Hotspot work order identification method and device, equipment, medium and program product

Abstract

The invention provides a hotspot work order identification method which can be applied to the technical field of intelligent customer service, the technical field of artificial intelligence and the technical field of big data, and the hotspot work order identification method comprises the steps that basic attribute information of at least one to-be-identified work order is determined, and the basic attribute information at least comprises a work order category; based on the work order category, primary work orders related to the work order to be identified are screened from historical work orders; calculating the similarity between the primary selection work orders, so as to screen out optimal work orders related to the work order to be identified from the primary selection work orders; clustering processing is carried out on a work order set composed of the work orders to be identified and the optimized work orders; and in response to a clustering processing result, under the condition that the number of the work orders in the target class cluster to which the to-be-identified work order belongs meets a preset numerical value condition, determining that the to-be-identified work order is a hotspot work order. The invention further provides a hotspot work order recognition device and equipment, a storage medium and a program product.

Description

Technical Field

[0001] The present disclosure relates to the fields of intelligent customer service technology, artificial intelligence technology, and big data technology, and more specifically, to a hot work order identification method, device, equipment, medium, and program product. Background Art

[0002] With the development and progress of society, users have higher and higher expectations and requirements for various service departments. When encountering problems, users expect to obtain services or solve problems more conveniently and efficiently. As a direct communication channel, the service hotline provides a platform for users to reflect their demands. Service departments can understand users' concerns and hot issues through the hotline, so as to quickly locate and classify users' needs, and provide more timely and effective solutions.

[0003] However, the problems reported by users may involve multiple fields or multiple scenarios. The diversity and complexity of user demands make it difficult to accurately classify and identify current work orders. In addition, the ambiguity and uncertainty of language expression increase the difficulty of work order identification. It can be seen that there are still many problems to be solved in the refinement of work order identification to further improve the quality and efficiency of services. Summary of the invention

[0004] In view of this, the present disclosure provides a hot work order identification method, apparatus, device, medium and program product.

[0005] One aspect of the present disclosure provides a method for identifying hot work orders, including: determining basic attribute information of at least one work order to be identified, wherein the basic attribute information includes at least a work order category; based on the work order category, screening out preliminary work orders related to the work order to be identified from historical work orders; calculating the similarity between each preliminary work order to screen out preferred work orders related to the work order to be identified from the preliminary work orders; clustering a work order set consisting of the work order to be identified and the preferred work orders; and, in response to the result of the clustering processing, determining that the work order to be identified is a hot work order when the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition.

[0006] According to an embodiment of the present disclosure, hot work order identification also includes: determining the target task category of the work order to be identified based on the results of clustering processing; and, based on the target task category, determining the target task execution strategy corresponding to the target task category from the preferred task execution strategies corresponding to the preferred work orders.

[0007] According to an embodiment of the present disclosure, the similarities between various preliminary work orders are calculated to screen out preferred work orders related to the work order to be identified from the preliminary work orders, including: determining a first keyword of the preliminary work order based on the preliminary work order text corresponding to the preliminary work order; determining the preliminary work order corresponding to the first keyword as a candidate work order; calculating the similarities between various candidate work orders; and determining a preferred work order related to the work order to be identified from the candidate work orders based on the similarities between the candidate work orders.

[0008] According to an embodiment of the present disclosure, based on the preliminary work order text corresponding to the preliminary work order, determining the first keyword of the preliminary work order includes: extracting multiple historical keywords from the historical work order text corresponding to the historical work order; calculating the historical reference weights of the historical keywords; extracting multiple initial keywords from the preliminary work order text; based on the historical reference weights, respectively calculating the initial weights corresponding to the initial keywords; and determining the first keyword of the preliminary work order based on the initial weights.

[0009] According to an embodiment of the present disclosure, the initial weights corresponding to each initial keyword are calculated based on the historical reference weights, including: when the correlation between the initial keyword and at least one historical keyword is greater than a preset correlation threshold, the initial weight corresponding to the initial keyword is calculated according to the proportion of at least one historical reference weight in the preliminary work order to which the initial keyword belongs; when the correlation between the initial keyword and each historical keyword is less than or equal to the preset correlation threshold, the initial weight corresponding to the initial keyword is calculated according to the historical reference weight and the similarity between the historical keyword corresponding to the historical reference weight and the initial keyword.

[0010] According to an embodiment of the present disclosure, a work order set consisting of work orders to be identified and preferred work orders is clustered, including: inputting the work order text to be identified corresponding to the work order to be identified into a semantic vector generation model to generate a first semantic vector; inputting the preferred work order texts corresponding to multiple preferred work orders into the semantic vector generation model to generate multiple second semantic vectors; and clustering the first semantic vector and the multiple second semantic vectors.

[0011] According to an embodiment of the present disclosure, a semantic vector generation model is trained by the following method: obtaining multiple groups of similar sample pairs and first sample labels of similar sample pairs; obtaining multiple groups of non-similar sample pairs and second sample labels of non-similar sample pairs; inputting at least one group of similar sample pairs into an initial model, and outputting a first similarity of similar sample pairs; inputting at least one group of non-similar sample pairs into the initial model, and outputting a second similarity of non-similar sample pairs; adjusting model parameters of the initial model according to the first similarity, the first sample label, and the second similarity, the second sample label, to obtain a semantic vector generation model.

[0012] According to an embodiment of the present disclosure, the initial model is a dual-branch model, which includes two first branch networks and second branch networks with the same network structure. At least one group of similar sample pairs is input into the initial model, and the first similarity of the similar sample pairs is output, including: inputting the first sample in the similar sample pair into the first branch network to generate a first sample semantic vector corresponding to the first sample; inputting the second sample in the similar sample pair into the second branch network to generate a second sample semantic vector corresponding to the second sample; and calculating the first sample similarity between the first sample semantic vector and the second sample semantic vector.

[0013] Another aspect of the present disclosure provides a hot work order identification device, including: a first determination module, used to determine basic attribute information of at least one work order to be identified, wherein the basic attribute information at least includes the work order category; a screening module, used to screen out preliminary work orders related to the work order to be identified from historical work orders based on the work order category; a calculation module, used to calculate the similarity between each preliminary work order, so as to screen out preferred work orders related to the work order to be identified from the preliminary work orders; a clustering module, used to cluster a work order set consisting of the work order to be identified and the preferred work orders; a second determination module, used to determine that the work order to be identified is a hot work order in response to the result of the clustering processing, when the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition.

[0014] Another aspect of the present disclosure provides an electronic device, including: one or more processors; and 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 implement the above method.

[0015] Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the above method when executed.

[0016] Another aspect of the present disclosure provides a computer program product, which includes computer executable instructions, and the instructions are used to implement the above method when executed.

[0017] According to the embodiments of the present disclosure, by obtaining the work order category of the work order to be identified, work orders with relevance in category can be roughly screened out from historical work orders according to the work order category, without comparing all historical work orders with the work order to be identified one by one, which can quickly narrow the scope of data calculation, reduce unnecessary analysis and processing, save time and resources, and improve the overall efficiency of work order processing. Further, in order to facilitate the provision of a more efficient and fast solution, starting from the characteristics of hot work orders (work orders of similar events that occur frequently within a certain period of time), preferred work orders with high similarity between two work orders are determined from the preliminary work orders. Specifically, on the basis of the similarity of the work order category, the similarity between each preliminary work order is calculated, so that each preferred work order is highly similar, and the work order that is semantically inconsistent with the work order to be identified is excluded from the preliminary work order, further narrowing the scope of data calculation. The task category in the historical work order is used to provide a reference for the new work order to be identified, avoiding repeated exploration and errors. Compared with the manual identification method, making decisions based on data and historical experience, rather than relying solely on subjective judgment, makes the decision more scientific and reasonable. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above 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:

[0019] Figure 1 An exemplary system architecture 100 to which a hotspot work order identification method according to an embodiment of the present disclosure can be applied is schematically shown;

[0020] Figure 2 A flowchart of a hotspot work order identification method according to an embodiment of the present disclosure is schematically shown;

[0021] Figure 3 A flowchart of a method for determining a preferred work order according to an embodiment of the present disclosure is schematically shown;

[0022] Figure 4 A schematic diagram schematically shows an initial model according to an embodiment of the present disclosure;

[0023] Figure 5 A schematic diagram of a method for determining a target task execution strategy according to an embodiment of the present disclosure is schematically shown;

[0024] Figure 6 A block diagram schematically shows a hotspot work order identification device according to an embodiment of the present disclosure; and

[0025] Figure 7 A block diagram of an electronic device suitable for implementing the hotspot work order identification method described above according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0026] 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.

[0027] 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 the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0028] 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.

[0029] 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.).

[0030] In the embodiments of the present disclosure, the collection, updating, analysis, processing, use, transmission, provision, disclosure, storage, etc. of the data involved (for example, including but not limited to user personal information) are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information and network security.

[0031] In the embodiments of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0032] As an important part of the workflow, hot work orders require more accurate and efficient identification methods to ensure the smooth progress of the business, so that the service department can provide corresponding problem-solving strategies in a timely manner according to the classification of hot work orders. For example, in the e-commerce industry, thousands of work orders are generated every day. Merchants need to quickly and accurately identify the categories of hot work orders that need to be processed, such as returns and exchanges, customer complaints, etc., so as to provide users with problem-solving strategies in a timely manner based on the identification results. For example, in the logistics industry, accurate identification of abnormalities in cargo transportation or problem categories of hot work orders can adjust transportation plans in a timely manner to avoid unnecessary delays and losses. Another example is the complaint handling of the service hotline. The service department needs to capture high-frequency events in a short period of time based on the hotline work orders, identify events that are closely related to residents' lives or may endanger the safety of residents' lives and property, so as to help the service department better understand the needs and concerns of the public, and be able to respond to emergencies in a timely manner to improve public satisfaction and trust.

[0033] Existing methods for identifying hot ticket categories often rely on manual judgment or are based on fixed rules. The drawback of manually answering the phone to determine the category of a ticket is that different manual customer service representatives may have different understandings and judgments of the same problem, resulting in inconsistent ticket classifications, and especially when the number of calls is large, it is easy to cause a backlog of tickets and processing delays. The drawback of identifying ticket categories based on fixed rules is that it cannot adapt to new types of problems or special situations, resulting in some tickets being unable to be accurately classified; fixed rules may not be able to make accurate judgments for tickets that are not clearly stated, complete, or ambiguous; relying solely on rules cannot understand the deep semantics and contextual information in the ticket, which may lead to classification errors.

[0034] In view of this, an embodiment of the present disclosure provides a hot work order identification method, including: determining basic attribute information of at least one work order to be identified, wherein the basic attribute information includes at least the work order category; based on the work order category, screening out preliminary work orders related to the work order to be identified from historical work orders; calculating the similarity between each preliminary work order to screen out preferred work orders related to the work order to be identified from the preliminary work orders; clustering the work order set consisting of the work order to be identified and the preferred work orders; and, in response to the result of the clustering processing, determining that the work order to be identified is a hot work order when the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition.

[0035] Figure 1 The exemplary system architecture 100 to which the hotspot work order identification method according to an embodiment of the present disclosure can be applied is schematically shown. It should be noted that: Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0036] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0037] The user may use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software, etc. (only for example).

[0038] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0039] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 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.

[0040] It should be noted that the hotspot work order identification method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the hotspot work order identification system provided in the embodiment of the present disclosure can generally be set in the server 105. The hotspot work order identification method 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 first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Correspondingly, the hotspot work order identification system 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 first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Alternatively, the hotspot work order identification method provided in the embodiment of the present disclosure can also be executed by the first terminal device 101, the second terminal device 102 or the third terminal device 103, or can also be executed by other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103. Accordingly, the hotspot work order identification system provided in the embodiment of the present disclosure may also be set in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0041] For example, the work order to be identified and the historical work order may be originally stored in any one of the first terminal device 101, the second terminal device 102 or the third terminal device 103 (for example, the first terminal device 101, but not limited thereto), or stored on an external storage device and may be imported into the first terminal device 101. Then, the first terminal device 101 may locally execute the hot work order identification method provided in the embodiment of the present disclosure, or send the work order to be identified and the historical work order to other terminal devices, servers, or server clusters, and the other terminal devices, servers, or server clusters that receive the initial data information may execute the hot work order identification method provided in the embodiment of the present disclosure.

[0042] 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.

[0043] Figure 2 A flowchart of a hotspot work order identification method according to an embodiment of the present disclosure is schematically shown.

[0044] like Figure 2 As shown, the hot work order identification method includes operations S210~S250.

[0045] In operation S210, basic attribute information of at least one work order to be identified is determined, wherein the basic attribute information at least includes a work order category.

[0046] According to an embodiment of the present disclosure, the work order to be identified may be a request from a user received at a certain time point or within a certain time range, for example, a telephone consultation from a user received through a telephone hotline.

[0047] According to an embodiment of the present disclosure, the work order category of at least one work order to be identified is determined. Specifically, the staff can establish an index for the work order to be identified by labeling according to the predefined work order category. Among them, the predefined work order category can be formulated according to relevant policies or relevant standards, or can be customized based on experience. Predefined task categories in public service scenarios include, for example, life services, education services, medical and health services, social security services, etc.; predefined task categories in security complaint scenarios include, for example, fire safety, public security management, food safety, etc.; predefined task categories in traffic consultation scenarios include, for example, traffic accidents, road congestion, damage to traffic facilities, etc.

[0048] According to an embodiment of the present disclosure, the present disclosure does not limit the hierarchy of the predefined work order category, which may be the life service category in the above example, or a subordinate category of the life service category, such as water outage, power outage, or network outage.

[0049] In operation S220 , based on the work order category, preliminary work orders related to the work order to be identified are screened from historical work orders.

[0050] According to an embodiment of the present disclosure, when acquiring historical work orders, an index is established for each historical work order by marking according to predefined work order categories.

[0051] According to an embodiment of the present disclosure, the work order to be identified may be the same as the work order category of the work order to be identified, or the work order category is similar to the work order to be identified. Taking the same work order category as the work order to be identified as an example, if the work order category of the work order to be identified is power outage, a part of the historical work orders with the work order category of power outage are screened out from the historical work orders and determined as the preliminary work orders. If there are multiple work orders to be identified, for example, the work order category of another work order to be identified is road congestion, another part of the historical work orders with the work order category of road congestion are screened out from the historical work orders and determined as another part of the preliminary work orders.

[0052] In operation S230 , the similarities between the preliminary selected work orders are calculated to screen out preferred work orders related to the work order to be identified from the preliminary selected work orders.

[0053] According to an embodiment of the present disclosure, the similarity between each preliminary work order is calculated, and the preferred work order related to the work order to be identified is screened out from the preliminary work orders according to the screening conditions. Specifically, the screening conditions may, for example, set a similarity threshold, and when the similarity between two preliminary work orders is greater than or equal to the similarity threshold, at least two work orders greater than or equal to the similarity threshold are determined as preferred work orders; or, for another example, the number of preferred work orders is set, and the preliminary work orders that meet the quantity requirements are determined as preferred work orders in descending order of similarity.

[0054] In operation S240 , clustering is performed on a work order set consisting of the work orders to be identified and the preferred work orders.

[0055] According to an embodiment of the present disclosure, the work order to be identified and multiple preferred work orders are combined into a work order set, and the work order set is clustered. The embodiment of the present disclosure does not specifically limit the clustering processing method, which can be a density-based clustering method, a level-based clustering method, a model-based clustering method, etc.

[0056] In operation S250 , in response to a result of clustering processing, if the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition, it is determined that the work order to be identified is a hot work order.

[0057] According to the embodiment of the present disclosure, after clustering processing, the classification results of each work order in the work order set and the target cluster to which the work order to be identified belongs are obtained.

[0058] According to an embodiment of the present disclosure, in response to the result of clustering processing, the clusters of each work order and the target cluster to which the work order to be identified belongs are obtained. Further, it is determined whether the number of work orders in the target cluster to which the work order to be identified belongs is greater than or equal to the reserved number. If so, it is determined that all the work orders in the target cluster to which the work order to be identified belongs are hot work orders, and the work order to be identified is a hot work order; if not, the opposite is true.

[0059] According to the embodiments of the present disclosure, by obtaining the work order category of the work order to be identified, the work orders with relevance in category can be roughly screened out from the historical work orders according to the work order category, without comparing all the historical work orders with the work order to be identified one by one, which can quickly narrow the scope, reduce unnecessary analysis and processing, save time and resources, and improve the overall efficiency of work order processing. Further, in order to facilitate the provision of a more efficient and fast solution, starting from the characteristics of hot work orders (work orders of similar events that occur frequently within a certain period of time), the preferred work orders with high similarity between two work orders are determined from the preliminary work orders. Specifically, on the basis of the similarity of the work order category, the similarity between each preliminary work order is calculated, so that the obtained preferred work orders are highly similar, and the work orders that are inconsistent with the semantics of the work order to be identified are excluded from the preliminary work orders, further narrowing the data calculation scope. The task category in the historical work order is used to provide a reference for the new work order to be identified, avoiding repeated exploration and errors. Compared with the manual identification method, making decisions based on data and historical experience, rather than relying solely on subjective judgment, makes the decision more scientific and reasonable.

[0060] According to an embodiment of the present disclosure, the basic attribute information of the work order to be identified may also include a time attribute and a location attribute, wherein the time attribute is used to characterize the time information of the event in the work order to be identified, and the location attribute is used to characterize the location information of the event in the work order to be identified. Similarly, the basic attribute information of each historical work order may also include a time attribute and a location attribute.

[0061] According to an embodiment of the present disclosure, three screening conditions are set according to the work order category, time attribute and location attribute of the work order to be identified, and preliminary work orders that simultaneously meet the above three screening conditions are screened out from historical work orders.

[0062] According to an embodiment of the present disclosure, the screening condition of the time attribute may be to determine the historical work orders that occurred at a certain time point as the preliminary work orders; or to determine the historical work orders corresponding to events that occurred within a certain time period as the preliminary work orders. For example, the historical work orders that occurred within 5 minutes before and after the time when the event in the work order to be identified occurred are determined as the preliminary work orders.

[0063] According to an embodiment of the present disclosure, the screening condition of the location attribute may be to determine the historical work orders that occurred at a certain location as the preliminary work orders; or to determine the historical work orders corresponding to the time that occurred in the adjacent geographical unit as the preliminary work orders, wherein the adjacent geographical unit may be defined according to the administrative area unit or according to the distance. For example, the same street as the location where the event in the work order to be identified occurred, or a range within 5 kilometers with the location where the event in the work order to be identified occurred as the center, etc.

[0064] According to the embodiments of the present disclosure, by combining the work order category, time attribute and location attribute for screening, it is possible to more accurately find historical work orders that are highly relevant to the work order to be identified. It is ensured that the screened preliminary work orders are similar to the current work order to be processed in multiple key dimensions, thereby improving the pertinence of subsequent analysis and strategy formulation.

[0065] According to the embodiments of the present disclosure, by filtering by time attributes, focusing on event work orders that occurred within a specific time range, the latest and most relevant processing experience can be timely learned. Ensure that the strategies and measures adopted are adapted to the current situation and improve the timeliness and effectiveness of processing. For example, when handling an emergency water outage, referring to the recent processing methods in the same time period can better cope with possible changes.

[0066] According to the embodiments of the present disclosure, considering location attributes helps to formulate more practical task execution strategies based on the characteristics and resource conditions of different regions. Different regions may have differences in infrastructure, environmental conditions, etc., so filtering work orders for specific locations can better adapt to local conditions. Assuming that water outages frequently occur in an old community, location filtering can be used to focus on analyzing the special conditions of the area and formulate special solutions.

[0067] According to an embodiment of the present disclosure, the hotspot work order identification method further includes operations S260 to S270.

[0068] In operation S260 , a target task category of the work order to be identified is determined according to the result of the clustering process.

[0069] According to an embodiment of the present disclosure, the task category is a subordinate category of the work order category, for example, it can be a subcategory of the work order category. There is a one-to-one correspondence between the task category and the cluster, that is, one cluster corresponds to one task category, and the target cluster to which the work order to be identified belongs corresponds to the target task category.

[0070] According to an embodiment of the present disclosure, a work order category may correspond to one or more task categories, wherein the task category may be used to characterize the cause of an event in a work order, or may be a subordinate category of the work order category. Exemplarily, the work order category is a life service category, and the task category may be a subcategory of the work order category, such as a task category of power outage. Furthermore, there are many reasons for power outages, such as line aging or damage, power load overload, animal contact, transformer failure, switch cabinet failure, power grid upgrade and transformation, and so on. Therefore, the task category may also be a subordinate category of the power outage, that is, there is a characterization of the cause of the occurrence of a specific task.

[0071] According to an embodiment of the present disclosure, based on the result of clustering processing, the target cluster to which the work order to be identified belongs is determined as the target task category of the work order to be identified.

[0072] In operation S270 , according to the target task category, a target task execution strategy corresponding to the target task category is determined from preferred task execution strategies corresponding to the preferred work orders.

[0073] According to an embodiment of the present disclosure, each task category corresponds to a task execution strategy, wherein the task execution strategy may be determined based on historical work orders, or may not be in the historical work orders but is configured in advance based on preset conditions.

[0074] According to the embodiments of the present disclosure, different task execution strategies correspond to task categories corresponding to different reasons. Exemplarily, different task execution strategies should be adopted for power outages caused by different reasons. For aging or damage of lines, replace aging and damaged lines; for power overload, change power dispatch and power resource allocation; for animal contact, install protective nets or bird repellents on line sections prone to contact; for equipment failure, maintain and inspect the equipment; for power grid upgrades and renovations, inform users of the specific power outage time and scope, and try to shorten the construction time and speed up the renovation progress to reduce the impact on users; and so on.

[0075] According to the embodiment of the present disclosure, the task execution strategy corresponding to the target task category is determined as the target task execution strategy, so as to provide guidance for solving problems to the staff in a timely manner and respond to user needs in a timely manner. The successful processing strategy in the historical work order is used to provide a reference for the new work order to be identified, avoiding repeated exploration and errors.

[0076] Figure 3 A flow chart of a method for determining a preferred work order according to an embodiment of the present disclosure is schematically shown.

[0077] like Figure 3 As shown, the similarities between the preliminary selected work orders are calculated to select preferred work orders related to the work order to be identified from the preliminary selected work orders, including operations S310 to S340.

[0078] In operation S310 , a first keyword of the preliminary work order is determined based on a preliminary work order text corresponding to the preliminary work order.

[0079] According to the embodiments of the present disclosure, considering that event descriptions often involve subjective analysis and contain a lot of noise information, such as adjectives or sentiment words, we need to give priority to the similarity of the core words of the sentence, that is, the similarity of the keywords, when generating the semantic vector. Specifically, based on the preliminary work order text corresponding to the preliminary work order, the first keyword of the preliminary work order is determined, and the noise information irrelevant to the semantics is excluded.

[0080] According to an embodiment of the present disclosure, the preliminary work order text corresponding to each preliminary work order is preprocessed to obtain the initial keywords corresponding to each preliminary work order text. The weight of each initial keyword is calculated, and the first keyword corresponding to each initial work order text is determined according to the keyword screening condition. Among them, the keyword screening condition can be to set a weight threshold, and determine the initial keyword with a weight greater than or equal to the weight threshold as the first keyword; or, set the number of keywords, sort them from large to small according to the weight, and determine the first keyword according to a predetermined number of keywords, etc.

[0081] In operation S320 , the preliminary selected work order corresponding to the first keyword is determined as a candidate work order.

[0082] According to an embodiment of the present disclosure, after determining the first keyword in the preliminary work order text, the preliminary work order to which the first keyword belongs is determined as a candidate work order, wherein each candidate work order contains at least one first keyword.

[0083] According to the embodiments of the present disclosure, by determining the first keyword, keywords with higher weights and corresponding candidate work orders are screened out from the preliminary work orders, that is, the interference caused by noise information is eliminated, and the accuracy of subsequent similarity calculation is improved. By calculating the keyword weight, the process of semantic filtering is realized, and those semantically irrelevant events are removed to refine the candidate work order list and improve the accuracy of work order recognition.

[0084] In operation S330 , the similarities between the candidate work orders are calculated.

[0085] According to an embodiment of the present disclosure, the first similarity of each candidate work order is calculated based on the word vector matrix of the first keyword corresponding to each candidate work order. The similarity calculation method is not limited here.

[0086] In operation S340 , a preferred work order related to the work order to be identified is determined from the candidate work orders according to the similarities between the candidate work orders.

[0087] According to an embodiment of the present disclosure, by comparing the similarity of two work orders to be identified, a preferred work order with higher similarity is determined from the candidate work orders. Specifically, the candidate work orders with a similarity threshold greater than or equal to the similarity threshold may be determined as the preferred work order. Alternatively, the number of preferred work orders may be set.

[0088] According to an embodiment of the present disclosure, based on the preliminary work order text corresponding to the preliminary work order, the first keyword of the preliminary work order is determined, which specifically includes steps 11 to 15.

[0089] In step 11, a plurality of historical keywords are extracted from the historical work order text corresponding to the historical work order.

[0090] According to an embodiment of the present disclosure, based on a plurality of historical work orders, a plurality of historical keywords are extracted from the historical work order texts corresponding to the historical work orders, and a vocabulary consisting of the plurality of historical keywords is constructed.

[0091] In step 12, the historical reference weights of the historical keywords are calculated.

[0092] According to an embodiment of the present disclosure, the historical reference weight of each historical keyword is calculated according to the number of historical work order texts in which the historical keyword appears and the total amount of historical work order texts, referring to formula (1):

[0093] Formula (1)

[0094] in, The weight of each historical keyword, i.e., the historical reference weight; The number of historical ticket texts where the historical keyword appears. The total amount of historical work order text.

[0095] In step 13, a plurality of initial keywords are extracted from the preliminary work order text.

[0096] According to an embodiment of the present disclosure, the preliminary work order text related to the preliminary work order is preprocessed, including removing irrelevant characters, word segmentation, removing stop words, etc., to obtain a word sequence. Using a preset word vector dictionary, the word sequence of the preliminary work order text is converted into a preliminary word vector matrix.

[0097] According to an embodiment of the present disclosure, keywords to be identified are extracted from work order text to be identified that is related to the work order to be identified.

[0098] According to an embodiment of the present disclosure, the work order text to be identified is preprocessed, including removing irrelevant characters, word segmentation, and stop words, to obtain a word sequence. The word sequence of the work order text to be identified is converted into a word vector matrix to be identified using a preset word vector dictionary.

[0099] In step 14, based on the historical reference weights, the initial weights corresponding to the initial keywords are calculated respectively, including the following two cases:

[0100] Case (1): When the correlation between the initial keyword and at least one historical keyword is greater than a preset correlation threshold, the initial weight corresponding to the initial keyword is calculated based on the proportion of at least one historical reference weight in the preliminary work order text to which the initial keyword belongs.

[0101] According to an embodiment of the present disclosure, the correlation between the initial keyword and at least one historical keyword can be calculated based on the similarity. The preset correlation threshold can be set based on experience, such as 75%, 85%, 95%, 100%, etc. That is, when the correlation between the initial keyword and at least one historical keyword is greater than the preset correlation threshold, the initial weight corresponding to the initial keyword is calculated using formula (2): .

[0102] Formula (2)

[0103] in, The length of the initial work order text.

[0104] According to an embodiment of the present disclosure, when the initial keyword is the same as at least one historical keyword, the initial weight corresponding to the initial keyword can be calculated using formula (2): .

[0105] Case (2): When the correlation between the initial keyword and each historical keyword is less than or equal to the preset correlation threshold, the initial weight corresponding to the initial keyword is calculated according to the historical reference weight and the similarity between the historical keyword corresponding to the historical reference weight and the initial keyword.

[0106] According to an embodiment of the present disclosure, when the correlation between the initial keyword and at least one historical keyword is less than or equal to a preset correlation threshold, the initial weight corresponding to the initial keyword is calculated using formula (3): .

[0107] Formula (3)

[0108] in, is the total number of initial keywords that appear in the initial ticket text, It is the similarity between the historical keywords and the initial keywords.

[0109] According to an embodiment of the present disclosure, when the initial keyword is different from at least one historical keyword, the initial weight corresponding to the initial keyword can be calculated using formula (3): .

[0110] In step 15, based on the initial weight, the first keyword of the preliminary work order is determined.

[0111] According to an embodiment of the present disclosure, based on the initial weight, according to the preset number of first keywords, or according to the weight threshold, the first keyword related to the execution strategy of the preliminary task in the preliminary work order is determined. Specifically, the initial keywords are sorted according to the initial weight, and according to the preset number of first keywords, the keywords with the highest ranking are determined as the first keywords corresponding to the preliminary work order text. Alternatively, a weight threshold is set, and the keywords that meet the threshold are determined as the first keywords corresponding to the preliminary work order text.

[0112] According to the embodiments of the present disclosure, the traditional algorithm for calculating keyword weights needs to recalculate the weights for all words every time a new document is added. This increases the amount of data calculation and the time consumed for data calculation. The method for calculating the first keyword weight in the embodiments of the present disclosure can quickly and simultaneously calculate the weights of each initial keyword, thereby reducing the data waiting time, improving data calculation efficiency, and improving the efficiency of processing and analyzing text data.

[0113] According to the embodiments of the present disclosure, the existing similarity calculation uses a semantic learning network to convert the original text into a vector representation to extract semantic information and generate a semantic vector, and uses the similarity of two work order vector representations to measure similar work orders. However, the work order text is the reporter's (hotspot caller) description of the event, which is affected by subjective factors and contains a large number of non-keywords, resulting in the semantic vector learned by the unsupervised network not being effective in judging text similarity. At the same time, similar events are not labeled in the data set of historical work orders, so it is impossible to use supervised training to learn semantic vectors to improve the ability to measure the distance or correlation between texts.

[0114] To solve the above problems, the embodiments of the present disclosure adopt a self-supervision method. Specifically, sample pairs for comparative training are generated based on existing work orders (historical work orders), and then a more efficient semantic learning network is trained using a comparative learning method. This new semantic learning network can map text to a semantic vector space with higher discrimination, thereby improving the ability to compare text similarity.

[0115] Specifically, multiple groups of similar sample pairs and first sample labels of similar sample pairs are obtained based on historical work order texts, wherein the historical work order texts are used as the first samples in the similar sample pairs, and the samples generated by the sample pair generation network are used as the second samples in the similar sample pairs.

[0116] Among them, the second sample for training the sample pair generation network is generated by data enhancement technology. Specifically, obtain the historical work order text; randomly shield some words in the historical work order text with a certain probability to obtain training samples, wherein the shielding method can be to delete some words or modify some words; use the training samples as the input of the model, and use the historical work order text as the output of the model to train the network to obtain the sample pair generation network. According to an embodiment of the present disclosure, based on the historical work order text, multiple groups of non-similar sample pairs and second sample labels of non-similar sample pairs are obtained. Specifically, two non-similar samples are randomly selected from the historical work order text as a group of non-similar sample pairs.

[0117] According to an embodiment of the present disclosure, an initial model is trained based on multiple groups of similar sample pairs and multiple groups of non-similar sample pairs to obtain a semantic vector generation model.

[0118] The first sample label is used to represent the similarity between similar samples, that is, y=1 represents two identical sample pairs. The second sample label is used to represent the similarity between non-similar samples, that is, y=0 represents two different sample pairs.

[0119] Figure 4 A schematic diagram of an initial model according to an embodiment of the present disclosure is schematically shown.

[0120] like Figure 4 As shown, the initial model is a double-branch model, which includes a first branch network and a second branch network with the same network structure.

[0121] Among them, the process of using similar sample pairs and non-similar sample pairs to train the initial model to obtain a semantic vector generation model, in addition to obtaining multiple groups of similar sample pairs and first sample labels of similar sample pairs; and obtaining multiple groups of non-similar sample pairs and second sample labels of non-similar sample pairs, also specifically includes operations S410 to S430.

[0122] In operation S410, at least one group of similar sample pairs is input into an initial model, and a first similarity of the similar sample pairs is output.

[0123] Specifically including step 21 to step 23.

[0124] In step 21, the first sample in the similar sample pair is input into the first branch network to generate a first sample semantic vector corresponding to the first sample.

[0125] In step 22, the second sample in the similar sample pair is input into the second branch network to generate a second sample semantic vector corresponding to the second sample.

[0126] In step 23, the first sample similarity between the first sample semantic vector and the second sample semantic vector is calculated. The vectors output by the two branch networks are concatenated together, and the result indicating whether they are similar is output through a "softmax layer".

[0127] In operation S420 , at least one group of non-similar sample pairs is input into the initial model, and a second similarity of the non-similar sample pairs is output.

[0128] Specifically including step 31 to step 33.

[0129] In step 31, the third sample in the non-similar sample pair is input into the first branch network to generate a third sample semantic vector corresponding to the third sample.

[0130] In step 32, the fourth sample in the non-similar sample pair is input into the second branch network to generate a fourth sample semantic vector corresponding to the fourth sample.

[0131] In step 33, a second sample similarity between the third sample semantic vector and the fourth sample semantic vector is calculated.

[0132] In operation S430, model parameters of the initial model are adjusted according to the first similarity and the first sample label, and the second similarity and the second sample label to obtain a semantic vector generation model.

[0133] According to an embodiment of the present disclosure, the initial model is trained using a contrast loss function, referring to formula (4):

[0134] Formula (4)

[0135] When y=1, it indicates that the first sample and the second sample of the input are the same sample pair, and when y=0, it indicates that the third sample and the fourth sample of the input are different sample pairs. The cosine distance of the output vector of the first branch network or the second branch network.

[0136] Furthermore, the trained semantic vector generation model is used to generate a first semantic vector corresponding to the work order to be identified and second semantic vectors corresponding to multiple preferred work orders, and clustering processing is performed, which specifically includes steps 41 to 43.

[0137] In step 41, the work order text to be identified corresponding to the work order to be identified is input into a semantic vector generation model to generate a first semantic vector.

[0138] In step 42, the preferred work order texts corresponding to the plurality of preferred work orders are input into a semantic vector generation model to generate a plurality of second semantic vectors.

[0139] In step 43, clustering is performed on the first semantic vector and the plurality of second semantic vectors.

[0140] According to an embodiment of the present disclosure, work orders are clustered based on the semantic vector corresponding to each work order. Exemplarily, a density-based clustering method, a hierarchical clustering method, a partition-based clustering method, or the like is adopted. Since the semantic vector of the work order and the keywords are obtained, work orders with similar task categories can be aggregated into the same category through clustering. Among them, the task category is used to characterize the semantics of the keyword. For example, if two work orders both describe the problem of garbage accumulation in a certain community, then the above two work orders may be aggregated into the same task category; if one work order is about garbage accumulation in the community and the other is about road damage, they may be aggregated into different task categories.

[0141] According to the embodiments of the present disclosure, clustering analysis of work order contents based on semantic vectors can accurately determine whether the work order to be identified and the preferred work order are events of the same task category. Furthermore, for work order times of the same task category, service personnel can promptly and accurately determine the target task execution strategy based on the corresponding task execution strategy in the preferred work order, thereby improving processing efficiency. At the same time, based on the clustering results, multiple work orders can also be processed centrally. For example, for multiple work orders reflecting the same road damage problem, maintenance personnel can be arranged to handle them at one time instead of processing each work order separately.

[0142] Figure 5 A schematic diagram of a method for determining a target task execution strategy according to an embodiment of the present disclosure is schematically shown.

[0143] like Figure 5 As shown, the method for determining the target task execution strategy includes operations S510 to S570.

[0144] In operation S510, based on basic attribute information of the work order to be identified, namely, the work order category, time attribute and location attribute, preliminary work orders that simultaneously meet the work order category condition, time condition and location condition are screened out from the historical work order pool.

[0145] In operation S520, pre-processing such as removing irrelevant characters, segmenting words, and removing stop words is performed on the preliminary selected work order text corresponding to the preliminary selected work order.

[0146] In operation S530, the weights of the various preliminary selected keywords in the preliminary selected work order text corresponding to the preliminary selected work order are calculated according to the historical reference weights, and the keywords that meet the predetermined conditions are determined as the first keywords, and the first keywords are used to represent the core keywords in the preliminary selected work order text that can represent the semantics of the sentence. The preliminary selected work order to which the first keyword belongs is determined as the candidate work order.

[0147] In operation S540, similarities between the candidate work orders are calculated to select preferred work orders with higher similarities from the candidate work orders according to a predetermined similarity condition.

[0148] In operation S550, a first semantic vector corresponding to the to-be-identified work order and a second semantic vector corresponding to the preferred work order are generated using the trained semantic vector generation model, and clustering processing is performed on the first semantic vector and the plurality of second semantic vectors;

[0149] In operation S560, in response to the result of the clustering process, if the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition, the work order to be identified is determined to be a hot work order, and at the same time, a target task category to which the work order to be identified belongs is determined;

[0150] In operation S570 , a target task execution strategy is determined according to the target task category to which the work order to be identified belongs.

[0151] Figure 6 A block diagram of a hotspot work order identification device according to an embodiment of the present disclosure is schematically shown.

[0152] like Figure 6 As shown, the hot work order identification device 600 includes a first determination module 610 , a screening module 620 , a calculation module 630 , a clustering module 640 and a second determination module 650 .

[0153] A first determination module 610 is used to determine basic attribute information of at least one work order to be identified, wherein the basic attribute information includes at least the work order category; a screening module 620 is used to screen out preliminary work orders related to the work order to be identified from historical work orders based on the work order category; a calculation module 630 is used to calculate the similarity between the preliminary work orders to screen out preferred work orders related to the work order to be identified from the preliminary work orders; a clustering module 640 is used to cluster a work order set consisting of the work order to be identified and the preferred work orders; and a second determination module 650 is used to determine, in response to the result of the clustering processing, that the work order to be identified is a hot work order when the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition.

[0154] According to an embodiment of the present disclosure, the hot work order identification device 600 also includes: a third determination module, used to determine the target task category of the work order to be identified based on the result of clustering processing; and a fourth determination module, used to determine the target task execution strategy corresponding to the target task category from the preferred task execution strategies corresponding to the preferred work orders based on the target task category.

[0155] According to an embodiment of the present disclosure, the calculation module 630 includes a first determination submodule, a second determination submodule, a calculation submodule and a third determination submodule: the first determination submodule is used to determine the first keyword of the preliminary work order based on the preliminary work order text corresponding to the preliminary work order; the second determination submodule is used to determine the preliminary work order corresponding to the first keyword as a candidate work order; the calculation submodule is used to calculate the similarity between each candidate work order; the third determination submodule is used to determine the preferred work order related to the work order to be identified from the candidate work orders based on the similarity between the candidate work orders.

[0156] According to an embodiment of the present disclosure, the first determining submodule includes: a first extracting unit, a first calculating unit, a second extracting unit, a second calculating unit and a determining unit.

[0157] The first extraction unit is used to extract multiple historical keywords from the historical work order text corresponding to the historical work order; the first calculation unit is used to calculate the historical reference weight of the historical keywords; the second extraction unit is used to extract multiple initial keywords from the preliminary work order text; the second calculation unit is used to calculate the initial weights corresponding to each initial keyword based on the historical reference weights; the determination unit is used to determine the first keyword of the preliminary work order based on the initial weight.

[0158] According to an embodiment of the present disclosure, the second computing unit includes: a first computing subunit and a second computing subunit.

[0159] The first calculation subunit is used to calculate the initial weight corresponding to the initial keyword according to the proportion of at least one historical reference weight in the preliminary work order to which the initial keyword belongs, when the correlation between the initial keyword and at least one historical keyword is greater than a preset correlation threshold; the second calculation subunit is used to calculate the initial weight corresponding to the initial keyword according to the historical reference weight and the similarity between the historical keyword corresponding to the historical reference weight and the initial keyword, when the correlation between the initial keyword and each historical keyword is less than or equal to the preset correlation threshold.

[0160] According to an embodiment of the present disclosure, the clustering module 640 includes a first generating submodule, a second generating submodule, and a clustering submodule.

[0161] The first generation submodule is used to input the work order text to be identified corresponding to the work order to be identified into the semantic vector generation model to generate a first semantic vector; the second generation submodule is used to input the preferred work order text corresponding to multiple preferred work orders into the semantic vector generation model to generate multiple second semantic vectors; the clustering submodule is used to cluster the first semantic vector and multiple second semantic vectors.

[0162] According to an embodiment of the present disclosure, the semantic vector generation model is trained by the following means:

[0163] The first acquisition module is used to acquire multiple groups of similar sample pairs and first sample labels of similar sample pairs; the second acquisition module is used to acquire multiple groups of non-similar sample pairs and second sample labels of non-similar sample pairs; the first output module is used to input at least one group of similar sample pairs into the initial model and output the first similarity of the similar sample pairs; the second output module is used to input at least one group of non-similar sample pairs into the initial model and output the second similarity of the non-similar sample pairs; the tuning module is used to adjust the model parameters of the initial model according to the first similarity, the first sample label, and the second similarity, the second sample label, so as to obtain a semantic vector generation model.

[0164] According to an embodiment of the present disclosure, the initial model is a dual-branch model, which includes a first branch network and a second branch network with the same network structure, and the first output module includes: a first input submodule, a second input submodule and a calculation submodule.

[0165] The first input submodule is used to input the first sample in the similar sample pair into the first branch network to generate a first sample semantic vector corresponding to the first sample; the second input submodule is used to input the second sample in the similar sample pair into the second branch network to generate a second sample semantic vector corresponding to the second sample; the calculation submodule is used to calculate the first sample similarity between the first sample semantic vector and the second sample semantic vector.

[0166] According to the embodiments of the present disclosure, the work order category of the work order to be identified is obtained through the first determination module, and the screening module can roughly screen out work orders that are related in category from historical work orders according to the work order category, without having to compare all historical work orders with the work order to be identified one by one, which can quickly narrow the scope of data calculation, reduce unnecessary analysis and processing, save time and resources, and improve the overall efficiency of work order processing. Further, in order to facilitate the provision of a more efficient and quick solution, starting from the characteristics of hot work orders (work orders for similar events that occur frequently within a certain period of time), preferred work orders that are highly similar to each other are determined from the preliminary work orders. Specifically, on the basis of the similarity of the work order category, the calculation module is used to calculate the similarity between each preliminary work order, so that the obtained preferred work orders are highly similar, while excluding the work orders that are semantically inconsistent with the work order to be identified from the preliminary work orders, further narrowing the scope of data calculation. The task category in the historical work order is used to provide a reference for the new work order to be identified, avoiding repeated exploration and errors. Compared with manual identification methods, decisions are made based on data and historical experience rather than subjective judgment, making them more scientific and reasonable.

[0167] According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (ASICs), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging circuits, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be performed.

[0168] For example, any multiple of the first determination module 610, the screening module 620, the calculation module 630, the clustering module 640 and the second determination module 650 can be combined in one module / unit / sub-unit for implementation, or any one of the modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of these modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units and implemented in one module / unit / sub-unit. According to an embodiment of the present disclosure, at least one of the first determination module 610, the screening module 620, the calculation module 630, the clustering module 640 and the second determination module 650 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 methods of software, hardware and firmware or in any appropriate combination of any of them. Alternatively, at least one of the first determination module 610 , the screening module 620 , the calculation module 630 , the clustering module 640 , and the second determination module 650 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0169] Figure 7 A block diagram of an electronic device suitable for implementing the hotspot work order identification method described above according to an embodiment of the present disclosure is schematically shown.

[0170] Figure 7The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0171] like Figure 7 As shown, the electronic device 700 according to an embodiment of the present disclosure includes a processor 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 to a random access memory (RAM) 703. The processor 701 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 701 may also include an onboard memory for caching purposes. The processor 701 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.

[0172] In RAM 703, various programs and data required for the operation of electronic device 700 are stored. Processor 701, ROM 702 and RAM 703 are connected to each other via bus 709. Processor 701 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 702 and / or RAM 703. It should be noted that the program can also be stored in one or more memories other than ROM 702 and RAM 703. Processor 701 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.

[0173] According to an embodiment of the present disclosure, the electronic device 700 may further include an input / output (I / O) interface 705, which is also connected to the bus 709. The electronic device 700 may further include one or more of the following components connected to the input / output (I / O) interface 705: an input portion 706 including a keyboard, a mouse, etc.; an output portion 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 708 including a hard disk, etc.; and a communication portion 709 including a network interface card such as a LAN card, a modem, etc. The communication portion 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the input / output (I / O) interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage portion 708 as needed.

[0174] According to an embodiment of the present disclosure, the method flow according to an embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the processor 701, the above-mentioned functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0175] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may 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 according to the embodiment of the present disclosure is implemented.

[0176] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it 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 that may be used by or in combination with an instruction execution system, apparatus, or device.

[0177] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 702 and / or the RAM 703 described above and / or one or more memories other than the ROM 702 and the RAM 703 .

[0178] An embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains a program code for executing the method provided by the embodiment of the present disclosure. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the hotspot work order identification method provided by the embodiment of the present disclosure.

[0179] When the computer program is executed by the processor 701, the above functions defined in the system / device of the embodiment of the present disclosure are executed. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0180] 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 709, and / or installed from the removable medium 711. 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.

[0181] 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).

[0182] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation 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 flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that 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 flowchart, and the combination of boxes in the block diagram or flowchart, 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. It can be understood by those skilled in the art that the features recorded in the various embodiments of the present disclosure can be combined and / or combined in a variety of ways, even if such a combination or combination is not explicitly recorded in the present disclosure. In particular, without departing from the spirit and teaching of the present disclosure, the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0183] 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, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. 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 hot work order identification method, comprising: Determine basic attribute information of at least one work order to be identified, wherein the basic attribute information at least includes a work order category; Based on the work order category, preliminary work orders related to the work order to be identified are obtained from historical work orders; Calculating the similarity between each of the preliminary selected work orders to select preferred work orders related to the work order to be identified from the preliminary selected work orders; performing clustering processing on a work order set consisting of the to-be-identified work orders and the preferred work orders; and, In response to the result of clustering processing, when the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition, it is determined that the work order to be identified is a hot work order.

2. The method according to claim 1, wherein: The method further comprises: Determining the target task category of the work order to be identified according to the result of the clustering process; and According to the target task category, the target task execution strategy corresponding to the target task category is determined from the preferred task execution strategies corresponding to the preferred work orders.

3. The method according to claim 1, wherein: Calculating the similarity between each of the preliminary selected work orders to select a preferred work order related to the work order to be identified from the preliminary selected work orders, including: Determining a first keyword of the preliminary work order based on a preliminary work order text corresponding to the preliminary work order; Determine the preliminary work order corresponding to the first keyword as a candidate work order; Calculate the similarity between each candidate work order; According to the similarities between the candidate work orders, a preferred work order related to the work order to be identified is determined from the candidate work orders.

4. The method according to claim 3, wherein: Determining a first keyword of the preliminary work order based on a preliminary work order text corresponding to the preliminary work order includes: Extracting a plurality of historical keywords from the historical work order text corresponding to the historical work order; Calculating the historical reference weight of the historical keyword; Extracting a plurality of initial keywords from the preliminary work order text; Based on the historical reference weights, respectively calculating initial weights corresponding to the initial keywords; Based on the initial weight, a first keyword of the preliminary work order is determined.

5. The method according to claim 4, wherein: Based on the historical reference weights, the initial weights corresponding to the initial keywords are calculated respectively, including: When the correlation between the initial keyword and at least one historical keyword is greater than a preset correlation threshold, the initial weight corresponding to the initial keyword is calculated according to the proportion of the at least one historical reference weight in the preliminary work order to which the initial keyword belongs; When the relevance between the initial keyword and each historical keyword is less than or equal to a preset relevance threshold, the initial weight corresponding to the initial keyword is calculated according to the historical reference weight and the similarity between the historical keyword corresponding to the historical reference weight and the initial keyword.

6. The method according to claim 1, wherein: Clustering a work order set consisting of the to-be-identified work orders and the preferred work orders includes: Inputting the work order text to be identified corresponding to the work order to be identified into a semantic vector generation model to generate a first semantic vector; Inputting the preferred work order texts corresponding to the plurality of preferred work orders into the semantic vector generation model to generate a plurality of second semantic vectors; Clustering is performed on the first semantic vector and the plurality of second semantic vectors.

7. The method according to claim 6, wherein: The semantic vector generation model is trained by the following method: Acquire multiple groups of similar sample pairs and first sample labels of the similar sample pairs; Acquire multiple groups of non-similar sample pairs and second sample labels of the non-similar sample pairs; Inputting at least one group of the similar sample pairs into an initial model, and outputting a first similarity of the similar sample pairs; Inputting at least one group of the non-similar sample pairs into the initial model, and outputting a second similarity of the non-similar sample pairs; According to the first similarity and the first sample label, and the second similarity and the second sample label, the model parameters of the initial model are adjusted to obtain the semantic vector generation model.

8. The method according to claim 7, wherein: The initial model is a double-branch model, which includes a first branch network and a second branch network having the same network structure. Inputting at least one group of the similar sample pairs into an initial model and outputting a first similarity of the similar sample pairs comprises: Inputting a first sample in the similar sample pair into the first branch network to generate a first sample semantic vector corresponding to the first sample; Inputting the second sample in the similar sample pair into the second branch network to generate a second sample semantic vector corresponding to the second sample; The first sample similarity between the first sample semantic vector and the second sample semantic vector is calculated and obtained.

9. A hotspot work order identification device, comprising: A first determination module, configured to determine basic attribute information of at least one work order to be identified, wherein the basic attribute information at least includes a work order category; A screening module, used for screening out preliminary work orders related to the work order to be identified from historical work orders based on the work order category; A calculation module, used for calculating the similarity between each of the preliminary selected work orders, so as to screen out preferred work orders related to the work order to be identified from the preliminary selected work orders; a clustering module, used for clustering a work order set consisting of the to-be-identified work orders and the preferred work orders; and The second determination module is used to determine that the work order to be identified is a hot work order in response to the result of clustering processing when the number of work orders in the target cluster to which the work order to be identified belongs meets a preset numerical condition.

10. An electronic device, comprising: one or more processors; a memory 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 implement the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the method according to any one of claims 1 to 8.

12. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images