Geographic learning enhancement method based on artificial intelligence

Abstract

The invention relates to an artificial intelligence-based geographic learning enhancement method, which relates to the technical field of artificial intelligence, and comprises the following steps: collecting geographic data and student learning behavior data, carrying out preprocessing and feature extraction on the data, storing the data in a cloud database, and constructing a generative adversarial network model; pre-processed geographic data and student data are used as a training set to perform deep learning training, the generative adversarial network model fuses multi-modal learning and a geographic knowledge graph, a reinforcement learning mechanism is adopted to continuously optimize a generator and a discriminator, and based on the learning progress, interest and demand of students, the learning process is optimized; the method comprises the steps of generating personalized geographic learning content through a trained generative adversarial network model, continuously collecting interaction data and feedback of students in a learning process, adjusting a learning path according to behaviors, answer results and feedback information of the students, forming a dynamic feedback mechanism, and realizing a personalized learning plan.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology. More specifically, the present invention relates to a method for enhancing geographical learning based on artificial intelligence. Background Art

[0002] With the continuous development of artificial intelligence technology, especially the breakthroughs in the fields of deep learning, natural language processing, and intelligent recommendation, the application of artificial intelligence in education has become increasingly widespread. Especially in the complex and interdisciplinary field of geography, the role of artificial intelligence has gradually emerged.

[0003] In the traditional process of geographical learning, students often rely on static resources such as books and maps. The learning method is relatively single, lacking interactivity and personalized support. While the geographical learning method based on artificial intelligence can better adapt to the different needs and learning styles of students through functions such as simulating human cognition, analyzing big data, and automated learning. Artificial intelligence technology can not only provide personalized learning plans according to the learning progress and comprehension ability of students, but also adjust the learning content through real-time feedback, making the learning process more flexible and efficient. Summary of the Invention

[0004] The present invention aims at the technical problems existing in the prior art, and provides a method for enhancing geographical learning based on artificial intelligence to solve the problems raised in the above background art.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: A method for enhancing geographical learning based on artificial intelligence specifically includes the following steps: Step 101: Collect geographical data and students' learning behavior data, preprocess and extract features of the data, and store them in a cloud database. The geographical data includes dynamic map data, climate data, urbanization process, and population distribution; the students' learning behavior data includes learning progress, problem feedback, and error types; Step 102: Construct a generative adversarial network model, and use the preprocessed geographical data and student data as a training set for deep learning training. The generative adversarial network model integrates multi-modal learning and geographical knowledge graph, and continuously optimizes the generator and discriminator using a reinforcement learning mechanism; Step 103: Based on the learning progress, interests, and needs of students, generate personalized geographical learning content through the trained generative adversarial network model. The geographical learning content includes scenario simulation, virtual experiment, and customized case analysis; Step 104: Continuously collect students' interaction data and feedback during the learning process, adjust the learning path according to students' behaviors, answering results, and feedback information, form a dynamic feedback mechanism, and realize a personalized learning plan.

[0006] In a preferred embodiment, in step 101, geographical data and students' learning behavior data are collected, preprocessed and feature extracted, and stored in a cloud database. The geographical data includes dynamic map data, climate data, urbanization process and population distribution; the students' learning behavior data includes learning progress, problem feedback and error types. The specific steps are as follows: Step A1, data collection: Obtain geographical data including dynamic map data, climate data, urbanization process and population distribution from publicly available data on the Internet through an API interface. Use an online education platform to collect students' learning progress, including completed course modules and test scores, and use an automated script to regularly record students' learning logs, including feedback documents submitted by students during the learning process, as well as the types and frequencies of errors made by students in tests; Step A2, data processing and storage: Data cleaning is performed on the geographical data and students' learning behavior data by removing duplicate data, filling in missing values and detecting outliers. The geographical data is converted to a geographical coordinate system, and climate patterns, urbanization speed, and population density features are extracted from the geographical data. Error frequency, learning duration, and feedback emotion features are extracted from the students' learning behavior data, and the extracted features are stored through a cloud database; Among them, the steps of the feedback emotion feature include: converting the feedback documents submitted by students into text format, removing punctuation marks, special characters and stop words, using a sentiment analysis model to perform sentiment scoring on the feedback text. For each feedback text, calculate its sentiment score as , where represents the sentiment score of the text, is the i-th word in the text, is the score of the word in the sentiment dictionary. According to the sentiment score, analyze the emotional tendency of the students' feedback. When the sentiment score , classify the feedback as positive. When the sentiment score , classify the feedback as negative. When the sentiment score , classify the feedback as neutral.

[0007] In a preferred embodiment, in step 102, a generative adversarial network model is constructed, and the preprocessed geographical data and student data are used as a training set for deep learning training. The generative adversarial network model integrates multi-modal learning and a geographical knowledge graph, and uses a reinforcement learning mechanism to continuously optimize the generator and discriminator. The specific steps are as follows: Step B1. Multimodal data fusion: Extract geographical entities from geographical data and label them. Using the extracted geographical entities, combined with the urbanization process and climate patterns, construct a geographical knowledge graph. Among them, the nodes of the knowledge graph are geographical entities and attributes, and the edges represent the relationships between entities. And adopt the graph embedding method to embed the nodes and edges in the knowledge graph into a continuous vector space to generate a vector representation of the graph. Fuse the feature vectors extracted from geographical data and students' learning behaviors, as well as the vector representation embedded in the geographical knowledge graph to form a multimodal dataset; Step B2. Construct a generative adversarial network model: Design a generator and a discriminator. By inputting the multimodal dataset into the generator, obtain the predicted learning behavior data of students. Input the predicted learning behavior data generated by the generator into the discriminator, and output a probability value indicating the probability that the input data comes from the true distribution. It further includes the following steps: Step B201. The network structure of the generator includes an input layer, a hidden layer, and an output layer. The input layer inputs the fused multimodal dataset X. The hidden layer extracts feature information from the input. The output layer generates the predicted learning behavior data of students. Set the size of the hidden layer to and , The formula from the input layer to the first hidden layer is: , The formula from the first hidden layer to the second hidden layer is: , The formula from the second hidden layer to the output layer is: , where is the weight matrix from the input layer to the first hidden layer, is the weight matrix from the first hidden layer to the second hidden layer, is the weight matrix from the second hidden layer to the output layer, f is the ReLU activation function, is the predicted learning behavior data of students generated, is the bias term from the input layer to the first hidden layer, is the bias term from the first hidden layer to the second hidden layer, is the bias term of the output layer; Step B202. The discriminator receives the predicted learning behavior data generated by the generator through the input layer, and outputs a probability value through the output layer. The specific formula is , where is the sigmoid activation function, which is used to map the output to the range, P is the true probability value output by the discriminator, are the weight matrices of each layer of the discriminator respectively, are the bias terms of each layer of the discriminator respectively; Step B3. Reinforcement learning optimization: Use real data and generated data for training, update the parameters of the generator and discriminator, and continuously optimize the generator and discriminator through the reward mechanism, which further includes the following steps: Step B301. Combine real data and generated data to form a training set, and define the reward function of the generator based on the probability value output by the discriminator , to evaluate the performance of the generator, where is the generated data, R is the reward function, is the discrimination probability of the discriminator for the generated data d; Step B302. Calculate the reward R according to the discrimination results of real data and generated data, and update the parameters of the generator using the reward information: , where is the learning rate, is the parameter updated by the generator, is the parameter of the generator, and R is the reward function; Step B303. Repeat the above steps, continuously compare the data generated by the generator with real data, optimize the generator and discriminator, and improve the model's prediction ability for learning behaviors.

[0008] In a preferred embodiment, in step 103, based on the student's learning progress, interests, and needs, personalized geographical learning content is generated through a trained generative adversarial network model. The geographical learning content includes scenario simulation, virtual experiment, and customized case analysis. The specific steps are as follows: Step C1. Represent the student's learning progress, interests, and needs as , extract relevant geographical features from the geographical knowledge graph to generate a knowledge vector: , and fuse the student feature S and the geographical knowledge feature to form an input vector , input the fused input vector Z into the trained generator to generate personalized learning content , where C is the generated learning content, is the feature fusion function, is the vector representation of the learning data, is the vector representation of the geographical entity, G represents the generator function, is the parameter of the generator; Step C2. According to the information in the input vector Z, generate geographical learning content of the types of scenario simulation, virtual experiment, and customized case analysis. The scenario simulation content related to the geographical environment is expressed as: ; Generate an experimental scenario related to geographical concepts by simulating ecological changes under different climate conditions, which is expressed as: ; Generate case analyses for specific geographical problems according to students' interests and needs, expressed as: ; Integrate the generated parts into a comprehensive learning module , and output a personalized geographical learning content module for students to study and explore.

[0009] In a preferred embodiment, in step 104, during the learning process, continuously collect students' interaction data and feedback, and adjust the learning path according to students' behaviors, answering results, and feedback information to form a dynamic feedback mechanism and implement a personalized learning plan. The specific steps are as follows: Step D1: Use an online platform to record the interaction data generated by students during the learning process, including learning duration, number of visits, answering results, students' evaluations, opinions, and suggestions on the learning content. Organize and analyze the recorded data, extract the learning efficiency and interest index, and calculate the students' learning satisfaction as , where J is the number of correct answers of the student in the test, T is the total learning time, is the time spent by the student in the jth visit, is the number of times the student conducts learning, is the upper limit of the number of visits, and are weight coefficients; Step D2: Evaluate the students' learning status based on the calculated learning satisfaction to form feedback information. According to the students' satisfaction and feedback information, implement the adjustment strategy of the learning path, combine the students' behavior patterns, interest points, and feedback information, generate personalized learning recommendation content, and push it to the students, expressed as , where, represents the personalized learning recommendation content, is the generator model, is the input feature vector, including the students' learning behaviors and feedback information.

[0010] The beneficial effects of the present invention are as follows: collecting geographical data and students' learning behavior data, preprocessing and feature extracting the data, storing it in a cloud database, constructing a generative adversarial network model, using the preprocessed geographical data and students' data as a training set for deep learning training. The generative adversarial network model integrates multi-modal learning and a geographical knowledge graph, and continuously optimizes the generator and discriminator using a reinforcement learning mechanism. Based on the students' learning progress, interests, and needs, personalized geographical learning content is generated through the trained generative adversarial network model. During the learning process, students' interaction data and feedback are continuously collected, and the learning path is adjusted according to the students' behaviors, answering results, and feedback information, forming a dynamic feedback mechanism to achieve a personalized learning plan. The present invention utilizes a real-time feedback mechanism to automatically adjust the learning path, making the teaching process more flexible and adaptable to the individual differences of students, enhancing the adaptability of learning, integrating the geographical knowledge graph with different forms of data, and being able to provide diverse learning resources, improving the students' learning experience and comprehension ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of protection of the present application.

[0013] In the description of the present application, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present application, "a plurality" means two or more, unless otherwise specifically defined.

[0014] In the description of the present application, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described in the present application as "for example" is not necessarily to be construed as more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without the use of these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but rather to be in line with the broadest scope consistent with the principles and features disclosed in the present application.

[0015] Embodiment 1 This embodiment provides a method for enhancing geographical learning based on artificial intelligence as shown in Figure 1 the following, which specifically includes the following steps: Step 101: Collect geographical data and students' learning behavior data, preprocess and extract features from the data, and store them in a cloud database. The geographical data includes dynamic map data, climate data, urbanization process, and population distribution; the students' learning behavior data includes learning progress, problem feedback, and error types. Step 102: Construct a generative adversarial network model, and use the preprocessed geographical data and student data as a training set for deep learning training. The generative adversarial network model integrates multimodal learning and a geographical knowledge graph, and continuously optimizes the generator and discriminator using a reinforcement learning mechanism. Step 103: Based on the students' learning progress, interests, and needs, generate personalized geographical learning content through the trained generative adversarial network model. The geographical learning content includes scenario simulation, virtual experiment, and customized case analysis. Step 104: Continuously collect students' interaction data and feedback during the learning process, adjust the learning path according to the students' behaviors, answer results, and feedback information, and form a dynamic feedback mechanism to achieve a personalized learning plan.

[0016] Preferably, in step 101, collect geographical data and students' learning behavior data, preprocess and extract features from the data, and store them in a cloud database. The geographical data includes dynamic map data, climate data, urbanization process, and population distribution; the students' learning behavior data includes learning progress, problem feedback, and error types. The specific steps are as follows: Step A1, Data Collection: Obtain geographical data including dynamic map data, climate data, urbanization process, and population distribution from publicly available data on the Internet through an API interface. Use an online education platform to collect students' learning progress, including completed course modules and test scores, and use an automated script to regularly record students' learning logs, including feedback documents submitted by students during the learning process, as well as the types and frequencies of errors made by students in tests. Step A2, Data Processing and Storage: Clean the geographical data and students' learning behavior data by removing duplicate data, filling in missing values, and detecting outliers. Convert the geographical data to a geographical coordinate system, and extract climate patterns, urbanization speed, and population density features from the geographical data. Extract error frequency, learning duration, and feedback emotion features from students' learning behavior data, and store the extracted features in a cloud database. Among them, the steps for the feedback emotion feature include: Convert the feedback documents submitted by students into text format, remove punctuation marks, special characters, and stop words, and use a sentiment analysis model to perform sentiment scoring on the feedback text. For each feedback text, calculate its sentiment score as , where represents the sentiment score of the text, is the i-th word in the text, is the score of this word in the sentiment dictionary. According to the sentiment score, analyze the emotional tendency of students' feedback. When the sentiment score , classify the feedback as positive. When the sentiment score , classify the feedback as negative. When the sentiment score , classify the feedback as neutral.

[0017] Preferably, in step 102, construct a generative adversarial network model, and use the preprocessed geographical data and student data as the training set for deep learning training. The generative adversarial network model integrates multi-modal learning and a geographical knowledge graph, and uses a reinforcement learning mechanism to continuously optimize the generator and discriminator. The specific steps are as follows: Step B1, Multi-modal Data Fusion: Extract geographical entities from the geographical data and label them. Use the extracted geographical entities, combined with the urbanization process and climate patterns, to construct a geographical knowledge graph. Among them, the nodes of the knowledge graph are geographical entities and attributes, and the edges represent the relationships between entities. Use a graph embedding method to embed the nodes and edges in the knowledge graph into a continuous vector space to generate a vector representation of the graph. Integrate the feature vectors extracted from the geographical data and students' learning behavior, as well as the vector representation embedded in the geographical knowledge graph, to form a multi-modal data set. Step B2. Construct a generative adversarial network model: Design a generator and a discriminator. By inputting the multi-modal dataset into the generator, the predicted learning behavior data of the students is obtained. Then, the predicted learning behavior data generated by the generator is input into the discriminator, and a probability value is output, indicating the probability that the input data comes from the true distribution. This further includes the following steps: Step B201. The network structure of the generator includes an input layer, a hidden layer, and an output layer. The input layer is the fused multi-modal dataset X after input. The hidden layer extracts feature information from the input, and the output layer generates the predicted learning behavior data of the students. Set the size of the hidden layer to and . The formula from the input layer to the first hidden layer is: . The formula from the first hidden layer to the second hidden layer is: . The formula from the second hidden layer to the output layer is: , where is the weight matrix from the input layer to the first hidden layer, is the weight matrix from the first hidden layer to the second hidden layer, is the weight matrix from the second hidden layer to the output layer, f is the ReLU activation function, is the generated predicted learning behavior data of the students, is the bias term from the input layer to the first hidden layer, is the bias term from the first hidden layer to the second hidden layer, is the bias term of the output layer; Step B202. The discriminator receives the predicted learning behavior data generated by the generator through the input layer, and outputs a probability value through the output layer. The specific formula is , where is the sigmoid activation function, which is used to map the output to the range. P is the true probability value output by the discriminator, are the weight matrices of each layer of the discriminator respectively, are the bias terms of each layer of the discriminator respectively; Step B3. Reinforcement learning optimization: Use the real data and the generated data for training, update the parameters of the generator and the discriminator, and continuously optimize the ability of the generator to generate the predicted learning behavior data and the recognition ability of the discriminator through the reward mechanism. This further includes the following steps: Step B301. Combine the real data and the generated data to form a training set, and define the reward function of the generator based on the probability value output by the discriminator to evaluate the performance of the generator. Among them, is the generated data, R is the reward function, is the discrimination probability of the discriminator for the generated data d; Step B302: Calculate the reward R based on the discrimination results of the real data and the generated data, and update the parameters of the generator using the reward information: , where is the learning rate, are the parameters updated by the generator, are the parameters of the generator, and R is the reward function; Step B303: Repeat the above steps, continuously compare the data generated by the generator with the real data, optimize the generator and the discriminator, and improve the model's prediction ability for learning behaviors.

[0018] Preferably, in step 103, based on the student's learning progress, interests, and needs, personalized geography learning content is generated through a trained generative adversarial network model. The geography learning content includes scenario simulation, virtual experiments, and customized case analysis. The specific steps are as follows: Step C1: Represent the student's learning progress, interests, and needs as , extract relevant geographical features from the geographical knowledge graph to generate knowledge vectors: , and fuse the student feature S and the geographical knowledge feature to form an input vector . Input the fused input vector Z into the trained generator to generate personalized learning content , where C is the generated learning content, is the feature fusion function, is the vector representation of the learning data, is the vector representation of the geographical entity, G represents the generator function, are the parameters of the generator; Step C2: According to the information in the input vector Z, generate geography learning content of the types of scenario simulation, virtual experiment, and customized case analysis. The scenario simulation content related to the geographical environment is expressed as: ; Generate experimental scenarios related to geographical concepts by simulating ecological changes under different climate conditions, expressed as: ; Generate case analysis for specific geographical problems according to the student's interests and needs, expressed as: ; And integrate the generated parts of the content into a comprehensive learning module , and output a personalized geography learning content module for the student to learn and explore.

[0019] Preferably, in step 104, during the learning process, the interaction data and feedback of students are continuously collected. According to the students' behaviors, answering results, and feedback information, the learning path is adjusted to form a dynamic feedback mechanism and implement a personalized learning plan. The specific steps are as follows: Step D1: Use an online platform to record the interaction data generated by students during the learning process, including learning duration, access times, answering results, students' evaluations, opinions, and suggestions on the learning content. Organize and analyze the recorded data, extract the learning efficiency and interest index, and calculate the learning satisfaction of students based on the collected data and analysis results as , where J is the number of correct answers of the student in the test, T is the total learning time, is the time spent by the student in the j-th access, is the number of times the student conducts learning, is the upper limit of access times, and are weight coefficients; Step D2: According to the calculated learning satisfaction, evaluate the learning status of students to form feedback information. According to the students' satisfaction and feedback information, implement the adjustment strategy of the learning path, combine the students' behavior patterns, interest points, and feedback information to generate personalized learning recommendation content, and push it to the students, expressed as , where, represents the personalized learning recommendation content, is the generator model, is the input feature vector, including the learning behaviors and feedback information of students.

[0020] It should be noted that in the above embodiments, the descriptions of each embodiment have their own focuses. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0021] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0022] The present invention will be described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0023] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0024] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0025] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.

[0026] Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A geographic learning enhancement method based on artificial intelligence, characterized in that: The specific steps include: Step 101: Collect geographic data and student learning behavior data, pre-process and extract features of the data, and store them in a cloud database, wherein the geographic data includes dynamic map data, climate data, urbanization process, and population distribution; and the student learning behavior data includes learning progress, problem feedback, and error type. Step 102: construct a generative adversarial network model, and use the preprocessed geographic data and student data as training sets to perform deep learning training. The generative adversarial network model integrates multimodal learning and geographic knowledge graphs, and uses a reinforcement learning mechanism to continuously optimize the generator and the discriminator. Step 103: Generate personalized geography learning content based on the student's learning progress, interests and needs through the trained generative adversarial network model, where the geography learning content includes scenario simulation, virtual experiments and customized case analysis; Step 104: Continuously collect students' interactive data and feedback during the learning process, adjust the learning path according to students' behaviors, answer results and feedback information, form a dynamic feedback mechanism, and implement a personalized learning plan.

2. The method for enhancing geographical learning based on artificial intelligence according to claim 1, characterized in that: In step 101, geographic data and student learning behavior data are collected, preprocessed and feature extracted, and stored in a cloud database. The geographic data includes dynamic map data, climate data, urbanization process, and population distribution; the student learning behavior data includes learning progress, problem feedback, and error type. The specific steps are as follows: Step A1, data collection: Obtain geographic data from the Internet through the API interface, use the online education platform to collect students' learning progress, including completed course modules and test scores, and use automated scripts to regularly record students' learning logs, including feedback documents submitted by students during the learning process, and the types and frequencies of students' errors in tests; Step A2, data processing and storage: perform data cleaning on geographic data and student learning behavior data, convert geographic data into geographic coordinate system, extract climate pattern, urbanization speed, and population density characteristics from geographic data, extract error frequency, learning time, and feedback emotion characteristics from student learning behavior data, and store the extracted characteristics in cloud database.

3. The method for enhancing geographical learning based on artificial intelligence according to claim 2, characterized in that: The step of feedback sentiment features includes: converting the feedback document submitted by the student into a text format, removing punctuation marks, special characters and stop words, using a sentiment analysis model to perform sentiment scoring on the feedback text, and for each feedback text, calculating its sentiment score as ,in, represents the sentiment score of the text, is the i-th word in the text, is the score of the word in the sentiment dictionary. According to the sentiment score, the sentiment tendency of the student feedback is analyzed. , classify the feedback as positive, when the sentiment score , classify the feedback as negative, when the sentiment score , classify the feedback as neutral.

4. The method for enhancing geographical learning based on artificial intelligence according to claim 1, characterized in that: In step 102, a generative adversarial network model is constructed, and the preprocessed geographic data and student data are used as training sets for deep learning training. The generative adversarial network model integrates multimodal learning and geographic knowledge graphs, and uses a reinforcement learning mechanism to continuously optimize the generator and the discriminator. The specific steps are as follows: Step B1, multimodal data fusion: extract geographic entities from geographic data and mark them, use the extracted geographic entities, combine urbanization process and climate pattern to build a geographic knowledge graph, where the nodes of the knowledge graph are geographic entities and attributes, and the edges represent the relationship between entities. Use graph embedding method to embed the nodes and edges in the knowledge graph into a continuous vector space to generate a vector representation of the graph, and fuse the feature vectors extracted from geographic data and student learning behavior with the vector representation embedded in the geographic knowledge graph to form a multimodal data set; Step B2, construct a generative adversarial network model: design a generator and a discriminator, input the multimodal data set into the generator to obtain the student's learning behavior prediction data, input the learning behavior prediction data generated by the generator into the discriminator, and output a probability value; Step B3, reinforcement learning optimization: Use real data and generated data for training, update the parameters of the generator and discriminator, and continuously optimize the generator and discriminator through the reward mechanism.

5. The method for enhancing geographical learning based on artificial intelligence according to claim 4 is characterized in that: In the step B2, in constructing a generative adversarial network model, a generator and a discriminator are designed, and the learning behavior prediction data of the students is obtained by inputting the multimodal data set into the generator, and the learning behavior prediction data generated by the generator is input into the discriminator to output a probability value, which further includes the following steps: Step B201, the network structure of the generator includes an input layer, a hidden layer and an output layer. The input layer is the multimodal data set X after input fusion, the hidden layer extracts feature information from the input, and the output layer generates the student's learning behavior prediction data. The hidden layer size is set to and , the formula from the input layer to the first hidden layer is: , the formula from the first hidden layer to the second hidden layer is: , the formula from the second hidden layer to the output layer is: ,in, is the weight matrix from the input layer to the first hidden layer, is the weight matrix from the first hidden layer to the second hidden layer, is the weight matrix from the second hidden layer to the output layer, f is the ReLU activation function, is the generated student learning behavior prediction data, is the input layer to the first hidden layer, is the bias from the first hidden layer to the second hidden layer, is the bias term of the output layer; Step B202: The discriminator receives the learning behavior prediction data generated by the generator through the input layer , and output a probability value through the output layer. The specific formula is: ,in, is the sigmoid activation function, which is used to map the output to range, P is the true probability value output by the discriminator, are the weight matrices of each layer of the discriminator, are the bias items of each layer of the discriminator.

6. The method for enhancing geographical learning based on artificial intelligence according to claim 4, characterized in that: In the step B3 reinforcement learning optimization, real data and generated data are used for training, the parameters of the generator and the discriminator are updated, and the generator and the discriminator are continuously optimized through the reward mechanism, which further includes the following steps: Step B301: Combine the real data and the generated data to form a training set, and define the reward function of the generator based on the probability value output by the discriminator. , to evaluate the performance of the generator, where is the generated data, R is the reward function, is the discriminant probability of the generated data d; Step B302: Calculate the reward R based on the discrimination results of the real data and the generated data, and use the reward information to update the parameters of the generator: ,in, is the learning rate, are the parameters updated by the generator, are the parameters of the generator, and R is the reward function; Step B303, repeat the above steps, continuously compare the data generated by the generator with the real data, optimize the generator and discriminator, and improve the model's ability to predict learning behavior.

7. The method for enhancing geographical learning based on artificial intelligence according to claim 1, characterized in that: In step 103, based on the student's learning progress, interests and needs, a trained generative adversarial network model is used to generate personalized geography learning content, which includes scenario simulation, virtual experiments and customized case analysis. The specific steps are as follows: Step C1: Express students’ learning progress, interests and needs as , extract relevant geographic features from the geographic knowledge graph and generate knowledge vectors: , and combine student characteristics S and geographical knowledge characteristics Fusion to form the input vector , input the fused input vector Z into the trained generator to generate personalized learning content , where C is the generated learning content, is the feature fusion function, is the vector representation of the learning data, is the vector representation of the geographic entity, G represents the generator function, are the parameters of the generator; Step C2: Generate geographical learning content of scenario simulation, virtual experiment and customized case analysis type according to the information in the input vector Z. The scenario simulation content related to the geographical environment is expressed as: ; By simulating ecological changes under different climatic conditions, experimental scenarios related to geographical concepts are generated, expressed as: ; Generate case studies for specific geographic issues based on student interests and needs, expressed as: ; and integrate the generated content into a comprehensive learning module , output personalized geography learning content modules to provide students with learning and exploration.

8. The method for enhancing geographical learning based on artificial intelligence according to claim 1, characterized in that: In step 104, the interactive data and feedback of students are continuously collected during the learning process, and the learning path is adjusted according to the students' behaviors, answer results and feedback information to form a dynamic feedback mechanism and realize a personalized learning plan. The specific steps are as follows: Step D1: Use the online platform to record the interactive data generated by students during the learning process, including learning time, number of visits, answer results, students' evaluation, opinions and suggestions on the learning content, organize and analyze the recorded data, extract the learning efficiency and interest index, and calculate the students' learning satisfaction based on the collected data and analysis results. , where J is the number of correct answers given by student J in the test, T is the total learning time, is the time the student spends on the jth visit, is the number of times students study. is the maximum number of visits. and is the weight coefficient; Step D2: Based on the calculated learning satisfaction, evaluate the student's learning status and form feedback information. According to the student's satisfaction and feedback information, implement the learning path adjustment strategy. Combined with the student's behavior pattern, interest points and feedback information, generate personalized learning recommendation content and push it to the student, which is expressed as ,in, Represents personalized learning recommendation content, is the generator model, is the input feature vector, which contains students’ learning behavior and feedback information.

Images