Intelligent meteorological answering model and system

Abstract

The invention provides an intelligent meteorological answering model and system, and relates to the technical field of meteorological services, and the method comprises the steps: obtaining meteorological data, and carrying out the preprocessing and integration of the meteorological data, so as to obtain a meteorological data set, and the meteorological data comprise satellite observation data, meteorological station data, and environment data submitted by a user; establishing a Transform model, performing pre-training by using the meteorological data set, and combining the pre-trained Transform model with a ReAct framework to obtain meteorological information; establishing an interaction module, obtaining a query signal through the interaction module, and identifying the query signal by using a natural language processing model to obtain a query intention; and acquiring meteorological information in real time according to the query intention, and transmitting the meteorological information to the user through the interaction module. According to the invention, the technical problems of low information response timeliness, poor user interaction experience, to-be-improved data analysis and prediction precision and the like in the existing meteorological service are solved.

Description

Technical Field

[0001] The present invention relates to the field of meteorological service technology, and in particular to an intelligent meteorological information answering model and system. Background Art

[0002] As society becomes more and more dependent on meteorological information, users have put forward higher requirements for the real-time and accuracy of meteorological services. In sudden weather events, fast and accurate weather forecasts are crucial to protecting people's lives and property. In recent years, artificial intelligence and machine learning technologies have been increasingly used in the field of meteorology. They can learn a large amount of historical meteorological data and extract useful features and information, thereby improving the accuracy and efficiency of weather forecasts. In order to obtain more comprehensive and accurate meteorological information, multi-source data fusion technology has become a hot topic in current research. This technology can effectively fuse meteorological data from different sensors and different observation platforms to improve the quality and availability of data. The current meteorological service platforms generally have the following shortcomings: long response time, users often need to wait for a long time to obtain specific meteorological information, affecting the timeliness of decision-making; unfriendly interactive interface: users often face complex interfaces and unintuitive operating procedures when using meteorological services; limited data processing and analysis capabilities: data processing in existing systems often cannot be performed in real time, and the accuracy of forecast results is limited by the limitations of the algorithms used. Summary of the invention

[0003] The present invention provides an intelligent weather information answering model and system, which solves the technical problems in existing weather services such as low information response timeliness, poor user interaction experience, and need to improve data analysis and prediction accuracy.

[0004] In order to solve the above technical problems, the technical solution of the present invention is as follows: In a first aspect, an intelligent weather information answering model and system includes: Acquire meteorological data, and perform preprocessing and integration to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, weather station data, and environmental data submitted by users; Establish a Transformer model and use the meteorological dataset for pre-training to obtain a pre-trained Transformer model; Combine the pre-trained Transformer model with the ReAct framework to obtain meteorological information; Establish an interactive module, obtain query signals through the interactive module, and use a natural language processing model to identify the query signals to obtain the query intent; According to the query intention, weather information is obtained in real time and transmitted to users through the interactive module; Obtain the user's query history and preference data, use machine learning combined with user portraits, and provide users with customized weather information.

[0005] Further, meteorological data is obtained, preprocessed and integrated to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, weather station data, and environmental data submitted by users, including: Access satellite observation data, weather station data, and user-submitted environmental data; Preprocessing satellite observation data, weather station data and environmental data submitted by users to obtain preprocessed data; The preprocessed data are integrated to obtain the meteorological data set.

[0006] Furthermore, a Transformer model is established and pre-trained using the meteorological dataset to obtain a pre-trained Transformer model, including: Based on the meteorological data set, a Transformer model is established; According to the meteorological data set, forward propagation is performed on each layer of the Transformer model to obtain the predicted output of the model; Use the cross entropy loss function to calculate the difference between the Transformer model's predicted output and the actual label to get the loss value; Back propagation is performed according to the loss value, the gradient of the Transformer model parameters is calculated, and the model parameters are updated to obtain the pre-trained Transformer model.

[0007] Furthermore, the pre-trained Transformer model is combined with the ReAct framework to obtain meteorological information, including: Load the pre-trained Transformer model into the ReAct framework; In the ReAct framework, the Transformer model is used to obtain the weather information forecast results and determine the required API parameters, including geographic location and time range; According to the API parameters, the ReAct framework extracts the meteorological information forecast results to obtain meteorological information.

[0008] Furthermore, an interaction module is established, and a query signal is obtained through the interaction module. The query signal is identified using a natural language processing model to obtain the query intent, including: Using the interactive module to obtain the query signal, and preprocessing it to obtain a preprocessed query signal; According to the preprocessed query signal, the natural language processing model performs language processing on the preprocessed query signal text to obtain a processed text, wherein the language processing includes word segmentation and part-of-speech tagging; The processed text is processed to obtain the query intent.

[0009] Furthermore, according to the query intent, weather information is obtained in real time and transmitted to the user through the interactive module, including: According to the query intent, determine the API interface to be called and the parameters to be passed, including the geographic location and timestamp; Obtain weather information in real time based on the API interface and the parameters passed; Analyze meteorological information to obtain key information, including temperature, humidity, wind speed, and weather conditions; Transmit key information to users through interactive modules.

[0010] Furthermore, we obtain the user's query history and preference data, use machine learning combined with user portraits, and provide users with customized weather information, including: Build a personalized profile of each user based on the acquired user query history and preference data, where the personalized profile includes the user's basic information, query habits and preferences; Based on the user's personalized portrait, use the machine learning algorithm for training to obtain a machine learning model; Based on the current user's profile and real-time weather data, a machine learning model is used to obtain customized weather information, including weather warnings, clothing suggestions, and travel tips; Customized weather information is pushed to users in real time through the interactive module.

[0011] An intelligent weather information answering system, comprising: An acquisition module is used to acquire meteorological data, and perform preprocessing and integration to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, meteorological station data, and environmental data submitted by users; Establish a module for establishing a Transformer model and pre-training it using a meteorological dataset to obtain a pre-trained Transformer model; The fusion module is used to combine the pre-trained Transformer model with the ReAct framework to obtain meteorological information; The processing module is used to establish an interactive module, obtain query signals through the interactive module, and use a natural language processing model to identify the query signals to obtain the query intent; obtain weather information in real time based on the query intent, and transmit it to the user through the interactive module; obtain the user's query history and preference data, use machine learning combined with user portraits, and provide users with customized weather information.

[0012] On the other hand, the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the intelligent weather information answering model as described above is implemented.

[0013] On the other hand, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned intelligent weather information response models.

[0014] On the other hand, the present invention further provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the intelligent weather information response model as described above is implemented.

[0015] The above solution of the present invention includes at least the following beneficial effects: According to the above scheme of the present invention, the intelligent weather information answering service can provide accurate weather information in real time, helping users make decisions in the shortest time, and can effectively support both daily life and emergency disaster response; the optimized system architecture and efficient algorithm processing ensure the rapid response of the service, and users can obtain the required weather data without waiting for a long time, thereby improving user experience and satisfaction; the design of an intuitive and easy-to-use interactive interface reduces the user's learning cost, so that users of different age groups and technical levels can easily get started and quickly query and obtain weather information; the interface supports personalized customization, and users can set the display content, layout and theme according to their needs and preferences to obtain a more intimate service experience; the use of intelligent algorithms to deeply mine and analyze massive meteorological data can more accurately capture the laws and trends of weather changes, thereby improving the accuracy of predictions and providing users with more reliable weather forecasts; the enhanced data processing and analysis capabilities enable the system to cope with a wider range of meteorological conditions and scenarios, including complex situations such as extreme weather and climate change, and provide users with more comprehensive meteorological services. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of the intelligent weather information answering model provided by an embodiment of the present invention.

[0017] Figure 2 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0019] like Figure 1 As shown, an embodiment of the present invention provides an intelligent weather information answering model, including: 11. Acquire meteorological data, and perform preprocessing and integration to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, weather station data, and environmental data submitted by users; 12. Establish a Transformer model and use the meteorological dataset for pre-training to obtain a pre-trained Transformer model; 13. Combine the pre-trained Transformer model with the ReAct framework to obtain meteorological information; 14. Establish an interactive module, obtain query signals through the interactive module, and use a natural language processing model to identify the query signals to obtain query intent; 15. According to the query intention, obtain weather information in real time and transmit it to the user through the interactive module; 16. Obtain the user's query history and preference data, use machine learning combined with user portraits, and provide users with customized weather information.

[0020] In an embodiment of the present invention, the collected raw meteorological data is preprocessed and integrated to form a structured meteorological data set, which improves the data quality and lays a solid foundation for subsequent model training and information extraction; utilizing the powerful feature extraction capability of the Transformer model, the system constructs a model architecture suitable for the meteorological field, and by using the integrated meteorological data set to pre-train the Transformer model, the model can learn the inherent laws and complex features of the meteorological data, thereby improving the performance of the model in meteorological information processing tasks; the system innovatively combines the pre-trained Transformer model with the ReAct framework, so that the model can respond more flexibly to changes in meteorological information and generate accurate and comprehensive meteorological information in real time; the introduction of the ReAct framework enhances the real-time and adaptability of the system, ensuring Ensure that users can obtain the latest and most accurate weather information at any time; the system has established a user-friendly interactive module to facilitate users to make weather query requests. Through the natural language processing model, the system can accurately capture the user's query intention by identifying and analyzing the user's query signal, providing an important basis for subsequent personalized services; according to the user's query intention, the system can obtain and generate corresponding weather information in real time. Through the interactive module, this weather information can be quickly and accurately transmitted to the user, meeting the user's demand for instant weather information; the system collects the user's query history and preference data, builds a user portrait, and uses machine learning technology to deeply explore user needs. Based on the analysis results of the user portrait and machine learning model, the system can provide users with more accurate and personalized weather information services, thereby improving user experience and satisfaction.

[0021] like Figure 1 As shown, 11, meteorological data is obtained, and preprocessed and integrated to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, weather station data, and environmental data submitted by users, including: 111, obtain satellite observation data, weather station data and user-submitted environmental data; 112, preprocessing the satellite observation data, the weather station data and the environmental data submitted by the user to obtain preprocessed data; 113, the preprocessed data are integrated to obtain a meteorological data set.

[0022] In the embodiment of the present invention, satellite observation data, weather station data and environmental data submitted by users are obtained. By collecting data from three different channels, namely satellites, weather stations and users, the comprehensiveness and diversity of meteorological information are ensured. Satellite observation data provides a wide range of macroscopic meteorological information, weather station data provides detailed observations at the ground level, and environmental data submitted by users adds real-time and region-specific information. This diversified data source enables the meteorological data set to contain richer information and more accurately reflect weather conditions and changing trends. The satellite observation data, weather station data and environmental data submitted by users are preprocessed to obtain preprocessed data. The preprocessing step usually includes data cleaning, which helps to remove noise, anomalies and other factors in the original data. Constant values ​​and missing data can be found to improve data quality. Through data standardization or normalization, data from different sources can be made comparable in value, which facilitates subsequent data integration and analysis. The original data can be converted into a form more suitable for analysis and model training as needed, such as time series data, image data, etc. The preprocessed data can be integrated to obtain a meteorological data set. The preprocessed data from different sources can be integrated to form a unified meteorological data set. Through integration, a more comprehensive, accurate and consistent meteorological data set can be obtained, which provides a high-quality data foundation for subsequent meteorological analysis, forecasting and services. The integrated meteorological data set is easier to manage and use, and can be directly input into the model for training or analysis, thereby improving work efficiency.

[0023] like Figure 1 As shown in 12, a Transformer model is established and pre-trained using a meteorological dataset to obtain a pre-trained Transformer model, including: 121, build a Transformer model based on the meteorological dataset; 122. According to the meteorological data set, forward propagation is performed on each layer of the Transformer model to obtain the predicted output of the model; 123, use the cross entropy loss function to calculate the difference between the Transformer model's predicted output and the actual label to get the loss value; 124, perform back propagation according to the loss value, calculate the gradient of the Transformer model parameters, and update the model parameters to obtain the pre-trained Transformer model.

[0024] In an embodiment of the present invention, a Transformer model is established according to a meteorological data set, and a suitable Transformer model structure is designed according to the characteristics and requirements of the meteorological data. The Transformer model is a deep learning model based on a self-attention mechanism, which is suitable for processing sequence data. When designing the model, it is necessary to consider parameters such as the dimension of the input data, the number of model layers, the number of heads (the number of heads in multi-head attention), and the dimension of the hidden layer. The meteorological data set may contain multiple features such as temperature, humidity, and wind speed. These data need to be properly preprocessed before they can be used as inputs to the Transformer model; according to the meteorological data set, each layer of the Transformer model is forward propagated to obtain the predicted output of the model, and the preprocessed meteorological data is input into the established Transformer model. The data is forward propagated through each layer of the model, including a multi-head self-attention layer, a feedforward neural network layer, etc., to finally obtain the predicted output of the model; the cross entropy loss function is used to calculate the difference between the predicted output of the Transformer model and the actual label to obtain the loss value. The cross entropy loss function is often used in classification problems to measure the difference between the probability distribution predicted by the model and the actual label, and to compare the predicted output of the model with the actual meteorological data. The data labels are compared and the loss value is calculated. The loss value reflects the accuracy of the model prediction: the smaller the loss value, the closer the model prediction is to the actual situation; back propagation is performed according to the loss value, the gradient of the Transformer model parameters is calculated, and the model parameters are updated to obtain the pre-trained Transformer model; back propagation is a commonly used optimization method in deep learning, which is used to adjust the parameters of the model according to the loss value to improve the prediction accuracy of the model. The model parameters are differentiated according to the calculated loss value to obtain the gradient of the parameters; an optimization algorithm (such as gradient descent, Adam, etc.) is used to update the model parameters according to the gradient. After multiple iterative updates, the model will gradually learn the inherent laws and patterns in the meteorological data, thereby improving its prediction accuracy. The final pre-trained Transformer model can be used for subsequent meteorological forecasting or classification tasks.

[0025] like Figure 1 As shown in 13, the pre-trained Transformer model is combined with the ReAct framework to obtain meteorological information, including: 131, load the pre-trained Transformer model into the ReAct framework; 132, in the ReAct framework, obtain the weather information forecast result through the Transformer model and determine the required API parameters, wherein the API parameters include the geographic location and the time range; 133. According to the API parameters, the ReAct framework extracts the meteorological information forecast results to obtain the meteorological information.

[0026] In an embodiment of the present invention, a pre-trained Transformer model is loaded into a ReAct framework. The ReAct framework may be a system framework for responsive data processing or application integration, supporting functions such as model loading, data processing, and result output. The Transformer model previously pre-trained with a meteorological data set is loaded into the ReAct framework so that it can predict and process meteorological information within the framework. In the ReAct framework, the meteorological information prediction result is obtained through the Transformer model and the required API parameters are determined. The API parameters include geographic location and time range. After the model is loaded, the ReAct framework will use the pre-trained Transformer model to process the input meteorological data and generate a prediction result of the meteorological information. Based on the demand or user input, the framework will determine what specific meteorological information needs to be obtained from the external API, which usually involves parameters such as geographic location (such as city, longitude and latitude, etc.) and time range (such as weather forecast for the next few hours or days); according to the API parameters, the ReAct framework extracts the meteorological information forecast results to obtain the meteorological information. With the API parameters, the ReAct framework will call the corresponding external meteorological service API and pass the parameters to obtain specific meteorological data; the framework will process the data returned from the API, which may include data cleaning, formatting, and integration with the prediction results of the Transformer model; finally, the ReAct framework will extract and generate meteorological information called, which may be weather forecasts, meteorological charts, or other forms of meteorological service products displayed in a user-friendly manner.

[0027] like Figure 1 As shown in 14, an interaction module is established, and a query signal is obtained through the interaction module, and the query signal is identified using a natural language processing model to obtain the query intent, including: 141, using the interactive module to obtain the query signal, and preprocessing it to obtain a preprocessed query signal; 142, according to the preprocessed query signal, the natural language processing model performs language processing on the preprocessed query signal text to obtain a processed text, wherein the language processing includes word segmentation and part-of-speech tagging; 143, the processed text is processed to obtain the query intent.

[0028] In the embodiment of the present invention, effect 14 describes in detail the process of obtaining user queries through the interaction module and identifying the query intent using the natural language processing model. The following are the detailed steps of the process: Use the interactive module to obtain the query signal and perform preprocessing to obtain the preprocessed query signal: The interaction module may be a user interface or an interface for receiving query signals input by a user, which signals are usually in text form, such as questions or requests input by the user.

[0029] In this step, the interaction module captures the original query signal input by the user and then preprocesses it. Preprocessing may include removing irrelevant characters, punctuation marks, stop words, and performing text normalization, etc., in order to improve the accuracy of subsequent natural language processing; based on the preprocessed query signal, the natural language processing model performs language processing on the preprocessed query signal text to obtain the processed text, and the language processing includes word segmentation and part-of-speech tagging. The preprocessed query signal is sent to the natural language processing model for further language analysis. Word segmentation is to divide continuous text into independent vocabulary units, which is especially important for languages ​​such as Chinese that do not have obvious vocabulary boundaries. Part-of-speech tagging is to assign a grammatical category to each vocabulary, which helps to understand the relationship between sentence structure and vocabulary. Through these language processing The model can understand the semantic content of the query signal more accurately through the processing step; the processed text is analyzed to obtain the query intent. The model conducts in-depth analysis on the text after word segmentation and part-of-speech tagging in order to identify the user's query intent. The query intent is the information the user wants to obtain or the operation he wants to perform when making a query. For example, the user may want to check the weather conditions, order air tickets, or learn more about a product. The model may use rule-based methods, machine learning algorithms, or deep learning technologies to identify query intent. The specific method depends on the complexity of the model and the quality of the training data; finally, the model will output one or more possible query intents, which will be used for subsequent information retrieval or service response.

[0030] like Figure 1 As shown in 15, according to the query intention, the meteorological information is obtained in real time and transmitted to the user through the interactive module, including: 151. Determine the API interface to be called and the parameters to be passed according to the query intent, wherein the parameters include a geographic location and a timestamp; 152, obtain weather information in real time according to the API interface and the parameters passed; 153, analyzing the meteorological information to obtain key information, wherein the key information includes temperature, humidity, wind speed, and weather conditions; 154, transmits the key information to the user through the interactive module.

[0031] In an embodiment of the present invention, the API interface to be called and the parameters to be passed are determined according to the query intent, and the parameters include the geographic location and timestamp. The system first analyzes the user's query intent and determines the type of meteorological information the user wants to obtain. According to the query intent, the system determines the specific API interface to be called. For example, if the user wants to query the current weather in a city, the system may select an API interface that provides real-time weather data. The system also determines the parameters that need to be passed to the API interface so that the API can return meteorological information for a specific area. In addition, the timestamp may also be passed as a parameter to request meteorological data for a specific point in time. According to the API interface and the passed parameters, the meteorological information is obtained in real time. The system obtains meteorological information in real time by calling the determined API interface and passing the corresponding parameters, and communicates with the server of the external meteorological service provider to obtain the latest meteorological data. "Real-time" means that the system can instantly obtain and update meteorological information , to meet the user's demand for the latest weather conditions; parse the meteorological information to obtain key information, which includes temperature, humidity, wind speed, and weather conditions. Once the meteorological information is obtained from the API interface, the system will parse and process this information. The parsing process mainly extracts key information from the meteorological data, such as temperature, humidity, wind speed, and weather conditions, which have an important impact on the user's daily activities and decisions; the system may organize and store this key information in a specific format or data structure to facilitate subsequent display and transmission, and transmit the key information to the user through an interactive module. Finally, the system transmits the parsed key meteorological information to the user through the interactive module. The interactive module may be a user interface (UI), which presents the meteorological information to the user intuitively through a graphical display method (such as a weather chart, icon, etc.), or it may be through other forms of output (such as voice broadcast, SMS notification, etc.) to ensure that the user can easily receive the required meteorological information.

[0032] like Figure 1 As shown in 16, the user's query history and preference data are obtained, and machine learning is used in combination with user portraits to provide users with customized weather information, including: 161. Build a personalized profile of each user based on the acquired user query history and preference data, where the personalized profile of each user includes basic information, query habits and preferences of the user; 162. Based on the user's personalized portrait, a machine learning algorithm is used for training to obtain a machine learning model; 163. Based on the current user's profile and real-time weather data, a machine learning model is used to obtain customized weather information, wherein the customized weather information includes weather warnings, clothing suggestions, and travel tips; 164, push customized weather information to users in real time through the interactive module.

[0033] In an embodiment of the present invention, a personalized portrait of each user is constructed based on the acquired user query history and preference data. The system first collects and analyzes the user's query history and preference data. These data may include cities, time periods, weather types, etc. that users frequently query, as well as the user's attention or feedback on specific weather information. Based on these data, the system constructs a personalized portrait for each user. This portrait not only contains the user's basic information (such as age, gender, geographic location, etc.), but also deeply reflects the user's query habits and preferences (such as frequently querying the morning weather, paying special attention to rainy and snowy weather, etc.); according to the user's personalized portrait, a machine learning algorithm is used for training to obtain a machine learning model. Using the user's personalized portrait as training data, the system selects a suitable machine learning algorithm for model training, which can learn and understand the user's personalized needs and preferences, thereby laying the foundation for the subsequent provision of customized meteorological information; according to the current user's portrait and real-time weather data, a machine learning model is used to obtain customized meteorological information. When a user requests weather information, the system will use the user's personalized portrait and the current real-time weather data to obtain customized meteorological information. The system uses real-time weather data and previously trained machine learning models to generate customized weather information. The customized weather information includes weather warnings for users (such as providing weather warnings at the destination for users who travel frequently), clothing suggestions (providing appropriate clothing recommendations based on the user's age and gender), travel tips (such as providing transportation suggestions based on the user's travel habits and current weather conditions), etc. The customized weather information is pushed to the user in real time through the interactive module. Once the customized weather information is generated, the system will immediately push this information to the user in real time through the interactive module. The interactive module may be a mobile application, web interface, or other form of user interface to ensure that the user can receive this personalized weather information conveniently and promptly.

[0034] Figure 2 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0035] like Figure 2 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the intelligent weather information answering model and system.

[0036] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0037] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the above-mentioned intelligent weather information response model and system.

[0038] On the other hand, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the above-mentioned intelligent weather information answering models and systems.

[0039] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0040] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the various embodiments or some parts of the embodiments.

[0041] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An intelligent weather information answering model, characterized in that: include: Acquire meteorological data, and perform preprocessing and integration to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, weather station data, and environmental data submitted by users; Establish a Transformer model and use the meteorological dataset for pre-training to obtain a pre-trained Transformer model; Combine the pre-trained Transformer model with the ReAct framework to obtain meteorological information; Establish an interactive module, obtain query signals through the interactive module, and use a natural language processing model to identify the query signals to obtain the query intent; According to the query intention, weather information is obtained in real time and transmitted to users through the interactive module; Obtain the user's query history and preference data, use machine learning combined with user portraits, and provide users with customized weather information.

2. The intelligent weather information model according to claim 1, characterized in that: Acquire meteorological data, preprocess and integrate them to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, weather station data, and environmental data submitted by users, including: Access satellite observation data, weather station data, and user-submitted environmental data; Preprocessing satellite observation data, weather station data and environmental data submitted by users to obtain preprocessed data; The preprocessed data are integrated to obtain the meteorological data set.

3. The intelligent weather information answering model according to claim 2 is characterized in that: Build a Transformer model and use the meteorological dataset for pre-training to obtain a pre-trained Transformer model, including: Based on the meteorological data set, a Transformer model is established; According to the meteorological data set, forward propagation is performed on each layer of the Transformer model to obtain the predicted output of the model; Use the cross entropy loss function to calculate the difference between the Transformer model's predicted output and the actual label to get the loss value; Back propagation is performed according to the loss value, the gradient of the Transformer model parameters is calculated, and the model parameters are updated to obtain the pre-trained Transformer model.

4. The intelligent weather information answering model according to claim 3 is characterized in that: Combine the pre-trained Transformer model with the ReAct framework to obtain meteorological information, including: Load the pre-trained Transformer model into the ReAct framework; In the ReAct framework, the Transformer model is used to obtain the weather information forecast results and determine the required API parameters, including geographic location and time range; According to the API parameters, the ReAct framework extracts the meteorological information forecast results to obtain meteorological information.

5. The intelligent weather information answering model according to claim 4 is characterized in that: An interactive module is established, and query signals are obtained through the interactive module. The query signals are identified using a natural language processing model to obtain query intent, including: Using the interactive module to obtain the query signal, and preprocessing it to obtain a preprocessed query signal; According to the preprocessed query signal, the natural language processing model performs language processing on the preprocessed query signal text to obtain a processed text, wherein the language processing includes word segmentation and part-of-speech tagging; The processed text is processed to obtain the query intent.

6. The intelligent weather information model according to claim 5, characterized in that: According to the query intent, weather information is obtained in real time and transmitted to users through interactive modules, including: According to the query intent, determine the API interface to be called and the parameters to be passed, including the geographic location and timestamp; Obtain weather information in real time based on the API interface and the parameters passed; Analyze meteorological information to obtain key information, including temperature, humidity, wind speed, and weather conditions; Transmit key information to users through interactive modules.

7. The intelligent weather information model according to claim 6, characterized in that: Obtain the user's query history and preference data, use machine learning combined with user portraits to provide users with customized weather information, including: Build a personalized profile of each user based on the acquired user query history and preference data, where the personalized profile includes the user's basic information, query habits and preferences; Based on the user's personalized portrait, use the machine learning algorithm for training to obtain a machine learning model; Based on the current user's profile and real-time weather data, a machine learning model is used to obtain customized weather information, including weather warnings, clothing suggestions, and travel tips; Customized weather information is pushed to users in real time through the interactive module.

8. An intelligent weather information answering system, characterized in that: include: An acquisition module is used to acquire meteorological data, and perform preprocessing and integration to obtain a meteorological data set, wherein the meteorological data includes satellite observation data, meteorological station data, and environmental data submitted by users; Establish a module for establishing a Transformer model and pre-training it using a meteorological dataset to obtain a pre-trained Transformer model; The fusion module is used to combine the pre-trained Transformer model with the ReAct framework to obtain meteorological information; A processing module is used to establish an interaction module, obtain a query signal through the interaction module, and use a natural language processing model to identify the query signal to obtain the query intent; According to the query intention, weather information is obtained in real time and transmitted to users through the interactive module; Obtain the user's query history and preference data, use machine learning combined with user portraits, and provide users with customized weather information.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the intelligent weather information answering model as claimed in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the intelligent weather information answering model as claimed in any one of claims 1 to 7 is implemented.

Images