Intelligent event allocation method and device based on machine learning

Abstract

The invention discloses an intelligent event allocation method and device based on machine learning, and the method comprises the steps: loading the weight and architecture of a pre-training model from a model library, and enabling the pre-training model to be a language representation model; the event text data is input into a pre-training model, a gradient descent method is used for updating the weight of the pre-training model, an intelligent event distribution model is obtained, and the training and deployment process of the intelligent event distribution model is directly completed at the front end based on an open source library TensorFlow.js; under the condition that the input event information is detected, word segmentation processing is carried out on the input event information according to the intelligent event distribution model to obtain a prediction result, and the prediction result comprises an event type corresponding to the input event information; and updating the intelligent event allocation model based on the input event information and a corresponding event type result obtained by interaction with the user.

Description

Technical Field

[0001] The present application relates to the field of machine learning technology, and more specifically, to a method and device for intelligent event distribution based on machine learning. Background Art

[0002] Existing intelligent event dispatching technologies usually pre-train the model first and then deploy the trained model. This approach may encounter challenges in actual deployment. For example, the complexity of the model and system may lead to high implementation and maintenance costs, limiting the application of the model in small or medium-sized systems.

[0003] However, although the intelligent event distribution model has a certain ability to obtain data in real time, it may lag behind in the data processing and analysis stage due to its reliance on traditional technical architecture. In a rapidly changing business environment, the model may not be able to adapt to the latest event types and processing requirements in a timely manner. In some event distribution systems, there is an imperfect user feedback mechanism, which makes it difficult for the system to optimize based on user feedback data, thus affecting long-term use results. Summary of the invention

[0004] In response to at least one defect or improvement need in the prior art, the present invention provides a method and device for intelligent event distribution based on machine learning, which solves the problem that the user feedback mechanism is not perfect and the system is difficult to optimize based on user feedback data. The intelligent event distribution model based on natural language processing achieves convenience and real-time performance.

[0005] To achieve the above-mentioned purpose, according to a first aspect of the present invention, a method for intelligent event distribution based on machine learning is provided, the method comprising: loading the weights and architecture of a pre-trained model from a model library, wherein the pre-trained model is a language representation model; inputting event text data into the pre-trained model, and updating the weights of the pre-trained model using a gradient descent method to obtain an intelligent event distribution model, wherein the training and deployment process of the intelligent event distribution model is directly completed on the front end based on the open source library TensorFlow.js; when input event information is detected, performing word segmentation processing on the input event information according to the intelligent event distribution model to obtain a prediction result, wherein the prediction result includes an event type corresponding to the input event information; and updating the intelligent event distribution model based on the input event information and the corresponding event type result obtained by interacting with the user.

[0006] In an exemplary embodiment, before loading the weights and architecture of the pre-trained model from the model library, the method also includes: converting the input event text data into a word vector through a word segmentation tool, wherein the event text data includes different types of event text descriptions and classification labels corresponding to the event text descriptions, and the word segmentation tool is used to convert the original continuous event text data into a numerical form.

[0007] In an exemplary embodiment, the event text data is input into the pre-trained model, and the weights of the pre-trained model are updated using the gradient descent method to obtain the event intelligent distribution model, including: defining a cross entropy loss function and defining an adaptive moment estimation optimizer, wherein the cross entropy loss function is suitable for multi-classification tasks; based on the cross entropy loss function and the adaptive moment estimation optimizer, the weights of the pre-trained model are updated to obtain the event intelligent distribution model.

[0008] In an exemplary embodiment, after the weights of the pre-trained model are updated based on the cross entropy loss function and the adaptive moment estimation optimizer to obtain the event intelligent distribution model, the method further includes: setting hyperparameters based on the input event information and selecting event types, wherein the hyperparameters include the number of iterations, the learning rate, and the batch size; updating the model parameters of the event intelligent distribution model based on the event type corresponding to the predicted input event information obtained by the event intelligent distribution model, in combination with the cross entropy loss function and the adaptive moment estimation optimizer, until the update stop condition is reached.

[0009] In an exemplary embodiment, when input event information is detected, after the input event information is segmented according to the event intelligent distribution model to obtain a prediction result, the method further includes: the event intelligent distribution model sorts the event keywords corresponding to the input event information in a sequence of high and low relevance and generates an event type selection option; and generates prompt information for prompting the user to select an event type from the provided event type selection options.

[0010] In an exemplary embodiment, when input event information is detected, performing word segmentation processing on the input event information according to the event intelligent distribution model to obtain a prediction result includes: writing logic in the front-end target code to call the event intelligent distribution model to perform word segmentation processing on the input event information to obtain a word segmentation processing result; performing event type prediction based on the word segmentation processing result to obtain the prediction result; and generating and displaying prediction information containing the prediction result on the front-end page.

[0011] According to the second aspect of the present invention, there is also provided an event intelligent distribution device based on machine learning, which includes: a loading unit, used to load the weights and architecture of a pre-trained model from a model library, wherein the pre-trained model is a language representation model; an acquisition unit, used to input event text data into the pre-trained model, and use the gradient descent method to update the weights of the pre-trained model to obtain an event intelligent distribution model, wherein the event intelligent distribution model training and deployment process is directly completed on the front end based on the open source library TensorFlow.js; a prediction unit, used to, when input event information is detected, perform word segmentation processing on the input event information according to the event intelligent distribution model to obtain a prediction result, wherein the prediction result includes the event type corresponding to the input event information; a first update unit, used to update the event intelligent distribution model based on the input event information and the corresponding event type result obtained by interaction with the user.

[0012] According to a third aspect of the present invention, there is also provided a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned event intelligent distribution method based on machine learning when running.

[0013] According to a fourth aspect of the present invention, there is also provided an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-mentioned event intelligent distribution method based on machine learning through the computer program.

[0014] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0015] (1) The present invention provides an event intelligent allocation method based on machine learning, which can directly define, train and run the open source library TensorFlow.js of the machine learning model in the browser and Node.js environment, train and deploy the event intelligent allocation model based on natural language processing, and achieve convenience and real-time performance. When the user inputs event information, the event intelligent allocation model sorts the event keywords from high to low according to the relevance, and provides the user with a choice of event types. This can not only increase the efficiency of event processing, but also further train and optimize the model through user selection.

[0016] (2) By adopting the event intelligent distribution method based on machine learning provided by the present invention, the accuracy of the model is continuously optimized and enhanced by using the real-time feedback of users as new input data, and incremental learning is achieved, that is, the model can continuously learn and adapt to new data without frequent retraining. This solution is not only convenient to deploy and implement, but also has the scalability to be applied to other types of event distribution in addition to public event distribution. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of a flow chart of an optional method for intelligent event allocation based on machine learning provided in an embodiment of the present application;

[0019] Figure 2 An optional event intelligent allocation model training deployment flow chart provided for an embodiment of the present application;

[0020] Figure 3 An optional incremental learning flow chart provided for an embodiment of the present application;

[0021] Figure 4 A schematic diagram of the structure of an optional event intelligent distribution device based on machine learning provided in an embodiment of the present application;

[0022] Figure 5 A schematic diagram of the structure of an optional electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0024] The terms "first", "second", "third", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0025] According to one aspect of the embodiments of the present application, a method for intelligent event distribution based on machine learning is provided. Figure 1 Describe the event intelligent distribution method based on machine learning provided in the embodiments of the present application.

[0026] Figure 1 is a flow chart of an optional event intelligent allocation method based on machine learning provided in an embodiment of the present application, such as Figure 1 As shown, the process of the method may include the following steps:

[0027] S102, loading the weight and architecture of a pre-trained model from a model library, wherein the pre-trained model is a language representation model;

[0028] S104, inputting the event text data into the pre-trained model, and using the gradient descent method to update the weight of the pre-trained model to obtain an event intelligent allocation model, wherein the event intelligent allocation model training and deployment process is directly completed on the front end based on the open source library TensorFlow.js;

[0029] S106, when input event information is detected, segmenting the input event information according to the event intelligent allocation model to obtain a prediction result, wherein the prediction result includes an event type corresponding to the input event information;

[0030] S108: Update the event intelligent allocation model based on the input event information and the corresponding event type results obtained through interaction with the user.

[0031] An event intelligent distribution method based on machine learning provided in an embodiment of the present application can be used in scenarios where different types of events are assigned to corresponding processing departments according to certain rules. Event intelligent distribution further utilizes artificial intelligence technology to achieve more efficient and accurate event identification and distribution. Event distribution technology has a wide range of applications in urban management, emergency response, customer service and other fields, and can direct various events to appropriate channels. With the increasing complexity of urban governance needs and the improvement of the level of intelligence, event intelligent distribution technology has emerged. Relying on advanced algorithm models, it has significantly improved the intelligence and automation level of the distribution process. The intelligent distribution system can self-learn, continuously optimize the distribution algorithm, improve the accuracy of identification and distribution, while reducing the manual burden and improving overall efficiency. Event intelligent distribution technology not only optimizes the event processing process, but also provides strong technical support for the continuous improvement of future urban governance and public services.

[0032] It should be noted that TensorFlow.js is an open source JavaScript library that allows machine learning models to be defined, trained, and run in browsers and Node.js environments. TensorFlow.js can run in browsers, mobile devices, Internet of Things (IoT) devices, and Node.js environments, which makes it possible to handle complex machine learning tasks directly on the front end without relying on server-side processing. TensorFlow.js provides an easy-to-use application programming interface (API) and pre-trained models, making machine learning development in a JavaScript environment easier and more efficient. These tools not only support model training and reasoning, but also allow developers to quickly build applications using existing pre-trained models.

[0033] For example, in this embodiment, the open source library TensorFlow.js, which can define, train and run machine learning models directly in the browser and Node.js environment, trains and deploys an event intelligent allocation model based on natural language processing, achieving convenience and real-time performance. When the user enters event information, the event intelligent allocation model sorts the event keywords from high to low according to the relevance, and provides the user with a choice of event types.

[0034] Alternatively, if Figure 2As shown, in the training and deployment phase of the event intelligent allocation model, time text data can be collected first, and then the weights and architecture of the pre-trained model can be loaded from the model library. Specifically, the weights of the pre-trained model can be updated using the gradient descent method to obtain the event intelligent allocation model. Here, the event intelligent allocation model is deployed on the front end based on the open source library TensorFlow.js. When input event information is detected, the input event information can be segmented according to the event intelligent allocation model to obtain a prediction result, and the event intelligent allocation model can be updated based on the input event information and the corresponding event type result obtained by interacting with the user.

[0035] Through the above steps S102 to S108, the weights and architecture of the pre-trained model are loaded from the model library, wherein the pre-trained model is a language representation model; the event text data is input into the pre-trained model, and the weights of the pre-trained model are updated using the gradient descent method to obtain an event intelligent distribution model, wherein the event intelligent distribution model training and deployment process is directly completed on the front end based on the open source library TensorFlow.js; when input event information is detected, the input event information is segmented according to the event intelligent distribution model to obtain a prediction result, wherein the prediction result includes an event type corresponding to the input event information; the event intelligent distribution model is updated based on the input event information and the corresponding event type result obtained by interaction with the user, which solves the problem that the user feedback mechanism is not perfect and the system is difficult to optimize according to the user feedback data. The event intelligent distribution model based on natural language processing achieves convenience and real-time performance.

[0036] In an exemplary embodiment, the method further comprises:

[0037] S11, converting the input event text data into word vectors through a word segmentation tool, wherein the event text data includes different types of event text descriptions and classification labels corresponding to the event text descriptions, and the word segmentation tool is used to convert the original continuous event text data into a numerical form.

[0038] In this embodiment, Figure 2 As shown in the figure, in order to transform the input event text data, you can first collect and organize the event text data for training, which contains different types of event descriptions and corresponding classification labels. The collected data set needs to be diverse and representative so that the model can learn the patterns of different events.

[0039] Furthermore, the text data can be segmented. The BertTokenizer segmentation tool officially provided by the Bidirectional Encoder Representations from Transformers (BERT) model can be used to convert the original continuous text data into a numerical form that the model can understand, as the input of the BERT model. BERT is a pre-trained language representation model that learns language representation through self-supervised learning and is suitable for various natural language processing tasks. BERT aims to accurately represent the semantic information in the text by introducing bidirectional context information, and can be quickly fine-tuned on a variety of downstream tasks. Download the BERT model from the model library officially provided by TensorFlow.js and load the model weights and architecture.

[0040] Through this embodiment, tensorflow.js is used to deploy an event intelligent distribution model based on natural language processing technology on the front end, thereby realizing the convenience of training and deployment of the event intelligent distribution model.

[0041] In an exemplary embodiment, event text data is input into a pre-trained model, and the weight of the pre-trained model is updated using a gradient descent method to obtain an event intelligent allocation model including:

[0042] S21, defining a cross entropy loss function and an adaptive moment estimation optimizer, wherein the cross entropy loss function is applicable to multi-classification tasks;

[0043] S22, based on the cross entropy loss function and the adaptive moment estimation optimizer, the weights of the pre-trained model are updated to obtain an event intelligent distribution model.

[0044] In the embodiment of the present application, when fine-tuning the model, the pre-trained BERT model can optionally be fine-tuned through an additional output layer without making extensive modifications to the task-specific architecture. During the fine-tuning process, pre-processed event text data is used as input data, and the model weights are updated using the gradient descent method through supervised learning. For the event intelligent allocation model, a cross-entropy loss function that is sensitive to probability distribution and suitable for multi-classification tasks is defined, and an Adam (adaptive moment estimation) optimizer with good convergence effect and high stability is defined to fine-tune the BERT model to achieve rapid convergence and good generalization performance of the model.

[0045] During the testing and optimization phase, you can use the test data set to test the performance of the model and make adjustments based on the feedback. You may need to return to the "model fine-tuning" step to adjust the model parameters to improve the accuracy and generalization of the model. In the process of model deployment and front-end integration, you can save the trained model to a local or online storage service, and write logic in the front-end JavaScript code to call the model for prediction. On the front-end page, when the user enters event information, the event intelligent allocation model performs word segmentation on the input information and gives the prediction results in real time. And based on the input data and feedback of users in actual applications, continuously iterate and improve the model, adjust the existing model parameters, and achieve continuous iteration.

[0046] In an exemplary implementation, the method further includes:

[0047] S31, setting hyper parameters based on input event information and selected event type, wherein the hyper parameters include number of iterations, learning rate, and batch size;

[0048] S32, based on the event type corresponding to the predicted input event information obtained by the event intelligent distribution model, the model parameters of the event intelligent distribution model are updated in combination with the cross entropy loss function and the adaptive moment estimation optimizer until the update stop condition is reached.

[0049] In the embodiment of the present application, Figure 3 An optional incremental learning flow chart provided in the embodiment of the present application is as follows: Figure 3 As shown in the figure, when performing reinforcement learning and model optimization, the model can be allowed to continuously learn new data based on existing knowledge without frequently retraining the entire model. The event information entered by the user and the event type selected are used as new input data to continuously update and optimize the model.

[0050] Specifically, you can determine the number of iterations, learning rate, batch size and other hyperparameters for incremental learning, and use the same loss function and optimizer as the original training. Randomly select a batch of data from the new dataset and input this batch of data into the event intelligent allocation model to obtain the prediction results. Use the loss function to calculate the gap between the prediction results and the true labels, and use the optimizer to update the model parameters according to the gradient. Repeat the process of predicting results and updating model parameters until the set number of iterations is reached or other stopping conditions are met. During the incremental learning process, regularly use the validation set to evaluate model performance to prevent the model from overfitting or underfitting.

[0051] Through this embodiment, according to the event information input by the user, multiple event type results are generated in real time by sorting from high to low according to the relevance. The user selects the event type, and the selection result is fed back in real time as new model input data. This can not only increase the efficiency of event processing, but also further train and optimize the model through user selection. The model can continuously learn and adapt to new data to achieve incremental learning.

[0052] In an exemplary embodiment, the method further comprises:

[0053] S41, the event intelligent allocation model sorts the event keywords corresponding to the input event information according to the sequence of high and low relevance and generates options for selecting event types;

[0054] S42, generating prompt information for prompting the user to select from the provided event type options.

[0055] In an embodiment of the present application, an event intelligent distribution model can be used to perform real-time event distribution and feedback. The event intelligent distribution model can be used to sort event keywords corresponding to input time information in a sequence of high and low relevance and generate options for selecting event type, and generate prompt information for prompting users to select from the provided event type options.

[0056] Specifically, when the user enters the event text information, the event intelligent allocation model sorts and displays the event keywords from high to low according to the relevance, and provides users with a choice of event types. For example, for urban governance events, when the user enters the event text information "There are electric vehicles on the roadway of the underground passage...", the keywords "traffic violations", "public security incidents", "urban and rural construction" and so on appear, and the user selects the most appropriate event type from them and submits it. This method can significantly improve the efficiency of event processing, allowing users and relevant departments to quickly locate the event type that needs to be processed.

[0057] The event text information entered by the user and the event type keywords selected are fed back to the event intelligent allocation model as new input data for further optimization of the model. This forms a closed-loop feedback mechanism that enables the model to continuously learn and adapt to new data environments.

[0058] Through this embodiment, the accuracy of the model is continuously optimized and enhanced by using real-time feedback from users as new input data, thereby achieving incremental learning, that is, the model can continuously learn and adapt to new data without the need for frequent retraining.

[0059] In an exemplary embodiment, when input event information is detected, performing word segmentation processing on the input event information according to the event intelligent allocation model to obtain a prediction result includes:

[0060] S51, writing logic in the front-end target code to call the event intelligent allocation model to perform word segmentation processing on the input event information to obtain the word segmentation processing result;

[0061] S52, predicting the event type based on the word segmentation processing result to obtain a prediction result;

[0062] S53, generating and displaying prediction information including prediction results on the front-end page.

[0063] In the embodiment of the present application, the trained model can be saved to a local or online storage service, and the logic calling model can be written in the front-end JavaScript code for prediction. In the front-end page, when the user enters event information, the event intelligent allocation model performs word segmentation on the input information and gives the prediction result in real time. According to the input data and feedback of users in actual applications, the model is continuously iterated and improved, and the existing model parameters are adjusted.

[0064] Through this embodiment, tensorflow.js is used to directly train and deploy the model on the front end without complex backend settings, which can achieve higher real-time interactivity. It can be completed directly on the client and is suitable for low-performance devices. It can achieve instant event classification and processing, and technically improve the distribution efficiency of the event intelligent distribution model.

[0065] According to another aspect of an embodiment of the present application, a distribution device for implementing the above-mentioned event intelligent distribution method based on machine learning is also provided. Figure 4 is a schematic diagram of the structure of an optional event intelligent distribution device based on machine learning according to an embodiment of the present application, such as Figure 4 As shown, the device may include:

[0066] A loading unit 402, used to load the weight and architecture of a pre-trained model from a model library, wherein the pre-trained model is a language representation model;

[0067] The acquisition unit 404 is used to input the event text data into the pre-trained model, and use the gradient descent method to update the weight of the pre-trained model to obtain an event intelligent allocation model, wherein the event intelligent allocation model training and deployment process is directly completed on the front end based on the open source library TensorFlow.js;

[0068] The prediction unit 406 is used to perform word segmentation processing on the input event information according to the event intelligent allocation model to obtain a prediction result when the input event information is detected, wherein the prediction result includes an event type corresponding to the input event information;

[0069] The first updating unit 408 is configured to update the event intelligent allocation model based on the input event information and the corresponding event type result obtained by interaction with the user.

[0070] It should be noted that the loading unit 402 in this embodiment can be used to execute the above step S102, the acquisition unit 404 in this embodiment can be used to execute the above step S104, the prediction unit 406 in this embodiment can be used to execute the above step S106, and the first update unit 408 in this embodiment can be used to execute the above step S108.

[0071] Through the above module, the weights and architecture of the pre-trained model are loaded from the model library, wherein the pre-trained model is a language representation model; the event text data is input into the pre-trained model, and the weights of the pre-trained model are updated using the gradient descent method to obtain an event intelligent distribution model, wherein the event intelligent distribution model training and deployment process is completed directly on the front end based on the open source library TensorFlow.js; when the input event information is detected, the input event information is segmented according to the event intelligent distribution model to obtain a prediction result, wherein the prediction result includes the event type corresponding to the input event information; the event intelligent distribution model is updated based on the input event information and the corresponding event type result obtained by interaction with the user, which solves the problem that the user feedback mechanism is not perfect and the system is difficult to optimize according to the user feedback data. The event intelligent distribution model based on natural language processing achieves convenience and real-time performance.

[0072] In an exemplary embodiment, the apparatus further comprises:

[0073] A conversion unit is used to convert the input event text data into a word vector through a word segmentation tool, wherein the event text data includes different types of event text descriptions and classification labels corresponding to the event text descriptions, and the word segmentation tool is used to convert the original continuous event text data into a numerical form.

[0074] In an exemplary embodiment, the acquiring unit includes:

[0075] A definition module, used to define a cross entropy loss function and an adaptive moment estimation optimizer, wherein the cross entropy loss function is applicable to multi-classification tasks;

[0076] An updating module is used to update the weight of the pre-trained model based on the cross entropy loss function and the adaptive moment estimation optimizer to obtain the event intelligent distribution model.

[0077] In an exemplary embodiment, the apparatus further comprises:

[0078] A setting unit, configured to set hyper parameters based on the input event information and the selected event type, wherein the hyper parameters include the number of iterations, the learning rate, and the batch size;

[0079] The second updating unit is used to update the model parameters of the event intelligent distribution model according to the event type corresponding to the predicted input event information obtained by the event intelligent distribution model, in combination with the cross entropy loss function and the adaptive moment estimation optimizer, until the update stop condition is reached.

[0080] In an exemplary embodiment, the apparatus further comprises:

[0081] A first generating unit, used for the event intelligent allocation model to sort the event keywords corresponding to the input event information according to a sequence of high and low relevance and generate an event type selection option;

[0082] The second generating unit is used to generate prompt information for prompting the user to select from the provided event type options.

[0083] In an exemplary embodiment, the prediction unit includes:

[0084] A word segmentation module is used to write logic in the front-end target code to call the event intelligent allocation model to perform word segmentation processing on the input event information to obtain a word segmentation processing result;

[0085] An acquisition module, used to predict the event type based on the word segmentation processing result to obtain the prediction result;

[0086] The generation module is used to generate and display prediction information including the prediction results in the front-end page.

[0087] It should be noted here that the examples and scenarios implemented by the above-mentioned modules and corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned embodiments. It should be noted that the above-mentioned modules as part of the device can run in a hardware environment and can be implemented by software or hardware, wherein the hardware environment includes a network environment.

[0088] According to another aspect of the embodiments of the present application, a storage medium is further provided. Optionally, in this embodiment, the storage medium can be used to execute the program code of any of the above-mentioned event intelligent allocation methods based on machine learning in the embodiments of the present application.

[0089] Optionally, in this embodiment, the storage medium is configured to store program codes for executing the following steps:

[0090] S1, load the weight and architecture of the pre-trained model from the model library, where the pre-trained model is a language representation model;

[0091] S2, input the event text data into the pre-trained model, use the gradient descent method to update the weight of the pre-trained model, and obtain the event intelligent allocation model. The event intelligent allocation model training and deployment process is completed directly on the front end based on the open source library TensorFlow.js;

[0092] S3, when input event information is detected, segmenting the input event information according to the event intelligent allocation model to obtain a prediction result, wherein the prediction result includes an event type corresponding to the input event information;

[0093] S4, updating the event intelligent allocation model based on the input event information and the corresponding event type results obtained through interaction with the user.

[0094] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, which will not be described in detail in this embodiment.

[0095] Among them, computer-readable storage media may include, but are not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0096] According to another aspect of an embodiment of the present application, an electronic device for implementing the above-mentioned event intelligent distribution method based on machine learning is also provided. The electronic device can be a server, a terminal, or a combination thereof.

[0097] Figure 5 is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application, such as Figure 5 As shown, it includes a processor 502, a communication interface 504, a memory 506 and a communication bus 508, wherein the processor 502, the communication interface 504, and the memory 506 communicate with each other through the communication bus 508, wherein:

[0098] Memory 506, used for storing computer programs;

[0099] The processor 502 is used to execute the computer program stored in the memory 506 to implement the following steps:

[0100] S1, load the weight and architecture of the pre-trained model from the model library, where the pre-trained model is a language representation model;

[0101] S2, input the event text data into the pre-trained model, use the gradient descent method to update the weight of the pre-trained model, and obtain the event intelligent allocation model. The event intelligent allocation model training and deployment process is completed directly on the front end based on the open source library TensorFlow.js;

[0102] S3, when input event information is detected, segmenting the input event information according to the event intelligent allocation model to obtain a prediction result, wherein the prediction result includes an event type corresponding to the input event information;

[0103] S4, updating the event intelligent allocation model based on the input event information and the corresponding event type results obtained through interaction with the user.

[0104] Optionally, the communication bus may be a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 The communication interface is used for communication between the electronic device and other devices.

[0105] The memory may include RAM, or may include non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0106] As an example, the memory 506 may include, but is not limited to, the loading unit 402, the acquisition unit 404, the prediction unit 406, and the first update unit 408 in the above-mentioned machine learning-based event intelligent distribution device. In addition, it may also include, but is not limited to, other module units in the above-mentioned machine learning-based event intelligent distribution device, which will not be repeated in this example.

[0107] The above-mentioned processor can be a general-purpose processor, which can include but not be limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; it can also be DSP (Digital Signal Processing), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0108] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0109] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0110] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0111] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0112] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0113] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0114] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, disk or optical disk and other media that can store program code.

[0115] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which can include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0116] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field not recorded in the present disclosure. The description and examples are regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

[0117] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0118] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An event intelligent allocation method based on machine learning, characterized in that: include: Loading the weights and architecture of a pre-trained model from a model library, wherein the pre-trained model is a language representation model; Input the event text data into the pre-trained model, and use the gradient descent method to update the weight of the pre-trained model to obtain an event intelligent allocation model, wherein the event intelligent allocation model training and deployment process is directly completed on the front end based on the open source library TensorFlow.js; When input event information is detected, segmenting the input event information according to the event intelligent allocation model to obtain a prediction result, wherein the prediction result includes an event type corresponding to the input event information; The event intelligent allocation model is updated based on the input event information and the corresponding event type result obtained by interaction with the user.

2. The method for intelligent event distribution based on machine learning as claimed in claim 1, characterized in that: Before loading the weights and architecture of the pre-trained model from the model library, the method further includes: The input event text data is converted into a word vector through a word segmentation tool, wherein the event text data includes different types of event text descriptions and classification labels corresponding to the event text descriptions, and the word segmentation tool is used to convert the original continuous event text data into a numerical form.

3. The method for intelligent event distribution based on machine learning as claimed in claim 1, characterized in that: The step of inputting the event text data into the pre-trained model and updating the weight of the pre-trained model using the gradient descent method to obtain the event intelligent allocation model includes: A cross entropy loss function is defined, and an adaptive moment estimation optimizer is defined, wherein the cross entropy loss function is applicable to multi-classification tasks; The weights of the pre-trained model are updated based on the cross entropy loss function and the adaptive moment estimation optimizer to obtain the event intelligent distribution model.

4. The method for intelligent event distribution based on machine learning as claimed in claim 3, characterized in that: After the weight of the pre-trained model is updated based on the cross entropy loss function and the adaptive moment estimation optimizer to obtain the event intelligent allocation model, the method further includes: Setting hyper parameters based on the input event information and the selected event type, wherein the hyper parameters include the number of iterations, the learning rate, and the batch size; According to the event type corresponding to the predicted input event information obtained by the event intelligent distribution model, the model parameters of the event intelligent distribution model are updated in combination with the cross entropy loss function and the adaptive moment estimation optimizer until the update stop condition is reached.

5. The method for intelligent event distribution based on machine learning as claimed in claim 1, characterized in that: In the case where input event information is detected, after performing word segmentation processing on the input event information according to the event intelligent allocation model to obtain a prediction result, the method further includes: The event intelligent allocation model sorts the event keywords corresponding to the input event information according to the sequence of high and low relevance and generates event type selection options; Prompt information is generated for prompting the user to select an event type from the provided event type selection options.

6. The method for intelligent event distribution based on machine learning as claimed in claim 1, characterized in that: When the input event information is detected, performing word segmentation processing on the input event information according to the event intelligent allocation model to obtain a prediction result includes: Writing logic in the front-end target code to call the event intelligent allocation model to perform word segmentation processing on the input event information to obtain a word segmentation processing result; Perform event type prediction based on the word segmentation processing result to obtain the prediction result; Generate and display prediction information including the prediction results on the front-end page.

7. An event intelligent distribution device based on machine learning, characterized in that: include: A loading unit, used to load the weight and architecture of a pre-trained model from a model library, wherein the pre-trained model is a language representation model; An acquisition unit is used to input event text data into the pre-trained model, and use the gradient descent method to update the weight of the pre-trained model to obtain an event intelligent allocation model, wherein the event intelligent allocation model training and deployment process is directly completed on the front end based on the open source library TensorFlow.js; A prediction unit, configured to, when input event information is detected, perform word segmentation processing on the input event information according to the event intelligent allocation model to obtain a prediction result, wherein the prediction result includes an event type corresponding to the input event information; The first updating unit is used to update the event intelligent allocation model based on the input event information and the corresponding event type result obtained by interaction with the user.

8. The event intelligent distribution device based on machine learning as claimed in claim 7, characterized in that: The device also includes: A conversion unit is used to convert the input event text data into a word vector through a word segmentation tool, wherein the event text data includes different types of event text descriptions and classification labels corresponding to the event text descriptions, and the word segmentation tool is used to convert the original continuous event text data into a numerical form.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 6 when executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 6 through the computer program.

Images