Intelligent voice interactive English dialogue teaching system

Abstract

The invention relates to the field of intelligent education, and discloses an intelligent voice interactive English dialogue teaching system, which is characterized in that text keywords extracted for the second time are primarily matched with an intention keyword list recorded in a corpus database, and then secondary matching is performed by combining the text keywords extracted for the first time; according to the method, a user intention keyword is screened out, a dialogue learning task of a corresponding category is automatically identified according to the user intention keyword, the dialogue learning task is allocated to the dialogue learning task of the corresponding category, and execution state data of the learning task and user learning behavior data are collected to analyze the execution state condition of the dialogue learning task. The learning difficulty in the execution process of the dialogue learning task is evaluated according to the learning result, then the dialogue difficulty degree is automatically adjusted, feedback is conducted to the user side, it is ensured that learning content closely fits the actual demand of the user, meanwhile, the learning efficiency of the user and the quality of learning results are improved, and the user experience is improved. Therefore, the interactive experience and the satisfaction degree of the user are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent education technology, and more specifically to an intelligent voice interactive English dialogue teaching system. Background Art

[0002] Traditional English teaching methods often focus on learning grammar and vocabulary, but lack the ability to cultivate students' oral expression. With the continuous advancement of artificial intelligence technology, speech recognition, natural language processing and other technologies have gradually matured, providing a technical foundation for the development of voice interactive English dialogue teaching systems, which can accurately recognize user voice input, understand user intentions, and generate corresponding answers, thereby realizing real-time interaction between humans and machines;

[0003] However, the above process still has the following disadvantages:

[0004] First, the existing voice interactive English dialogue teaching system only determines the system response based on the user's current answer and context information, and calls the corresponding service or API to generate the answer. It cannot accurately capture the user's learning intention, resulting in incorrect answer generation and insensitive dialogue response.

[0005] Second, the existing voice interactive English dialogue teaching system lacks the ability to automatically adjust the difficulty of learning content by real-time monitoring of the task execution process and analyzing the difficulty level. It is unable to start learning tasks according to user intentions and intelligently guide users according to user feedback and progress during task execution. It is therefore unable to further improve users' learning efficiency and the quality of their learning outcomes, which leads to a decline in users' interactive experience and satisfaction. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent voice interactive English dialogue teaching system to solve the problems existing in the above-mentioned background technology.

[0007] The present invention provides the following technical solution: an intelligent voice interactive English dialogue teaching system, comprising:

[0008] Data input module: used to identify the voice and text keywords input by the user, convert the input voice into text, extract keywords from the first and second input texts respectively, and transmit the first and second extracted text keywords to the data screening module;

[0009] Data screening module: The user's intended keywords are screened out by performing a preliminary match between the text keywords extracted for the second time and the list of intended keywords recorded in the corpus database, and then performing a secondary match with the text keywords extracted for the first time, calculating the keyword matching degree;

[0010] User intention analysis module: Through the matching analysis of the two selected keywords and calculating the similarity, it is determined whether the user intention keywords are accurate as the description of the user intention, and the user intention analysis results are transmitted to the task learning module;

[0011] Task learning module: automatically identifies the corresponding category of dialogue learning tasks according to the user's intended keywords, and assigns them to the corresponding category of dialogue learning tasks, wherein the dialogue learning tasks include word dialogue learning units, phrase dialogue learning units, sentence dialogue learning units, and paragraph dialogue learning units, and monitors and collects dialogue learning task execution status data, and transmits the dialogue learning task execution status data to the task learning analysis module;

[0012] User data collection module: used to collect user learning behavior data in real time during the process of performing dialogue learning tasks, and transmit the collected user learning behavior data to the task learning analysis module;

[0013] Task learning analysis module: Analyzes the execution status of the dialogue learning task through the execution status data of the learning task and the user's learning behavior data, obtains the dialogue fluency change index, action activity change index and learning task execution efficiency change index, and transmits the execution status analysis results of the dialogue learning task to the learning difficulty assessment module;

[0014] Learning difficulty evaluation module: Analyzes the learning difficulty during the execution of the dialogue learning task based on the execution status analysis result of the dialogue learning task, obtains the learning difficulty evaluation coefficient, evaluates the difficulty of the dialogue learning task through the learning difficulty evaluation coefficient, and transmits the evaluation result to the guidance switching module;

[0015] Guidance switching module: The guidance recommendation coefficient is calculated by comprehensively calculating the learning difficulty assessment coefficient and the keyword matching degree, and the dialogue difficulty is automatically adjusted according to the guidance recommendation coefficient, and the dialogue difficulty adjustment result is transmitted to the information feedback module;

[0016] Information feedback module: Feedback the adjustment results of the dialogue learning task to the user end, and provide feedback to the user in the form of voice prompts and interface displays.

[0017] Preferably, the data input module automatically recognizes the user's recorded voice through a microphone, converts the recorded voice into text, decomposes the user's text input into words and phrases, and extracts key words and phrases therefrom.

[0018] Preferably, the data screening module comprehensively analyzes the keyword matching degree C by calculating the word frequency of each keyword in the user text and the inverse document frequency in the intended keyword list, and compares the keyword matching degree C with a preset matching degree threshold θ, thereby screening out the user intended keywords;

[0019] The specific analysis method of the keyword matching degree is as follows:

[0020] Step S211: Calculate the frequency of each keyword extracted for the second time in the user text N(t, d2) represents the number of times the keyword appears in the second input user text t, N 总 represents the total number of words in the second user text t;

[0021] Step S212: Calculate the inverse document frequency in the intended keyword list M 总 represents the total number of keywords in the intent keyword list, and M(t) represents the number of intent keyword lists containing word t;

[0022] Step S213: Calculate the initial matching degree of each keyword FC=TF(t, d2)×IDF(t) by combining TF(t, d2) and IDF(t);

[0023] Step S214: by combining the initial matching result with the first input keyword, a secondary matching is performed to calculate the keyword matching degree C=FC+TF(t, d1)×IDF(t);

[0024] When the keyword matching degree C>the preset matching degree threshold θ, the keyword extracted at this time is used as the user intention keyword.

[0025] Preferably, the user intention analysis module filters out the similarity between keywords by analyzing the two texts. If it is determined that the relationship between the keywords of the two texts is similar, the user intention keywords are accurate as a description of the user intention, and the filtered user intention keywords are transmitted to the task learning module. If it is determined that the relationship between the keywords of the two texts is not similar, the user is prompted again to re-input voice.

[0026] Preferably, the task learning module associates user intention keywords with corresponding dialogue learning task categories by creating a mapping table, and uses the mapping table to assign dialogue learning tasks according to user intention keywords, and then monitors the execution process of the assigned dialogue learning tasks, and collects dialogue learning task execution status data, including the average conversation time, question frequency, number of errors and learning task progress.

[0027] Preferably, the user data collection module uses a camera to capture in real time the user's body movements and facial expressions during the execution of the dialogue learning task, thereby collecting the user's user learning behavior data, including the duration of the expression, the frequency of the movement, the speed of the movement, and the duration of the movement.

[0028] Preferably, the task learning analysis module performs a comprehensive analysis on the dialogue learning task execution status data and the user learning behavior data during the user's execution of the dialogue learning task, and monitors the execution status changes of the dialogue learning task;

[0029] The conversation fluency change index is analyzed by recording the number of pauses, pause time, and speech speed changes of the user in the conversation, and the conversation fluency change index is calculated as U t represents the total time of the conversation, U p Indicates the pause time in the conversation, P r represents the pause rate;

[0030] The action activity change index is calculated by analyzing the collected user's body and facial motion data and calculating the action activity change index as follows: E1 represents the time the expression is maintained, E 总 represents the total time of the conversation, F1 represents the actual action frequency, and F max Indicates the maximum action frequency, V1 indicates the actual action speed, V max represents the maximum action speed, D1 represents the action duration, β1, β2, β3, β4 are weight coefficients;

[0031] The learning task execution efficiency change index is analyzed by recording the number of errors in the time when the user completes the task, and the learning task execution efficiency change index is calculated as T user represents the average time it takes for a user to complete a task, R errors represents the number of errors made by the user in the dialogue learning task, R total Represents the total number of attempts in the dialogue learning task.

[0032] Preferably, the learning difficulty assessment module analyzes the comprehensive dialogue fluency change index X, the action activity change index A and the learning task execution efficiency change index Z, and calculates the learning difficulty assessment coefficient using the following calculation formula: a1, a2, a3 are weight coefficients, and e represents a constant.

[0033] Preferably, the guidance switching module analyzes the user's learning intention and learning difficulty in the dialogue learning task, thereby dynamically adjusting the difficulty of the dialogue learning task in real time;

[0034] The specific calculation formula of the guidance recommendation coefficient is: C represents the keyword matching degree, G represents the learning difficulty evaluation coefficient, θ represents the preset matching degree threshold, and κ represents the preset learning difficulty threshold;

[0035] By setting the upper limit guidance recommendation threshold η1 and the lower limit guidance recommendation threshold η2 respectively, and comparing the guidance recommendation coefficient W with the upper limit guidance recommendation threshold η1 and the lower limit guidance recommendation threshold η2 respectively, the difficulty of the dialogue learning task can be adjusted in real time; if the lower limit guidance recommendation threshold η2<guidance recommendation coefficient W<upper limit guidance recommendation threshold η1, there is no need to adjust the difficulty of the dialogue learning task; if the guidance recommendation coefficient W≤lower limit guidance recommendation threshold η2 or the guidance recommendation coefficient W≥upper limit guidance recommendation threshold η1, the difficulty of the dialogue learning task needs to be adjusted according to the current category of dialogue learning tasks.

[0036] Preferably, the information feedback module automatically generates corresponding voice prompt information according to the result of the dialogue difficulty adjustment, plays the voice prompt through the speaker to inform the user of the current dialogue difficulty level, and displays the corresponding information through the interface.

[0037] Technical effects and advantages of the present invention:

[0038] The present invention extracts keywords from the first and second text inputs respectively, and then performs an initial match between the text keywords extracted for the second time and the list of intent keywords recorded in the corpus database, and then performs a secondary match in combination with the text keywords extracted for the first time, thereby screening out user intent keywords, and judging whether the user intent keywords are accurate as descriptions of user intent through the matching degree of the keywords screened out twice, and then automatically identifying corresponding categories of dialogue learning tasks according to the user intent keywords, and assigning them to corresponding categories of dialogue learning tasks, and monitoring and collecting dialogue learning task execution status data, by real-time collecting user learning behavior data of users in the process of executing dialogue learning tasks, analyzing the execution status of dialogue learning tasks with the execution status data of learning tasks and user learning behavior data, based on dialogue learning The execution status analysis results of the task evaluate the learning difficulty during the execution of the dialogue learning task, and then comprehensively calculate the guidance recommendation coefficient through the learning difficulty assessment coefficient and the keyword matching degree, automatically adjust the difficulty of the dialogue, and provide feedback to the user end, which is conducive to real-time monitoring of the difficulty of learning task execution. Through multiple screening of user intention keywords, it can accurately capture the user's learning intention and customize personalized learning tasks accordingly to ensure that the learning content is closely aligned with the user's actual needs, thereby improving the pertinence and effectiveness of learning. By intelligently guiding users according to their feedback and progress during task execution, it is conducive to automatically switching to more appropriate learning tasks, ensuring that the learning process is coherent and efficient, while improving the user's learning efficiency and the quality of learning outcomes, thereby enhancing the user's interactive experience and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 The present invention is a flow chart of an intelligent voice interactive English dialogue teaching system. DETAILED DESCRIPTION

[0040] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. In addition, the forms of the various structures recorded in the following embodiments are merely illustrative. The intelligent voice interactive English dialogue teaching system involved in the present invention is not limited to the various structures recorded in the following embodiments. All other implementations obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0041] like Figure 1 This embodiment provides an intelligent voice interactive English dialogue teaching system, including:

[0042] Data input module: used to identify the voice and text keywords entered by the user, by converting the entered voice into text, extracting keywords from the first and second entered texts respectively, and transmitting the first and second extracted text keywords to the data screening module.

[0043] In this embodiment, the data input module automatically recognizes the user's recorded voice through a microphone, converts the recorded voice into text, decomposes the user's text input into words and phrases, and extracts key words and phrases therefrom.

[0044] It should be specifically noted that in the extraction of keywords from the user's text, the part of speech of each word after decomposition needs to be determined and marked, including nouns, verbs, and adjectives, and common meaningless words such as "and", "the", and "in" need to be removed.

[0045] Data screening module: The user’s intended keywords are screened out by initially matching the text keywords extracted for the second time with the list of intended keywords recorded in the corpus database, and then performing a secondary match with the text keywords extracted for the first time, calculating the keyword matching degree.

[0046] In this embodiment, the data screening module calculates the word frequency of each keyword in the user text and the inverse document frequency in the intended keyword list to comprehensively analyze the keyword matching degree C, and compares the keyword matching degree C with the preset matching degree threshold θ to screen out the user intended keywords;

[0047] The specific analysis method of keyword matching is as follows:

[0048] Step S211: Calculate the frequency of each keyword extracted for the second time in the user text N(t, d2) represents the number of times the keyword appears in the second input user text t, N 总 represents the total number of words in the second user text t;

[0049] Step S212: Calculate the inverse document frequency in the intended keyword list M 总 represents the total number of keywords in the intent keyword list, and M(t) represents the number of intent keyword lists containing word t;

[0050] Step S213: Calculate the initial matching degree of each keyword FC=TF(t, d2)×IDF(t) by combining TF(t, d2) and IDF(t);

[0051] Step S214: by combining the initial matching result with the first input keyword, a secondary matching is performed to calculate the keyword matching degree C=FC+TF(t, d1)×IDF(t);

[0052] When the keyword matching degree C>the preset matching degree threshold θ, the keyword extracted at this time is used as the user intention keyword.

[0053] It should be specifically noted that in the data screening module, the goal is to improve the intention screening process of tourism English conversation ability:

[0054] First entry: The user says: "I want to visit Paris next summer. I've always dreamed of seeing the Eiffel Tower." The system extracts keywords including "visit", "Paris", "summer", "Eiffel Tower", and "dreamed";

[0055] Second entry: The system responds: "That sounds exciting! What do you want to do in Paris?" The user answers: "I want to take a river cruise on the Seine and visit the Louvre Museum." The keywords extracted by the system include "river cruise", "Seine", "visit", and "Louvre Museum";

[0056] Initial match: In the intent keyword list, the keywords related to "travel plan" are "visit", "travel", "museum", "river", "cruise", and "tower". The results of the initial match between the keywords entered for the second time and the intent keyword list are: "river cruise matches", "Seine does not match", "visit matches", and "Louvre Museum matches";

[0057] Secondary matching: The keywords entered for the first time are combined with the initial matching results for secondary matching. Then, "visit" appears both for the first and second time, and the matching degree increases; "Paris" appears for the first time, which may increase the matching degree; "summer" appears for the first time, but is not directly related to the travel plan intent keywords; "Eiffel Tower" appears for the first time, and is matched; "dreamed" appears for the first time, but is not directly related to the travel plan intent keywords; "river cruise" appears for the second time, and is matched; "Seine" appears for the second time, and is matched; "Louvre Museum" appears for the second time, and is matched;

[0058] After calculating the matching degree of each keyword, the most relevant keywords are selected according to the matching degree, including “visit”, “river cruise”, “Seine”, and “Louvre Museum”.

[0059] User intention analysis module: Through the two selected keyword matching analysis and calculation of similarity, it is determined whether the user intention keyword is accurate as a description of the user intention, and the user intention analysis results are transmitted to the task learning module.

[0060] In this embodiment, the user intention analysis module filters out the similarity between keywords by analyzing the two texts. If it is determined that the relationship between the keywords of the two texts is similar, the user intention keywords are accurate in describing the user intention, and the filtered user intention keywords are transmitted to the task learning module. If it is determined that the relationship between the keywords of the two texts is not similar, the user is prompted again to re-input voice.

[0061] It should be noted that the specific calculation formula for the similarity of user intention keywords in the user intention analysis module is: FC(t1, d) represents the initial matching degree of keyword t1 in text d, FC(t2, d) represents the initial matching degree of keyword t2 in text d, and D represents the set of all texts containing t1 and t2;

[0062] By comparing the similarity Sim(t1, t2) with the similarity recognition threshold γ, if the similarity Sim(t1, t2)>intention recognition threshold γ, it is recognized that the user intention keyword is accurate in describing the user intention; if the user intention I≤intention recognition threshold γ, it is recognized that the user intention keyword is inaccurate in describing the user intention.

[0063] Task learning module: automatically identifies the corresponding category of dialogue learning tasks based on user intention keywords and assigns them to the corresponding category of dialogue learning tasks. The dialogue learning tasks include word dialogue learning units, phrase dialogue learning units, sentence dialogue learning units and paragraph dialogue learning units. It also monitors and collects dialogue learning task execution status data, and transmits the dialogue learning task execution status data to the task learning analysis module.

[0064] In this embodiment, the task learning module creates a mapping table to associate user intention keywords with corresponding dialogue learning task categories, uses the mapping table to assign dialogue learning tasks based on user intention keywords, monitors the execution process of the assigned dialogue learning tasks, and collects dialogue learning task execution status data, including the average conversation time, question frequency, number of errors, and learning task progress.

[0065] User data collection module: used to collect user learning behavior data in real time during the process of performing dialogue learning tasks, and transmit the collected user learning behavior data to the task learning analysis module.

[0066] In this embodiment, the user data collection module uses a camera to capture the user's body movements and facial expressions in real time during the performance of the dialogue learning task, thereby collecting the user's learning behavior data, including the duration of the expression, the frequency of the movement, the speed of the movement, and the duration of the movement.

[0067] Task learning analysis module: Analyzes the execution status of the dialogue learning task through the execution status data of the learning task and the user learning behavior data, obtains the dialogue fluency change index, action activity change index and learning task execution efficiency change index, and transmits the execution status analysis results of the dialogue learning task to the learning difficulty assessment module.

[0068] In this embodiment, the task learning analysis module performs a comprehensive analysis of the dialogue learning task execution status data and the user learning behavior data during the user's execution of the dialogue learning task, and monitors the execution status changes of the dialogue learning task;

[0069] The conversation fluency change index is analyzed by recording the number of pauses, pause time, and speech speed changes of users in the conversation, and the conversation fluency change index is calculated as U t represents the total time of the conversation, U p Indicates the pause time in the conversation, P r represents the pause rate;

[0070] The action activity change index is calculated by analyzing the collected user's body and facial movement data. E1 represents the time the expression is maintained, E 总 represents the total time of the conversation, F1 represents the actual action frequency, and F max Indicates the maximum action frequency, V1 indicates the actual action speed, V max represents the maximum action speed, D1 represents the action duration, β1, β2, β3, β4 are weight coefficients;

[0071] The learning task execution efficiency change index is analyzed by recording the number of errors in the time when users complete the task, and the learning task execution efficiency change index is calculated as T user represents the average time it takes for a user to complete a task, R errors represents the number of errors made by the user in the dialogue learning task, R total Represents the total number of attempts in the dialogue learning task.

[0072] Learning difficulty assessment module: Based on the execution status analysis results of the dialogue learning task, the learning difficulty during the execution of the dialogue learning task is analyzed to obtain the learning difficulty assessment coefficient. The difficulty of the dialogue learning task is evaluated by the learning difficulty assessment coefficient, and the evaluation results are transmitted to the guidance switching module.

[0073] In this embodiment, the learning difficulty assessment module analyzes the comprehensive dialogue fluency change index X, the action activity change index A, and the learning task execution efficiency change index Z, and calculates the learning difficulty assessment coefficient using the following formula: a1, a2, a3 are weight coefficients, and e represents a constant.

[0074] Guidance switching module: The guidance recommendation coefficient is calculated by comprehensively calculating the learning difficulty assessment coefficient and the keyword matching degree, the difficulty of the conversation is automatically adjusted according to the guidance recommendation coefficient, and the dialogue difficulty adjustment result is transmitted to the information feedback module.

[0075] In this embodiment, the guidance switching module analyzes the user's learning intention and learning difficulty in the dialogue learning task, thereby dynamically adjusting the difficulty of the dialogue learning task in real time;

[0076] The specific calculation formula of the guidance recommendation coefficient is: C represents the keyword matching degree, G represents the learning difficulty evaluation coefficient, θ represents the preset matching degree threshold, and κ represents the preset learning difficulty threshold;

[0077] By setting the upper limit guidance recommendation threshold η1 and the lower limit guidance recommendation threshold η2 respectively, and comparing the guidance recommendation coefficient W with the upper limit guidance recommendation threshold η1 and the lower limit guidance recommendation threshold η2 respectively, the difficulty of the dialogue learning task can be adjusted in real time; if the lower limit guidance recommendation threshold η2<guidance recommendation coefficient W<upper limit guidance recommendation threshold η1, there is no need to adjust the difficulty of the dialogue learning task; if the guidance recommendation coefficient W≤lower limit guidance recommendation threshold η2 or the guidance recommendation coefficient W≥upper limit guidance recommendation threshold η1, the difficulty of the dialogue learning task needs to be adjusted according to the current category of dialogue learning tasks.

[0078] It should be specifically explained that the process of adjusting the difficulty of the current category of dialogue learning tasks is as follows: when the guided recommendation coefficient W≤the lower limit guided recommendation threshold η2, the difficulty of the dialogue learning task needs to be reduced, including reducing uncommon words or professional terms in the dialogue, providing more prompts and help information, and reducing the complexity of the questions and reducing the number of questions; when the guided recommendation coefficient W≥the upper limit guided recommendation threshold η1, the difficulty of the dialogue learning task needs to be increased, including increasing uncommon words and professional terms in the dialogue, reducing information prompts during the dialogue, and appropriately increasing the number and difficulty of questions.

[0079] Information feedback module: Feedback the adjustment results of the dialogue learning task to the user end, and provide feedback to the user in the form of voice prompts and interface displays.

[0080] In this embodiment, the information feedback module automatically generates corresponding voice prompt information according to the result of the dialogue difficulty adjustment, plays the voice prompt through the speaker to inform the user of the current dialogue difficulty level, and displays the corresponding information through the interface.

[0081] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0082] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An intelligent voice interactive English dialogue teaching system, characterized by: include: Data input module: used to identify the voice and text keywords input by the user, convert the input voice into text, extract keywords from the first and second input texts respectively, and transmit the first and second extracted text keywords to the data screening module; Data screening module: The user's intended keywords are screened out by performing a preliminary match between the text keywords extracted for the second time and the list of intended keywords recorded in the corpus database, and then performing a secondary match with the text keywords extracted for the first time, calculating the keyword matching degree; User intention analysis module: Through the matching analysis of the two selected keywords and calculating the similarity, it is determined whether the user intention keywords are accurate as the description of the user intention, and the user intention analysis results are transmitted to the task learning module; Task learning module: automatically identifies the corresponding category of dialogue learning tasks according to the user's intended keywords, and assigns them to the corresponding category of dialogue learning tasks, wherein the dialogue learning tasks include word dialogue learning units, phrase dialogue learning units, sentence dialogue learning units, and paragraph dialogue learning units, and monitors and collects dialogue learning task execution status data, and transmits the dialogue learning task execution status data to the task learning analysis module; User data collection module: used to collect user learning behavior data in real time during the process of performing dialogue learning tasks, and transmit the collected user learning behavior data to the task learning analysis module; Task learning analysis module: Analyzes the execution status of the dialogue learning task through the execution status data of the learning task and the user's learning behavior data, obtains the dialogue fluency change index, action activity change index and learning task execution efficiency change index, and transmits the execution status analysis results of the dialogue learning task to the learning difficulty assessment module; Learning difficulty evaluation module: Analyzes the learning difficulty during the execution of the dialogue learning task based on the execution status analysis result of the dialogue learning task, obtains the learning difficulty evaluation coefficient, evaluates the difficulty of the dialogue learning task through the learning difficulty evaluation coefficient, and transmits the evaluation result to the guidance switching module; Guidance switching module: The guidance recommendation coefficient is calculated by comprehensively calculating the learning difficulty assessment coefficient and the keyword matching degree, and the dialogue difficulty is automatically adjusted according to the guidance recommendation coefficient, and the dialogue difficulty adjustment result is transmitted to the information feedback module; Information feedback module: Feedback the adjustment results of the dialogue learning task to the user end, and provide feedback to the user in the form of voice prompts and interface displays.

2. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The data input module automatically recognizes the user's input voice through a microphone, converts the input voice into text, decomposes the user's text input into words and phrases, and extracts key words and phrases therefrom.

3. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The data screening module calculates the word frequency of each keyword in the user text and the inverse document frequency in the intended keyword list to comprehensively analyze the keyword matching degree C, and compares the keyword matching degree C with the preset matching degree threshold θ to screen out the user intended keywords; The specific analysis method of the keyword matching degree is as follows: Step S211: Calculate the frequency of each keyword extracted for the second time in the user text N(t, d2) represents the number of times the keyword appears in the second input user text t, N 总 represents the total number of words in the second user text t; Step S212: Calculate the inverse document frequency in the intended keyword list M 总 represents the total number of keywords in the intent keyword list, and M(t) represents the number of intent keyword lists containing word t; Step S213: Calculate the initial matching degree of each keyword FC=TF(t, d2)×IDF(t) by combining TF(t, d2) and IDF(t); Step S214: by combining the initial matching result with the first input keyword, a secondary matching is performed to calculate the keyword matching degree C=FC+TF(t, d1)×IDF(t); When the keyword matching degree C>the preset matching degree threshold θ, the keyword extracted at this time is used as the user intention keyword.

4. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The user intention analysis module filters out the similarity between keywords by analyzing the two texts. If it is determined that the relationship between the keywords of the two texts is similar, the user intention keywords are accurate in describing the user intention, and the filtered user intention keywords are transmitted to the task learning module. If it is determined that the relationship between the keywords of the two texts is not similar, the user is prompted again to re-input voice.

5. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The task learning module creates a mapping table to associate user intent keywords with corresponding dialogue learning task categories, uses the mapping table to assign dialogue learning tasks based on user intent keywords, monitors the execution process of the assigned dialogue learning tasks, and collects dialogue learning task execution status data, including the average conversation time, question frequency, number of errors, and learning task progress.

6. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The user data collection module uses a camera to capture the user's body movements and facial expressions in real time during the dialogue learning task, thereby collecting the user's user learning behavior data, including the duration of expression maintenance, movement frequency, movement speed, and movement duration.

7. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The task learning analysis module performs a comprehensive analysis on the dialogue learning task execution status data and the user learning behavior data during the user's dialogue learning task execution, and monitors the execution status changes of the dialogue learning task; The conversation fluency change index is analyzed by recording the number of pauses, pause time, and speech speed changes of the user in the conversation, and the conversation fluency change index is calculated as U t represents the total time of the conversation, U p Indicates the pause time in the conversation, P r represents the pause rate; The action activity change index is calculated by analyzing the collected user's body and facial motion data and calculating the action activity change index as follows: E1 represents the time the expression is maintained, E 总 represents the total time of the conversation, F1 represents the actual action frequency, and F max Indicates the maximum action frequency, V1 indicates the actual action speed, V max represents the maximum action speed, D1 represents the action duration, β1, β2, β3, β4 are weight coefficients; The learning task execution efficiency change index is analyzed by recording the number of errors in the time when the user completes the task, and the learning task execution efficiency change index is calculated as T user represents the average time it takes for a user to complete a task, R errors represents the number of errors made by the user in the dialogue learning task, R total Represents the total number of attempts in the dialogue learning task.

8. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The learning difficulty evaluation module analyzes the comprehensive dialogue fluency change index X, action activity change index A and learning task execution efficiency change index Z, and calculates the learning difficulty evaluation coefficient using the following formula: a1, a2, a3 are weight coefficients, and e represents a constant.

9. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The guidance switching module analyzes the user's learning intention and learning difficulty in the dialogue learning task, thereby dynamically adjusting the difficulty of the dialogue learning task in real time; The specific calculation formula of the guidance recommendation coefficient is: C represents the keyword matching degree, G represents the learning difficulty evaluation coefficient, θ represents the preset matching degree threshold, and κ represents the preset learning difficulty threshold; By setting the upper limit guidance recommendation threshold η1 and the lower limit guidance recommendation threshold η2 respectively, and comparing the guidance recommendation coefficient W with the upper limit guidance recommendation threshold η1 and the lower limit guidance recommendation threshold η2 respectively, the difficulty of the dialogue learning task can be adjusted in real time; if the lower limit guidance recommendation threshold η2<guidance recommendation coefficient W<upper limit guidance recommendation threshold η1, there is no need to adjust the difficulty of the dialogue learning task; if the guidance recommendation coefficient W≤lower limit guidance recommendation threshold η2 or the guidance recommendation coefficient W≥upper limit guidance recommendation threshold η1, the difficulty of the dialogue learning task needs to be adjusted according to the current category of dialogue learning tasks.

10. The intelligent voice interactive English dialogue teaching system according to claim 1 is characterized by: The information feedback module automatically generates corresponding voice prompt information according to the result of the dialogue difficulty adjustment, plays the voice prompt through the speaker to inform the user of the current dialogue difficulty level, and displays the corresponding information through the interface.

Images