Language field simulation intelligent teaching system

Abstract

The invention belongs to the technical field of language teaching, and particularly discloses a language field simulation intelligent teaching system, which is characterized in that an optimization period is set, and then use records of various language situations in the teaching system are called in the optimization period to focus the interaction state of a virtual interaction object, so that the interaction interference degree of the language situations is evaluated; the method can achieve the timely and periodic continuous recognition of the interaction interference of the language situation in a teaching system, thereby providing a reliable guarantee for the optimization opportunity of the optimization of the language situation. When the virtual interaction object in the language situation provided by the teaching system is identified to have interaction interference, interaction interference trend analysis and interaction interference and student performance influence correlation analysis are carried out on the virtual interaction object; the language situation needing to be optimized is selected under the double consideration of the interaction interference state and the correlation of the interaction interference influence on the student language performance, and the selection result ensures the accuracy and pertinence of the optimization strategy.

Description

Technical Field

[0001] The invention belongs to the technical field of language teaching, and specifically discloses a language on-site simulation intelligent teaching system. Background Art

[0002] Modern society attaches more and more importance to language application ability, while traditional language classroom teaching that emphasizes the imparting of theoretical knowledge and grammatical rules can no longer meet the teaching needs of language application. In this situation, language on-site simulation teaching applications are born. It can simulate a variety of real contexts and use virtual interactive objects to interact with students for students to practice language skills, allowing learners to practice and respond in real life, thereby enhancing their language application ability.

[0003] Because of the teaching advantages provided by on-site language simulation teaching, on-site language simulation teaching is currently widely used in educational institutions to measure students' language application level. However, when measuring students' language application, the focus is mostly on students, and insufficient attention is paid to the interactive interference of virtual interactive objects in the language context provided by the simulation teaching system, which makes it difficult to timely identify language contexts that need to be optimized. This will affect students' actual language practice experience to a certain extent, and then the student performance scores obtained in such language contexts cannot objectively reflect the students' actual ability level, thus affecting the fairness and credibility of the evaluation.

[0004] Furthermore, when the interaction interference of virtual interactive objects in the language context is identified, the existing selection of the language context to be optimized in the simulation teaching system usually takes the state of the interaction interference itself as the selection basis, ignoring the correlation between the impact of the interaction interference on the students' language performance. On the one hand, it directly causes inaccurate selection results, because some interaction interferences may be mild in degree but have a greater negative impact on students' language performance. This impact may be implicit or complex and not easy to measure directly by the degree of interference. On the other hand, the optimization of the language context may lose its pertinence and effectiveness, because the optimized language context may still have problems that have a substantial impact on students' performance. These problems have not been effectively solved, and students' learning experience and results have not been effectively improved. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a language scene simulation intelligent teaching system, which effectively makes up for the deficiency of the simulation teaching system in optimizing the language context provided by placing the focus of language simulation teaching on virtual interactive objects.

[0006] The purpose of the present invention can be achieved through the following technical solutions: The present invention provides a language scene simulation intelligent teaching system, comprising the following modules: a language context usage record retrieval module, which is used to set an optimization cycle, and then retrieve the usage records corresponding to each language context in the teaching system within the optimization cycle, and extract the usage time, using students and dialogue voice from the usage records.

[0007] The interaction state recognition module is used to focus the recorded conversation voice on the virtual interaction object and extract the interaction state indicators of the virtual interaction object, including the interaction response time, interaction interruption frequency and interaction clarity.

[0008] The interaction interference evaluation module is used to evaluate the interaction interference degree of each language context in each usage record based on the interaction state indicator of the virtual interaction object.

[0009] The interaction interference trend analysis module is used to perform interaction interference trend analysis based on the interaction interference degree of each language context in each usage record.

[0010] The student performance score extraction module is used to retrieve the student performance score records related to the usage records corresponding to each language context, and extract the student performance score values ​​from the score records.

[0011] The interaction interference impact association module is used to classify the usage records corresponding to the same interaction interference degree in each language context, obtain the usage records classified with different interaction interference degrees in each language context, and combine them with the student performance score values ​​in the scoring records involved in the usage records to perform an interaction interference and student performance impact association analysis, and obtain the interaction interference and student performance impact correlation corresponding to each language context.

[0012] The language context optimization selection module is used to select the language context that needs to be optimized within the optimization cycle by combining the interaction interference trend corresponding to each language context and the correlation between the interaction interference corresponding to each language context and the impact on student performance.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The present invention sets an optimization cycle, and then retrieves the usage records of each language context in the teaching system within the optimization cycle to focus on the interaction status of the virtual interactive object, thereby evaluating the degree of interaction interference of the language context, and can achieve timely and periodic continuous identification of language context interaction interference in the teaching system, helping teaching system managers to promptly discover possible problems in the system, thereby providing reliable guarantee for the optimization timing of the language context optimization, which is conducive to minimizing the impact on students' language performance scores.

[0014] 2. When identifying the existence of interactive interference among virtual interactive objects in the language context provided by the teaching system, the present invention performs interactive interference trend analysis and correlation analysis on the impact of interactive interference on student performance. This enables the selection of language contexts that need to be optimized under dual considerations of the interactive interference state itself and the correlation between the interactive interference and the impact of the interactive interference on student language performance. The selection result ensures the accuracy and pertinence of the optimization strategy, and is closer to actual application needs, which helps the teaching system to continuously provide students with real and effective language contexts after optimization.

[0015] 3. After screening out the language contexts that need to be optimized, the present invention adds an evaluation prediction of the optimization direction. By performing a correlation analysis on the interactive interference of the language contexts that need to be optimized and the conversation duration, and then predicting whether there is an optimization direction based on the correlation analysis results, it can provide a reliable reference for the optimization implementation of the language context, avoid unnecessary optimization experiments and adjustments, ensure that each optimization is well-founded and has the expected effect, and more effectively utilize optimization resources, which is conducive to improving the efficiency of language context optimization implementation and the guarantee of the effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0017] Figure 1 It is a schematic diagram of the connection of each module of the system of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] See also Figure 1 As shown, the present invention proposes a language on-site simulation intelligent teaching system, including a language context usage record retrieval module, an interaction state recognition module and a student performance score extraction module connected to the language context usage record retrieval module, an interaction interference assessment module connected to the interaction state recognition module, an interaction interference trend analysis module and an interaction interference impact association module connected to the interaction interference assessment module, an interaction interference impact association module connected to the student performance score extraction module, and a language context optimization selection module connected to the interaction interference trend analysis module and the interaction interference impact association module.

[0020] The language context usage record retrieval module is used to set an optimization cycle, and then retrieve the usage records corresponding to each language context in the teaching system within the optimization cycle, and extract the usage time, the students using it, and the conversation voice from the usage records.

[0021] It should be understood that since the use of language contexts is continuously generated in real time, when retrieving the usage records of language contexts, it should be taken into account that the use of language contexts and interactive interference may change over time. Therefore, only the most recent records need to be retrieved to more accurately reflect the actual needs of current students. In order to avoid the difficulty of retrieving older usage records to meet current optimization needs, by setting an optimization cycle, such as retrieving usage records within 6 months of the current date, the most recent usage records can be retrieved in real time to ensure that optimization strategies and adjustments are based on the latest usage records. At the same time, retrieving the most recent usage records can simplify the data management and analysis process. This allows you to focus on the most recent and relevant data and avoid the additional complexity and resource consumption brought about by processing large amounts of outdated data.

[0022] In the example of the above scheme, the language contexts provided by the teaching system include but are not limited to debates, business negotiations, job interviews, etc.

[0023] The interaction state recognition module is used to focus the recorded conversation voice on the virtual interaction object, and extract the interaction state indicators of the virtual interaction object, including the interaction response duration, interaction interruption frequency and interaction clarity.

[0024] It needs to be explained that the reason for using interaction response time, interaction interruption frequency and interaction clarity as interaction status indicators is that these indicators can objectively quantify and measure the interaction quality of virtual interactive objects and identify whether there is interaction interference. These quantitative indicators make the evaluation of interaction interference more scientific and reliable.

[0025] Applied to the above embodiment, the process of refining the interaction response duration in the interaction status indicator is as follows: the conversation speech is divided into conversation groups according to the interaction, and the student's conversation time and the virtual interaction object's conversation time are extracted from each conversation group respectively, and the two conversation times are subtracted to obtain the interval duration of the conversation between the virtual interaction object and the student in each conversation group.

[0026] The interaction mentioned above is to identify whether there is interaction in the conversational speech, and divide the conversational speech with interaction into conversation groups. The conversation group includes two conversation subjects, the student and the virtual interaction object, which can be specifically understood as a question-and-answer format.

[0027] The interaction response time is obtained by calculating the average of the intervals between the virtual interaction objects and students in each dialogue group.

[0028] Further applied to the above embodiment, the interaction interruption frequency in the interaction status indicator is refined as follows: according to the dialogue subject, the dialogue voice whose dialogue subject is the virtual interaction object is screened out from the dialogue voice.

[0029] Perform interruption detection on the screened conversation voices and count the frequency of interaction interruptions.

[0030] As an example implementation of the above scheme, the interruption detection of conversational speech can adopt an energy threshold detection method, specifically by setting a sound energy threshold. When the sound energy is detected to exceed the set threshold, the speech is considered to have started. When the sound energy is lower than the threshold for a certain period of time, the system determines that the speech has ended. In this way, whether the speech is interrupted is judged based on the speech start time and the speech end time.

[0031] As another example implementation of the above scheme, the interruption detection of conversational speech can use an endpoint detection algorithm, such as the short-time energy zero-crossing rate, to detect the start time and end time of the speech signal, and then judge whether the language is interrupted based on the start time and end time of the speech signal.

[0032] It should be noted that, since the detection of voice interruption is relatively mature in the prior art, the present invention does not describe the specific operation of voice interruption detection in detail.

[0033] In a further innovation of the above scheme, when a conversation voice interruption is detected, the interruption frequency is recorded and divided by the length of the conversation voice to obtain the interaction interruption frequency, wherein the length of the conversation voice refers to the duration of the voice signal in time, which can be specifically understood as the duration.

[0034] Further applied to the above embodiment, the interaction clarity in the interaction status indicator is refined as follows: the dialogue voice with the screened dialogue subject as the virtual interaction object is subjected to waveform conversion and spectrum conversion using audio processing software to obtain a dialogue voice waveform graph and a dialogue voice spectrum graph, wherein the dialogue voice waveform graph reflects the amplitude change of the voice signal on the time axis, and the dialogue voice spectrum graph uses frequency as the horizontal axis and energy consumption or amplitude as the vertical axis, reflecting the energy distribution of the voice signal at different frequencies.

[0035] The peaks and troughs are marked from the dialogue speech waveform, and the dialogue speech waveform is divided into bands, each band corresponds to a peak and a trough, and then the transition slope between the peaks and troughs in each band is obtained.

[0036] The transition slope between the peak and the trough in each band is compared with the transition slope corresponding to the known high-quality speech sample to calculate the smoothness of the dialogue speech waveform. The high-quality speech sample usually has a smoother and continuous waveform transition. The specific calculation formula is: , where Indicates The transition slope between the peaks and troughs in the band, Indicates the band number, , It represents the transition slope corresponding to the known high-quality speech sample, where the closer the transition slope between the peak and the trough in each band is to the transition slope corresponding to the known high-quality speech sample, Represents a natural constant, the greater the smoothness of the conversational speech waveform.

[0037] It should be noted that speech with high clarity should have clear peaks and troughs on the waveform, and the transition between the peaks and troughs should be relatively continuous and natural. The transition slope between the peaks and troughs is obtained, which describes the rate of sound change between the peaks and troughs. If the transition slope is relatively gentle, it means that the speech changes are relatively natural and clear.

[0038] The flatness of the conversational speech spectrum is calculated from the conversational speech spectrogram. Specifically, the spectral flatness refers to the uniformity of the energy distribution of the speech signal in the entire spectrum, which reflects the ratio of high-frequency components to low-frequency components in the spectrum. If the energy of the speech signal in all frequency bands is relatively uniformly distributed, the spectral flatness is high; conversely, if there are some frequency bands with high energy and other frequency bands with low energy, the spectral flatness is low. The spectral flatness is usually expressed by calculating the ratio of the geometric mean to the arithmetic mean of the spectrum.

[0039] It should be noted that clear speech should have a clear energy peak in the core frequency range, the spectrum should be relatively flat, without excessive noise or distortion, and high spectrum flatness usually means that the energy distribution of the speech signal in all frequency bands is relatively uniform, which may help improve the naturalness and audibility of the speech. Clear speech usually has a relatively flat spectrum without obvious frequency distortion or uneven energy.

[0040] The interaction clarity is calculated by taking a weighted average of the smoothness of the dialogue speech waveform and the flatness of the dialogue speech spectrum. In a specific implementation example, the weight coefficients corresponding to the smoothness of the dialogue speech waveform and the flatness of the dialogue speech spectrum may be 0.6 and 0.4.

[0041] When analyzing the clarity of conversational speech, the present invention does not study the conversational speech itself, but provides an intuitive display of the speech signal structure by converting the conversational speech into a waveform graph and a frequency spectrum. It can not only observe and understand the characteristics of the speech signal more clearly, but also provide a rich data basis and visual support for speech processing, recognition and its application. Herein, analysis is performed based on two indicators: the smoothness of the speech waveform and the flatness of the speech spectrum. Each of them reflects the characteristics of the speech signal from different angles, and can achieve a more comprehensive and accurate evaluation and clarity of the conversational speech.

[0042] The interaction interference evaluation module is used to analyze the interaction interference degree of each language context in each usage record based on the interaction state indicator of the virtual interaction object. Specifically, the interaction state indicator of the virtual interaction object in each language context in each usage record is substituted into the analysis formula , we get the degree of interaction interference of each language context in each usage record, where , , They respectively represent the interaction response time, interaction interruption frequency, and interaction clarity of the virtual interaction object in the usage record. , , They respectively represent the ideal interaction response time, ideal interaction interruption frequency, and ideal interaction clarity of the teaching system scenario simulation. The ideal interaction response time, ideal interaction interruption frequency, and ideal interaction clarity need to be reasonably set when the teaching system is designed and constructed according to the specific usage scenario, target audience, and teaching objectives. The specific settings can be obtained from the system design documents.

[0043] It should be added that as the number of users and the frequency of use in the teaching system increase, the system load may increase. If the system infrastructure (such as servers, network bandwidth) or software design is not sufficient to support high loads, the interactive performance may decline, which can be manifested as longer interactive response time, higher interaction interruption frequency, and lower interaction clarity.

[0044] The interaction interference trend analysis module is used to perform interaction interference trend analysis based on the interaction interference degree of each language context in each usage record. The specific analysis is as follows: a two-dimensional coordinate system is constructed with the usage time as the horizontal axis and the interaction interference degree as the vertical axis. Several points are marked in the constructed two-dimensional coordinate system according to the interaction interference degree of each language context in each usage record and the usage time corresponding to the usage record, so as to form an interaction interference change curve corresponding to each language context.

[0045] The change rate is statistically calculated from the interaction interference change curve corresponding to each language context, where the change rate can be calculated by averaging the slopes of each point on the change curve, which is the interaction interference trend corresponding to each language context. This trend intuitively reflects the state of the interaction interference itself. Since the change rate can be positive or negative, when the change rate is positive, the interaction interference tends to increase, and when the change rate is negative, the interaction interference tends to decrease.

[0046] The student performance score extraction module is used to retrieve the student performance score records related to the usage records corresponding to each language context, and extract the student performance score values ​​from the score records.

[0047] The interaction interference impact association module is used to classify the usage records corresponding to the same interaction interference degree in each language context, obtain the usage records classified with different interaction interference degrees in each language context, and combine them with the student performance score values ​​in the score records involved in the usage records to perform an interaction interference and student performance impact association analysis, and obtain the interaction interference and student performance impact correlation corresponding to each language context.

[0048] Preferably, the correlation analysis of the impact of interaction interference and student performance is implemented as follows: the student performance score values ​​in the score records involved in each usage record classified into different levels of interaction interference in each language context are compared, the highest score value and the lowest score value are extracted, and then the highest score value and the lowest score value are subtracted, and the difference is used as the exponent with three quarters as the base to obtain the student performance stability corresponding to each level of interaction interference in each language context, wherein the smaller the difference between the highest score value and the lowest score value, the greater the student performance stability, indicating that the student performance is less affected by the interaction interference.

[0049] A two-dimensional coordinate system is constructed with the degree of interaction interference as the horizontal axis and the stability of student performance as the vertical axis. Several points are marked in the constructed two-dimensional coordinate system according to the stability of student performance corresponding to each degree of interaction interference in each language context to form a scatter plot. A regression line is drawn for the scatter plot, and the slope of the regression line is obtained.

[0050] The absolute value of the slope of the regression line is taken as the correlation between the interaction interference and student performance corresponding to each language situation. The larger the absolute value of the slope of the regression line, the greater the correlation between the interaction interference and student performance, indicating that student performance is more affected by the interaction interference.

[0051] The language context optimization selection module is used to select the language contexts that need to be optimized within the optimization cycle by combining the interactive interference trends corresponding to each language context and the correlation between the interactive interference and the student's performance. The specific selection is as follows: the interactive interference trends corresponding to each language context and the correlation between the interactive interference and the student's performance corresponding to each language context are statistically calculated to obtain the optimization requirement corresponding to each language context. The specific statistical formula is: , where the greater the interactive interference tendency of the language context, the greater the correlation of its impact on student performance, and the greater the need for optimization.

[0052] The optimization requirement corresponding to each language context is compared with the system initial setting threshold. For example, the system initial setting threshold is 0.7, and the language contexts with optimization requirements greater than the system initial setting threshold are selected as the language contexts that need to be optimized within the optimization cycle.

[0053] In the innovative implementation of the above scheme, after the language context optimization selection module selects the language context that needs to be optimized, it also judges whether the selected voice context has a certain optimization direction. The specific judgment is as follows: the selected voice context is recorded as the target language context, and then the conversation duration of the conversation content in each usage record corresponding to the target language context is counted.

[0054] It should be pointed out that the conversation duration mentioned above refers to the duration of the entire conversation voice.

[0055] The degree of interaction interference and conversation duration of the target language context in each usage record were correlated and analyzed to obtain the correlation coefficient between interaction interference and conversation duration corresponding to the target language context.

[0056] Specifically, the above-mentioned related analysis is as follows: a two-dimensional coordinate system is constructed with usage time as the horizontal axis and conversation duration as the vertical axis, and then a number of points are marked in the constructed two-dimensional coordinate system according to the conversation duration in each usage record in the target language context and the usage time corresponding to the usage record, so as to form a conversation duration change curve corresponding to the target language context.

[0057] The correlation coefficient between the interaction interference change curve corresponding to the target language context and the conversation duration change curve is calculated.

[0058] It should be added that the correlation coefficient usually refers to the Pearson correlation coefficient, which can measure the strength and direction of the linear relationship between two variables. Here it is used to calculate the correlation coefficient between two changing curves. If the correlation coefficient is close to 1, it means that there is a strong positive correlation between the two curves, that is, they change in the same direction. If the correlation coefficient is close to -1, it means that there is a strong negative correlation between the two curves, that is, they change in the opposite direction. If the correlation coefficient is close to 0, it means that there is no correlation between the two curves.

[0059] In calculating the correlation coefficient between the interaction interference change curve and the conversation duration change curve, the curves can be regarded as time series data of two variables, which have corresponding values ​​at the same time point. Therefore, the two curves are divided into two arrays according to the corresponding values ​​at the same time point, and then the covariance and standard deviation of the two arrays are calculated respectively. Finally, the covariance and standard deviation are used to calculate the Pearson correlation coefficient.

[0060] The absolute value of the correlation coefficient between the interaction interference and the conversation duration corresponding to the target language context is compared with the effective correlation coefficient of the coefficient configuration. Exemplarily, the effective correlation coefficient can be 0.8. If the absolute value of the correlation coefficient between the interaction interference and the conversation duration corresponding to the target language context is greater than or equal to the effective correlation coefficient, indicating that the interaction interference and the conversation duration in the target language context have an obvious correlation, then it is judged that the target language context has a definite optimization direction, namely, limiting the conversation duration. Otherwise, it is analyzed that the target language context does not have a definite optimization direction.

[0061] When analyzing the correlation between the interaction interference change curve and the conversation duration change curve, the present invention uses the correlation coefficient to analyze the correlation by using curve comparison, which can intuitively display the correlation between the interaction interference and the conversation duration. For example, if the correlation coefficient is positive and close to 1, it means that as the conversation duration increases, the interaction interference increases. On the contrary, if the correlation coefficient is negative and close to -1, it means that as the conversation duration increases, the interaction interference decreases. This intuitive data interpretation helps to quickly understand and optimize implementation.

[0062] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. A language on-site simulation intelligent teaching system, characterized in that: Includes the following modules: The language context usage record retrieval module is used to set the optimization cycle, and then retrieve the usage records corresponding to each language context in the teaching system within the optimization cycle, and extract the usage time, students using it, and conversation voice from the usage records; An interaction state recognition module is used to focus the recorded conversation voice on the virtual interaction object and extract the interaction state indicators of the virtual interaction object, including the interaction response time, interaction interruption frequency and interaction clarity; An interaction interference evaluation module, used to evaluate the interaction interference degree of each language context in each usage record based on the interaction state indicator of the virtual interaction object; An interaction interference trend analysis module is used to perform interaction interference trend analysis based on the interaction interference degree of each language context in each usage record; A student performance score extraction module is used to retrieve the student performance score records related to the usage records corresponding to each language context, and extract the student performance score values ​​from the score records; The interaction interference impact association module is used to classify the usage records corresponding to the same interaction interference degree in each language context, obtain the usage records classified with different interaction interference degrees in each language context, and combine them with the student performance score values ​​in the score records involved in the usage records to perform interaction interference and student performance impact association analysis, and obtain the interaction interference and student performance impact correlation degree corresponding to each language context; The language context optimization selection module is used to select the language context that needs to be optimized within the optimization cycle by combining the interaction interference trend corresponding to each language context and the correlation between the interaction interference corresponding to each language context and the impact on student performance.

2. A language on-site simulation intelligent teaching system as claimed in claim 1, characterized in that: The process of extracting the interaction response time in the interaction status indicator is as follows: The conversation speech is divided into conversation groups according to the interaction, and the conversation time of the student and the conversation time of the virtual interaction object are extracted from each conversation group, thereby obtaining the interval length of the conversation between the virtual interaction object and the student in each conversation group; The interaction response time is obtained by calculating the average of the intervals between the virtual interaction objects and students in each dialogue group.

3. A language on-site simulation intelligent teaching system as claimed in claim 2, characterized in that: The interaction interruption frequency in the interaction status indicator is refined as follows: According to the dialogue subject, the dialogue voices whose dialogue subject is a virtual interaction object are selected from the dialogue voices; Perform interruption detection on the screened conversation voices and count the frequency of interaction interruptions.

4. A language on-site simulation intelligent teaching system as claimed in claim 3, characterized in that: The interaction clarity in the interaction status indicator is summarized as follows: The dialogue voices selected as the dialogue subjects of the virtual interaction objects are converted into waveforms and spectra using audio processing software to obtain dialogue voice waveform graphs and dialogue voice spectrograms; Mark the peak and trough positions from the conversation speech waveform, thereby dividing the conversation speech waveform into bands, each band corresponds to a peak and a trough, and then obtain the transition slope between the peak and the trough in each band; The smoothness of the dialogue speech waveform is calculated by comparing the transition slope between the peak and the trough in each band with the transition slope corresponding to the known high-quality speech sample; Calculate the conversational speech spectrum flatness from the conversational speech spectrogram; The interaction clarity is calculated by taking the weighted average of the smoothness of the dialogue speech waveform and the flatness of the dialogue speech spectrum.

5. A language on-site simulation intelligent teaching system as claimed in claim 1, characterized in that: The degree of interaction interference of each language context in each usage record is analyzed as follows: Substitute the interaction status indicators of the virtual interaction objects in each language context in each usage record into the analytical formula , we get the degree of interaction interference of each language context in each usage record, where , , They respectively represent the interaction response time, interaction interruption frequency, and interaction clarity of the virtual interaction object in the usage record. , , They respectively represent the ideal interaction response time, ideal interaction interruption frequency, and ideal interaction clarity of the teaching system scenario simulation.

6. A language on-site simulation intelligent teaching system as claimed in claim 1, characterized in that: The mutual interference trend analysis is as follows: A two-dimensional coordinate system is constructed with the usage time as the horizontal axis and the interaction interference degree as the vertical axis. A number of points are marked in the constructed two-dimensional coordinate system according to the interaction interference degree of each language context in each usage record and the usage time corresponding to the usage record, so as to form an interaction interference change curve corresponding to each language context. The change rate is calculated from the interaction interference change curve corresponding to each language context as the interaction interference trend corresponding to each language context.

7. A language on-site simulation intelligent teaching system as claimed in claim 1, characterized in that: The correlation analysis of the impact of interactive interference on student performance is implemented as follows: Compare the student performance scores in the scoring records of the usage records classified by different levels of interaction interference in each language context, extract the highest and lowest scores, and then subtract the highest and lowest scores, and use three quarters as the base, and use the difference as the index to calculate the student performance stability corresponding to each interaction interference level in each language context; A two-dimensional coordinate system is constructed with the degree of interaction interference as the horizontal axis and the stability of student performance as the vertical axis. Several points are marked in the constructed two-dimensional coordinate system for the stability of student performance corresponding to each degree of interaction interference in each language context to form a scatter plot, and a regression line is drawn for the scatter plot, and the slope of the regression line is obtained at the same time; The absolute value of the slope of the regression line is taken as the correlation between the interaction interference corresponding to each language context and the impact on students' performance.

8. A language on-site simulation intelligent teaching system as claimed in claim 1, characterized in that: The operation of selecting the language context that needs to be optimized during the optimization cycle is as follows: The optimization requirement corresponding to each language situation is calculated by calculating the correlation between the interaction interference trend corresponding to each language situation and the influence of the interaction interference corresponding to each language situation on the student's performance; The optimization requirement corresponding to each language context is compared with the system initial setting threshold, and the language contexts with optimization requirements greater than the system initial setting threshold are selected as the language contexts that need to be optimized within the optimization cycle.

9. A language on-site simulation intelligent teaching system as claimed in claim 8, characterized in that: After selecting the language context to be optimized, the selected speech context is judged whether it has a certain optimization direction. The specific judgment is as follows: The selected speech context is recorded as the target language context, and then the conversation duration of the target language context corresponding to the conversation content in each usage record is counted; The degree of interaction interference and the duration of conversation in the target language context in each usage record are correlated and analyzed to obtain the correlation coefficient between interaction interference and conversation duration corresponding to the target language context; The absolute value of the correlation coefficient between the interaction interference and the conversation duration corresponding to the target language context is compared with the effective correlation coefficient of the coefficient configuration. If the absolute value of the correlation coefficient between the interaction interference and the conversation duration corresponding to the target language context is greater than or equal to the effective correlation coefficient, then it is judged that the target language context has a definite optimization direction, namely, limiting the conversation duration. Otherwise, it is analyzed that the target language context does not have a definite optimization direction.

10. A language on-site simulation intelligent teaching system as claimed in claim 6, characterized in that: The process of performing correlation analysis on the interaction interference degree and conversation duration of the target language context in each usage record is as follows: A two-dimensional coordinate system is constructed with the usage time as the horizontal axis and the conversation duration as the vertical axis, and then a number of points are marked in the constructed two-dimensional coordinate system according to the conversation duration in each usage record of the target language context and the usage time corresponding to the usage record, so as to form a conversation duration change curve corresponding to the target language context; The correlation coefficient between the interaction interference change curve corresponding to the target language context and the conversation duration change curve is calculated.

Images