to extract the associated data between the topic data from the work order text data to be analyzed; the associated data includes the complaining subject, the responsible subject, the disposal measures, and the time and space location;

[0146] The knowledge graph generation submodule is used to generate a knowledge graph of topic data and related data to obtain related topic data.

[0147] In an embodiment of the present invention, it also includes:

[0148] The multi-dimensional information acquisition module is used to obtain multi-dimensional information of each work order topic, including ranking, number of requests, month-on-month comparison, and ranking changes;

[0149] The hot topic screening module is used to screen hot topics from work order topics based on ranking, number of requests, month-on-month growth and ranking changes.

[0150] Further, in the embodiment, the entire operation process of the work order topic analysis device is monitored, that is, whether each module in the work order topic analysis device has an abnormality can be monitored in real time; when an abnormality occurs, the abnormal type and abnormal content are determined according to the module where the abnormality occurs, and abnormal alarm information is generated in real time according to the abnormal type and abnormal content, and the abnormal alarm information is sent to the operation and maintenance personnel for early warning in the form of text messages, emails, etc. It can be understood that the judgment of the specific abnormal module can be that each module corresponds to a data interface, and if the data interface fails to successfully transmit data (input or output), it is considered that the module corresponding to the data interface has an abnormality. It can be understood that the abnormal alarm information corresponding to the work order topic analysis device can also be a traffic light alarm signal, that is, each module of the work order topic analysis device is correspondingly provided with an indicator light, when the work order topic analysis device is operating normally, the indicator light is green, and when the work order topic analysis device is operating abnormally, the indicator light is red.

[0151] The work order topic analysis device of the embodiment of the present invention extracts effective information from the work order text data to be analyzed through a representation learning algorithm; extracts characteristic key information from the effective information; performs word segmentation processing on the characteristic key information to obtain characteristic word segmentation; uses characteristic word segmentation and a pre-trained topic discovery model to extract topic data from the work order text data to be analyzed; performs association analysis on the topic data to obtain related topic data; clusters the related topic data to obtain topic clusters; performs semantic matching on the clusters to obtain the work order topics corresponding to each topic cluster. This improves the accuracy of work order topic classification, and facilitates subsequent problem warning, automatic dispatching, and hotline analysis of similar topics.

[0152] See also Figure 6 The embodiment of the present invention further provides an electronic device, which includes a processor 601 and a memory 602; the memory 602 is used to store program codes and transmit the program codes to the processor 601; the processor 601 is used to execute the work order topic analysis method of the embodiment of the present invention according to the instructions in the program code.

[0153] The embodiment of the present invention further provides a computer-readable storage medium, which is used to store program codes, and the program codes are used to execute the work order topic analysis method of the embodiment of the present invention.

[0154] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0155] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0156] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0157] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0158] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0159] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0160] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0161] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0162] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A work order topic analysis method, characterized in that: include: Extract valid information from the work order text data to be analyzed through a representation learning algorithm, the valid information includes: topic content, relationship, complaint subject, responsible subject, disposal measures and time and space location; Extracting characteristic key information from the effective information; Performing word segmentation processing on the characteristic key information to obtain characteristic word segmentation; Using the feature segmentation and pre-trained topic discovery model, extracting topic data from the work order text data to be analyzed; Performing correlation analysis on the topic data to obtain related topic data; Clustering the related topic data to obtain topic clusters; Perform semantic matching on the topic clusters to obtain the work order topics corresponding to each topic cluster; The step of performing correlation analysis on the topic data to obtain correlated topic data includes: Extracting the associated data between the topic data from the work order text data to be analyzed; the associated data includes the complaining subject, the responsible subject, the disposal measures, and the time and space location; Generate a knowledge graph of the topic data and the associated data to obtain associated topic data.

2. The method according to claim 1, characterized in that: The step of extracting effective information from the work order text data to be analyzed by using a representation learning algorithm includes: Determine redundant information in the work order text data to be analyzed by a representation learning algorithm; The redundant information is eliminated to obtain valid information in the work order text data to be analyzed.

3. The method according to claim 1, characterized in that The step of extracting topic data from the work order text data to be analyzed by using the feature segmentation and the pre-trained topic discovery model includes: Encode the feature word segmentation using a semantic representation algorithm to obtain a text vector; Mapping the text vector to a low-dimensional space using a word embedding algorithm to obtain a low-dimensional text vector; The low-dimensional text vector is input into a pre-trained topic discovery model to output topic data.

4. The method according to claim 1, characterized in that: Also includes: Obtain multi-dimensional information of each work order topic, including ranking, number of requests, month-on-month comparison, and ranking change; Hot topics are filtered from the work order topics according to the ranking, the number of requests, the month-on-month change and the ranking change.

5. A work order topic analysis device, characterized in that: include: An effective information extraction module is used to extract effective information from the work order text data to be analyzed through a representation learning algorithm. The effective information includes: topic content, relationship, complaint subject, responsible subject, disposal measures, and time and space location; A feature key information extraction module, used to extract feature key information from the valid information; A word segmentation module is used to perform word segmentation processing on the characteristic key information to obtain characteristic word segmentations; A topic data extraction module, used to extract topic data from the work order text data to be analyzed by using the feature segmentation and the pre-trained topic discovery model; A correlation analysis module, used to perform correlation analysis on the topic data to obtain related topic data; A clustering module, used for clustering the related topic data to obtain topic clusters; A semantic matching module, used to perform semantic matching on the topic clusters to obtain the work order topics corresponding to each topic cluster; Correlation analysis module, including: A related data extraction submodule is used to extract related data between the topic data from the work order text data to be analyzed; the related data includes the complaining subject, the responsible subject, the disposal measures, and the time and space location; The knowledge graph generation submodule is used to generate a knowledge graph of the topic data and the associated data to obtain associated topic data.

6. The device according to claim 5, characterized in that The effective information extraction module comprises: A redundant information determination submodule, used to determine the redundant information in the work order text data to be analyzed by using a representation learning algorithm; The effective information extraction submodule is used to remove the redundant information and obtain the effective information in the work order text data to be analyzed.

7. The device according to claim 5, characterized in that The topic data extraction module comprises: An encoding submodule, used for encoding the feature segmentation words using a semantic representation algorithm to obtain a text vector; A low-dimensional text vector acquisition submodule is used to map the text vector to a low-dimensional space using a word embedding algorithm to obtain a low-dimensional text vector; The topic data output submodule is used to input the low-dimensional text vector into the pre-trained topic discovery model and output the topic data.

8. An electronic device, characterized in that: The device comprises a processor and a memory; The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the work order topic analysis method according to any one of claims 1-4 according to the instructions in the program code.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the work order topic analysis method according to any one of claims 1 to 4.

Images