Model training method, event dispatching processing method, equipment and medium
By training event allocation prediction models and real-time provisioning engineers, the service efficiency problem caused by unpredictable event allocation in the prior art is solved, and more efficient event processing and service efficiency are achieved.
Patent Information
- Application Number
- CN202110940636.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-08-17
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In the prior art, event allocation is not predicted, resulting in engineers needing to handle both remote and on-site events, which makes service efficiency inefficient.
By training the event allocation prediction model, the title, symptoms and solution path of historical events are used to generate coding vectors, input the convolutional neural network for training, predict the solution of the event, and determine the processing plan for the event to be allocated based on the engineer's status data. Real-time provisioning of the number of engineers to match event requirements.
Improve the accuracy and service efficiency of event resolution, ensure that engineers focus on handling remote or on-site events, reduce frequent movements between workstations and users, and improve work efficiency.
Smart Images

Figure CN113743746B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of Internet technology, and in particular to a model training method, an event dispatching processing method, a device and a medium. Background Art
[0002] In today's large enterprises, the IT (Internet technology) department has a dedicated service team to support service requests for various types of IT-related issues among employees within the enterprise. In this scenario, when engineers help users solve problems, they can be divided into two forms: "remote solution" and "on-site solution". For example, problems such as software installation and permission settings can be directly supported through remote desktop, telephone, email, etc. For hardware, operating system and other related failures, engineers need to walk from their own workstations to the user's workstation for one-on-one on-site support.
[0003] Therefore, the existing technology does not prejudge the solutions to all events, and directly assigns engineers without grouping them, which inevitably leads to engineers having to handle remote and on-site events at the same time, resulting in low service efficiency. Summary of the invention
[0004] The technical problem to be solved by the present invention is to overcome the defect of low event processing efficiency caused by directly allocating events to engineers in the prior art, and to provide a model training method, event dispatching processing method, device and medium.
[0005] The present invention solves the above technical problems through the following technical solutions:
[0006] According to a first aspect of the present invention, a method for training an event allocation prediction model is provided, the training method comprising:
[0007] Acquire a plurality of texts to be segmented corresponding to historical events to be trained; the texts to be segmented include title texts, symptom texts, and solution texts;
[0008] Performing full word segmentation processing on the text to be segmented to generate a first word segmentation corresponding to the title text, a second word segmentation corresponding to the symptom text, and a third word segmentation corresponding to the solution text;
[0009] Acquire a first encoding vector corresponding to the first word segmentation and a second encoding vector corresponding to the second word segmentation based on a pre-constructed target dictionary;
[0010] Associating the first encoding vector with the second encoding vector to generate a third encoding vector;
[0011] Obtain a fourth encoding vector corresponding to the third word segmentation;
[0012] The third encoding vector and the fourth encoding vector corresponding to each of the historical events are input into a convolutional neural network model for training to generate an event allocation prediction model.
[0013] Preferably, before the step of acquiring the first encoding vector corresponding to the first word segmentation and the second encoding vector corresponding to the second word segmentation based on the pre-built target dictionary, the step includes:
[0014] Taking the first word segmentation whose number of occurrences in a preset time period is greater than a first set number as a first feature word, and encoding the first feature word according to a first preset algorithm to generate a corresponding first dictionary label vector;
[0015] Taking the second participle whose number of occurrences in the preset time period is greater than the second set number as the second feature word, and encoding the second feature word according to the second preset algorithm to generate a corresponding second dictionary label vector;
[0016] A target dictionary is generated according to the first dictionary label vector and the second dictionary label vector.
[0017] Preferably, the first preset algorithm includes a TF-IDF (Term Frequency–Inverse Document Frequency) algorithm, and / or the second preset algorithm includes a TF-IDF algorithm.
[0018] Preferably, the step of performing full word segmentation processing on the text to be segmented includes:
[0019] At least one of auxiliary words, quantifiers and line breaks of a preset length in the text to be segmented is deleted.
[0020] According to a second aspect of the present invention, the present invention provides a method for processing event dispatching, the method comprising:
[0021] Get the basic data corresponding to the event to be assigned;
[0022] Inputting the basic data into an event allocation prediction model trained by any one of the methods in the first aspect to obtain a corresponding prediction result;
[0023] Get the status data corresponding to the engineer information;
[0024] A processing scheme for the event to be assigned is determined according to the status data and the prediction result.
[0025] Preferably, the status data includes a group name and status information, and the group name includes an on-site group and a remote group;
[0026] The step of obtaining status data corresponding to the engineer information includes:
[0027] Acquire the status information of the engineer corresponding to the remote group or the on-site group; the status information includes offline working and online waiting;
[0028] The step of determining a processing scheme for the event to be allocated according to the state data and the prediction result comprises:
[0029] The online waiting engineers corresponding to the remote group or the on-site group are used as engineers to be assigned;
[0030] One of the engineers to be assigned is selected as a target engineer, and the event to be assigned is processed.
[0031] Preferably, the step of selecting one of the engineers to be assigned as a target engineer includes:
[0032] Calculating the work saturation value of each of the engineers to be assigned; the work saturation value represents the ratio of the number of events currently processed by the engineer to be assigned to the processing capacity value;
[0033] The engineer to be assigned corresponding to the minimum work saturation value is used as the target engineer.
[0034] Preferably, after the step of obtaining the status data corresponding to the engineer information, the following steps are included:
[0035] Calculating the group work saturation values corresponding to the remote group and the on-site group; the group work saturation value represents the ratio of the sum of the number of all currently processed events in the group and the number of pending events in the group to the processing capacity of all engineers in the group;
[0036] Personnel deployment is performed between the remote group and the on-site group according to the group work saturation value.
[0037] Preferably, the step of allocating personnel between the remote group and the on-site group according to the group work saturation value comprises:
[0038] comparing a first group of work saturation values corresponding to the remote group with a second group of work saturation values corresponding to the on-site group;
[0039] If the work saturation value of the first group is greater than the work saturation value of the second group, the engineers of the on-site group are transferred to the remote group until the work saturation value of the first group is close to the work saturation value of the second group;
[0040] If the work saturation value of the first group is less than the work saturation value of the second group, engineers of the remote group are transferred to the on-site group until the work saturation value of the first group is close to the work saturation value of the second group.
[0041] According to the third aspect of the present invention, the present invention also provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the training method of the event allocation prediction model as described in the first aspect, or the event dispatch processing method as described in the second aspect is implemented.
[0042] According to the fourth aspect of the present invention, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the training method of the event allocation prediction model as described in the first aspect, or the event dispatch processing method as described in the second aspect.
[0043] The positive and progressive effects of the present invention are: providing a model training method, an event dispatch processing method, a device and a medium. In the model training method, by obtaining the first coding vector corresponding to the title text of the historical event to be trained, the second coding vector corresponding to the symptom text and the third coding vector corresponding to the solution text, the first coding vector, the second coding vector and the third coding vector are input into the convolutional neural network for training, thereby improving the accuracy of predicting the solution to the event. The trained event allocation prediction model is used to predict the events to be assigned, and the remote solution or on-site solution is determined according to the prediction results. The work saturation of the remote group and the on-site group is tracked in real time, and the number of engineers in each group is dynamically allocated to match the number of engineers corresponding to the solution required for each type of event. The work saturation of each group of engineers is tracked in real time to determine the target engineer corresponding to each event, thereby improving the overall service efficiency of the team. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 This is a first flow chart of the method for training an event allocation prediction model according to Embodiment 1 of the present invention.
[0045] Figure 2 This is a second flow chart of the method for training the event allocation prediction model of embodiment 1 of the present invention.
[0046] Figure 3 This is a first flow chart of the event dispatch processing method of embodiment 2 of the present invention.
[0047] Figure 4 This is a first flow chart of the event dispatch processing method of embodiment 3 of the present invention.
[0048] Figure 5 This is a flowchart of step S42 of the event dispatch processing method of embodiment 3 of the present invention.
[0049] Figure 6 This is a second flow chart of the event dispatch processing method according to the third embodiment of the present invention.
[0050] Figure 7 This is a schematic diagram of the hardware structure of an electronic device according to Embodiment 4 of the present invention. DETAILED DESCRIPTION
[0051] The present invention is further described below by way of examples, but the present invention is not limited to the scope of the examples.
[0052] Example 1
[0053] like Figure 1 As shown, this embodiment provides a training method for an event allocation prediction model, and the training method includes the following steps:
[0054] S110, obtaining texts to be segmented corresponding to multiple historical events to be trained; the texts to be segmented include title texts, symptom texts, and solution texts.
[0055] In this embodiment, the text to be segmented may include but is not limited to text in service request information of various types of IT-related issues. For example, the text to be segmented 1 may be the title text of "software failure", the symptom text of "email cannot be opened", and the solution text of "remote solution"; the text to be segmented 2 may be the title text of "hardware failure", the symptom text of "computer cannot be turned on", and the solution text of "on-site solution"; the text to be segmented 3 may be the title text of "authorization application", the symptom text of "application for remote desktop authorization", and the solution text of "remote solution".
[0056] S120, performing full word segmentation processing on the text to be segmented, generating a first word segmentation corresponding to the title text, a second word segmentation corresponding to the symptom text, and a third word segmentation corresponding to the solution text.
[0057] For example, the segmented text 1 is fully segmented to generate two first segmented words, "software" and "fault", two second segmented words, "mailbox" and "cannot open", and a third segmented word, "remote solution". The segmented text 2 is fully segmented to generate two first segmented words, "hardware" and "fault", two second segmented words, "computer" and "cannot boot", and a third segmented word, "on-site solution". The segmented text 3 is fully segmented to generate two first segmented words, "authority" and "application", four second segmented words, "application", "opening", "remote", and "desktop authority", and a third segmented word, "remote solution".
[0058] S130 , obtaining a first encoding vector corresponding to the first word segmentation and a second encoding vector corresponding to the second word segmentation based on a pre-constructed target dictionary.
[0059] For example, two first coding vectors corresponding to "software" and "fault", two second coding vectors corresponding to "mailbox" and "cannot open", and a third coding vector corresponding to "remote solution" are obtained from the target dictionary.
[0060] S140, associating the first encoding vector and the second encoding vector to generate a third encoding vector.
[0061] The two first encoding vectors in the segmented text 1 are associated with the two second encoding vectors. For example, the association method includes but is not limited to the method of connecting the vectors end to end to generate a third encoding vector.
[0062] S150. Obtain a fourth encoding vector corresponding to the third word segmentation.
[0063] In this embodiment, if the third word segmentation is "on-site solution" or "remote solution", the fourth coding vector corresponding to "on-site solution" and the fourth coding vector corresponding to "remote solution" are obtained. It should be noted that the setting of the fourth coding vector can be limited according to the situation. For example, the coding vector corresponding to "on-site solution" is [0.1], and the coding vector corresponding to "remote solution" is [1,0].
[0064] S160: Input the third coding vector and the fourth coding vector corresponding to each historical event into the convolutional neural network model for training to generate an event allocation prediction model.
[0065] When the third encoding vector and the fourth encoding vector in the text to be segmented 1 are input into the DNN (Deep Neural Networks, deep neural network model) for training, the output state is M states (M=2), and the hidden layer H dimension of the network model does not need to be too large. For example, H=2M is sufficient, and the learning speed and number of iterations of the model can be specifically debugged according to the actual sample data, and no specific restrictions are made here.
[0066] After the model training is completed, when there are W pending events to be predicted, obtain the word list of each pending event, substitute it into the DNN network model for calculation, and obtain a W*M prediction result matrix. Calculate the sum of the squares of the differences between W 1*m matrices and each output result (1*M) vector to obtain W prediction results. For example, the output vector obtained by the model for a certain event is (a, b): P 远程 =(a-1)^2+b^2; P 现场 =(a^2+(b-1)^2), if P 远程 <P 现场, then this event is resolved remotely; otherwise, it is resolved on-site.
[0067] It should be noted that this DNN model can be composed of a fully connected layer, an activation layer, and a fully connected output layer, or can be set according to actual situations, which is not limited herein.
[0068] Among them, in this embodiment, step S120 specifically includes:
[0069] Delete at least one of the function words, quantifiers, and line breaks with a preset length in the text to be segmented.
[0070] Specifically, filter function words, quantifiers, line breaks, etc. such as "de", "ge", "tai", "\r\n" in the text to be segmented, and then perform full word segmentation on the filtered text.
[0071] In an implementable solution, as Figure 2 shown, after step 120, it further includes:
[0072] S121. Use the words in the first word segmentation that appear more than a first set number within a preset time period as the first feature words, and encode the first feature words according to a first preset algorithm to generate corresponding first dictionary label vectors.
[0073] S122. Use the words in the second word segmentation that appear more than a second set number within a preset time period as the second feature words, and encode the second feature words according to a second preset algorithm to generate corresponding second dictionary label vectors;
[0074] S123. Generate a target dictionary according to the first dictionary label vector and the second dictionary label vector.
[0075] Among them, the first preset algorithm includes the TF-IDF algorithm, and / or the second preset algorithm includes the TF-IDF algorithm.
[0076] Specifically, the first set number can be 3, 4, 5, etc., and the second set number can be 3, 4, 5, etc. Screen out the words in all the texts to be segmented where each first word segmentation appears more than 3 times, and rank them in descending order of the number of appearances. Take the top 200 words as the first feature words; screen out the words in all the texts to be segmented where each second word segmentation appears more than 3 times, and rank them in descending order of the number of appearances. Take the top 200 words as the second feature words; the first feature words and the second feature words can be encoded according to the same preset algorithm or different preset algorithms to generate dictionary label vectors, and further generate a target dictionary.
[0077] In this embodiment, a method for training an event allocation prediction model is provided. The present invention trains the model based on vectors corresponding to title texts of historical events, vectors corresponding to symptom texts, and vectors corresponding to solution texts, thereby improving the accuracy of model predictions.
[0078] Example 2
[0079] like Figure 3 As shown, this embodiment provides a method for processing event dispatching, and the method comprises the following steps:
[0080] S1. Obtain basic data corresponding to the event to be assigned.
[0081] S2. Input the basic data into the event allocation prediction model trained by the method of Example 1 to obtain corresponding prediction results.
[0082] S3. Obtain status data corresponding to the engineer information.
[0083] S4. Determine a processing plan for the event to be assigned based on the status data and the prediction results.
[0084] In this embodiment, the list of events to be assigned can be obtained in real time from the enterprise's event submission system, and the engineer information can be obtained from the enterprise's engineer sign-in and sign-out system in real time, and the status data of engineers in each group can be obtained based on the engineer information. The title text, symptom text, and solution text of each event in the event list are input into the event allocation prediction model trained by the method of Example 1 to obtain the prediction result of each event to be assigned, and the processing plan for each event to be assigned is determined based on the status data and the prediction result.
[0085] This embodiment provides a method for handling event dispatch, which uses a trained model to predict the dispatched events for remote or on-site resolution, tracks the work saturation of the remote group and the on-site group in real time, and dynamically adjusts the number of engineers in the group, thereby improving service efficiency. It enables engineers in various departments within the enterprise to be dedicated to their specific tasks, and different events are handled by group, ensuring that each engineer can focus on handling remote or on-site events within a preset time period, avoiding frequent movement between workstations and users, and improving work efficiency.
[0086] Example 3
[0087] Based on Example 2, this example provides a method for processing event dispatching, such as Figure 4 As shown, compared with Example 2, improvements are made, specifically:
[0088] In one feasible solution, the method further comprises the following steps:
[0089] S1. Obtain basic data corresponding to the event to be assigned.
[0090] S2. Input the basic data into the event allocation prediction model trained by the method of Example 1 to obtain corresponding prediction results.
[0091] S31. Obtain status information of engineers corresponding to the remote group or the on-site group; the status information includes offline working and online waiting.
[0092] S41. The online waiting engineers of the remote group or the on-site group are designated as engineers to be assigned.
[0093] S42. Select an engineer to be assigned as a target engineer and process the event to be assigned.
[0094] For example, obtain basic data 1 corresponding to event 1 to be assigned and basic data 2 corresponding to event 2 to be assigned, input basic data 1 and basic data 2 into the model for prediction, and the prediction result is that event 1 to be assigned is solved remotely, and event 2 to be assigned is solved on-site. Obtain the status information of each engineer in the remote group or the on-site group. For example, the status information can be expressed as (remote group, engineer Wang, offline work), (remote group, engineer Zhang, online waiting), (on-site group, engineer Li, online waiting), (on-site group, engineer Lin, offline work). Select Engineer Zhang in the remote group as the engineer to be assigned to handle event 1 to be assigned, or select Engineer Li in the on-site group as the engineer to be assigned to handle event 2 to be assigned.
[0095] In one feasible solution, Figure 5 As shown, step S42 specifically includes:
[0096] Step S421, calculating the work saturation value of each engineer to be assigned; the work saturation value represents the ratio of the number of events currently processed by the engineer to be assigned to the processing capacity value.
[0097] Step S422: The engineer to be assigned corresponding to the minimum work saturation value is taken as the target engineer.
[0098] Specifically, the processing capacity of each group of engineers can be set according to the situation. For example, there are 2 engineers in the remote group, and the processing capacity value of each engineer is 4. The number of events currently processed by engineer A is 2, the number of events currently processed by engineer B is 1, and the number of events currently processed by engineers C, D, and E is 4. Since the work saturation value of engineer A is 0.5 and the work saturation value of engineer B is 0.25, engineer B is selected as the target engineer of the remote group.
[0099] In one feasible solution, Figure 6 As shown, after step S3, the following steps are also included:
[0100] Step S32, calculating the group work saturation values corresponding to the remote group and the on-site group; the group work saturation value represents the ratio of the sum of the number of currently processed events in the group and the number of pending events in the group to the processing capacity of all engineers in the group.
[0101] Step S33: deploy personnel between the remote group and the on-site group according to the group work saturation value.
[0102] For example, the processing capacity value of the remote group engineer is 5, and the processing capacity value of the on-site group engineer is 5. The remote group has currently been assigned 12 events, including events being processed and events that have been assigned to be processed. The on-site group has currently been assigned 8 events. The group work saturation value corresponding to the remote group is 12 / (5*4)=0.6, and the group work saturation value corresponding to the on-site group is 8 / (5*4)=0.4. It can be seen that the group work saturation of the on-site group and the remote group is different. The personnel of the on-site group and the remote group are deployed so that the work saturation values of the two groups are approximately the same.
[0103] Among them, in the above-mentioned possible implementation scheme, step S33 specifically includes:
[0104] Step S331: compare the first group of working saturation values corresponding to the remote group with the second group of working saturation values corresponding to the on-site group.
[0105] Step S332: If the working saturation value of the first group is greater than the working saturation value of the second group, the engineers of the on-site group are transferred to the remote group until the working saturation value of the first group is close to the working saturation value of the second group.
[0106] Step S333: If the working saturation value of the first group is less than the working saturation value of the second group, engineers of the remote group are transferred to the on-site group until the working saturation value of the first group is close to the working saturation value of the second group.
[0107] Specifically, the first group work saturation value corresponding to the remote group is 0.6, and the second group work saturation value corresponding to the on-site group is 0.4. After transferring one engineer from the on-site group to the remote group, the first group work saturation value corresponding to the remote group is 12 / (6*4)=0.5, and the second group work saturation value corresponding to the on-site group is 8 / (4*4)=0.5. At this time, the work saturation values corresponding to the remote group and the on-site group are the same, and the personnel transfer is completed.
[0108] This embodiment provides a method for handling event dispatch, which uses a trained model to predict the dispatched events for remote or on-site resolution, tracks the work saturation of the remote group and the on-site group in real time, and dynamically adjusts the number of engineers in the group, thereby improving service efficiency. It enables engineers in various departments within the enterprise to be dedicated to their specific tasks, and different events are handled by group, ensuring that each engineer can focus on handling remote or on-site events within a preset time period, avoiding frequent movement between workstations and users, and improving work efficiency.
[0109] Example 4
[0110] Figure 7 A schematic diagram of the structure of an electronic device provided in Example 4 of the present invention. The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the program, the training method of the event allocation prediction model in Example 1 or the event allocation processing method in Example 2 or Example 3 is implemented.
[0111] Figure 7 The electronic device 60 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.
[0112] like Figure 7 As shown, the electronic device 60 may be in the form of a general-purpose computing device, for example, it may be a server device. The components of the electronic device 60 may include, but are not limited to: at least one processor 61, at least one memory 62, and a bus 63 connecting different system components (including the memory 62 and the processor 61).
[0113] The bus 63 includes a data bus, an address bus, and a control bus.
[0114] The memory 62 may include a volatile memory, such as a random access memory (RAM) 621 and / or a cache memory 622 , and may further include a read-only memory (ROM) 623 .
[0115] The memory 62 may also include a program / utility 625 having a set (at least one) of program modules 624, such program modules 624 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.
[0116] The processor 61 executes various functional applications and data processing by running the computer program stored in the memory 62, such as the training method of the event allocation prediction model in Example 1 of the present invention or the event dispatch processing method in Example 2 or Example 3.
[0117] The electronic device 60 may also communicate with one or more external devices 64 (e.g., keyboards, pointing devices, etc.). Such communication may be performed via an input / output (I / O) interface 65. Furthermore, the electronic device 60 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 66. Figure 7 As shown, the network adapter 66 communicates with other modules of the electronic device 60 via the bus 63. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device 60, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems, etc.
[0118] It should be noted that although several units / modules or sub-units / modules of the electronic device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into multiple units / modules to be embodied.
[0119] Example 5
[0120] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the program implements the steps of the training method of the event allocation prediction model in Embodiment 1 or the steps of the event dispatch processing method in Embodiment 2 or Embodiment 3.
[0121] The readable storage medium may include but is not limited to: a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the above.
[0122] In a possible implementation, the present invention can also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to enable the terminal device to execute the steps of the training method of the event allocation prediction model in Example 1 or the steps of the event dispatch processing method in Example 2 or Example 3.
[0123] Among them, the program code for executing the present invention can be written in any combination of one or more programming languages, and the program code can be executed completely on the user device, partially on the user device, as an independent software package, partially on the user device and partially on a remote device, or completely on the remote device.
[0124] Although the specific embodiments of the present invention are described above, it should be understood by those skilled in the art that this is only for illustration and the protection scope of the present invention is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but these changes and modifications all fall within the protection scope of the present invention.
Claims
1. A training method for an event allocation prediction model, characterized in that: The training method comprises: Acquire a plurality of texts to be segmented corresponding to historical events to be trained; the texts to be segmented include title texts, symptom texts, and solution texts; Performing full word segmentation processing on the text to be segmented to generate a first word segmentation corresponding to the title text, a second word segmentation corresponding to the symptom text, and a third word segmentation corresponding to the solution text; Acquire a first encoding vector corresponding to the first word segmentation and a second encoding vector corresponding to the second word segmentation based on a pre-constructed target dictionary; Associating the first encoding vector with the second encoding vector to generate a third encoding vector; Obtain a fourth encoding vector corresponding to the third word segmentation; The third encoding vector and the fourth encoding vector corresponding to each of the historical events are input into a convolutional neural network model for training to generate an event allocation prediction model.
2. The method for training an event distribution prediction model according to claim 1, characterized in that: Before the step of acquiring the first encoding vector corresponding to the first word segmentation and the second encoding vector corresponding to the second word segmentation based on the pre-built target dictionary, the method includes: Taking the first word segmentation whose number of occurrences in a preset time period is greater than a first set number as a first feature word, and encoding the first feature word according to a first preset algorithm to generate a corresponding first dictionary label vector; Taking the second participle whose number of occurrences in the preset time period is greater than the second set number as the second feature word, and encoding the second feature word according to the second preset algorithm to generate a corresponding second dictionary label vector; A target dictionary is generated according to the first dictionary label vector and the second dictionary label vector.
3. The method for training an event distribution prediction model according to claim 2, characterized in that: The first preset algorithm includes a TF-IDF algorithm, and / or the second preset algorithm includes a TF-IDF algorithm.
4. The method for training an event distribution prediction model according to claim 1, wherein: The step of performing full word segmentation processing on the text to be segmented includes: At least one of auxiliary words, quantifiers and line breaks of a preset length in the text to be segmented is deleted.
5. A method for processing event dispatch, characterized in that: The processing method comprises: Obtaining basic data corresponding to the event to be assigned; the basic data includes a title text and a symptom text of the event to be assigned; Inputting the basic data into an event allocation prediction model trained by the method according to any one of claims 1 to 4 to obtain a corresponding prediction result; Get the status data corresponding to the engineer information; Determine a processing scheme for the event to be assigned according to the status data and the prediction result; The status data includes group names and status information, and the group names include on-site groups and remote groups; The step of obtaining status data corresponding to the engineer information includes: Acquire the status information of the engineer corresponding to the remote group or the on-site group; the status information includes offline working and online waiting; The step of determining a processing scheme for the event to be allocated according to the state data and the prediction result comprises: The online waiting engineers corresponding to the remote group or the on-site group are used as engineers to be assigned; One of the engineers to be assigned is selected as a target engineer, and the event to be assigned is processed.
6. The event dispatch processing method according to claim 5, characterized in that: The step of selecting an engineer to be assigned as a target engineer includes: Calculating the work saturation value of each of the engineers to be assigned; the work saturation value represents the ratio of the number of events currently processed by the engineer to be assigned to the processing capacity value; The engineer to be assigned corresponding to the minimum work saturation value is used as the target engineer.
7. The event dispatch processing method according to claim 5, characterized in that: After the step of obtaining the status data corresponding to the engineer information, the following steps are included: Calculating the group work saturation values corresponding to the remote group and the on-site group; the group work saturation value represents the ratio of the sum of the number of all currently processed events in the group and the number of pending events in the group to the processing capacity of all engineers in the group; Personnel deployment is performed between the remote group and the on-site group according to the group work saturation value.
8. The event dispatch processing method according to claim 7, characterized in that: The step of allocating personnel between the remote group and the on-site group according to the group work saturation value comprises: comparing a first group of work saturation values corresponding to the remote group with a second group of work saturation values corresponding to the on-site group; If the work saturation value of the first group is greater than the work saturation value of the second group, the engineers of the on-site group are transferred to the remote group until the work saturation value of the first group is close to the work saturation value of the second group; If the work saturation value of the first group is less than the work saturation value of the second group, engineers of the remote group are transferred to the on-site group until the work saturation value of the first group is close to the work saturation value of the second group.
9. An electronic device, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the training method of the event allocation prediction model as described in any one of claims 1 to 4, or the processing method of event dispatch as described in any one of claims 5 to 8.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the training method of the event allocation prediction model as described in any one of claims 1-4, or the event dispatch processing method as described in any one of claims 5-8.
