Multi-dimensional oral language ability quantitative evaluation and weakness tracing system

Abstract

The invention relates to the technical field of spoken language calibration, and discloses a multi-dimensional spoken language ability quantitative evaluation and weakness tracing system which comprises a data acquisition module, a model correction module, an ability evaluation module, a weakness tracing module and a feedback module. A more comprehensive, accurate and intelligent evaluation scheme is provided, the system can identify the advantages and weak links of the learner more accurately through multi-dimensional data acquisition and refined spoken language ability analysis, the disadvantage that a traditional evaluation method excessively depends on manual scoring or simple speech recognition is avoided, and the evaluation efficiency is improved. Through dynamic model adjustment and detailed weakness tracing analysis, the system can provide a customized improvement scheme for learners, significantly improve the accuracy and practicability of spoken language ability evaluation, and promote scientization and personalization of spoken language ability training.

Description

Technical Field

[0001] The present invention relates to the technical field of oral calibration, and specifically to a multi-dimensional oral ability quantitative evaluation and weakness tracing system. Background Technique

[0002] With the popularization of intelligent voice devices and the continuous development of human-computer interaction technology, speech ability evaluation and language quality analysis have become important research directions. Especially in industries such as education, medical care, and customer service, the evaluation of oral ability has become increasingly important. More specifically, with the increasing demand for language learning, especially the cultivation of the oral ability of English learners, the role of oral ability evaluation in the field of education has become crucial. By accurately quantifying the oral ability of learners, it can help educators better understand the strengths and weaknesses of learners, so as to conduct effective tutoring.

[0003] Currently, traditional oral ability evaluations usually rely on manual scoring or simple speech recognition technologies, and these methods have many limitations. Manual scoring is greatly affected by subjective factors, and the scoring results may deviate due to different personal understandings and standards of evaluators. And existing speech recognition systems mostly focus on the clarity and accuracy of speech, ignoring more detailed oral ability characteristics, such as speech rate, pitch, syllable amplitude, pause time, vocabulary, etc. This results in traditional evaluation methods being unable to comprehensively and accurately reflect the oral ability of learners, especially in rapidly changing oral expressions, and being unable to timely feedback the deficiencies and potential problems in language expressions. Moreover, existing evaluation systems often lack the function of tracing weaknesses and cannot deeply analyze specific problems in learners' oral language.

[0004] Therefore, we propose a multi-dimensional oral ability quantitative evaluation and weakness tracing system to solve the above-mentioned problems. Summary of the Invention

[0005] The purpose of the present invention is to provide a multi-dimensional oral ability quantitative evaluation and weakness tracing system to solve the problems in the above-mentioned background technique that traditional oral ability evaluations usually rely on manual scoring or simple speech recognition technologies, and these methods have many limitations. Manual scoring is greatly affected by subjective factors, and the scoring results may deviate due to different personal understandings and standards of evaluators.

[0006] To achieve the above purpose, the present invention provides the following technical solution: A multi-dimensional oral ability quantitative evaluation and weakness tracing system, including a data acquisition module, a model correction module, an ability evaluation module, a weakness tracing module, and a feedback module; The data acquisition module is used to obtain multi-source data of the user and the detection device, and preprocess the obtained data to generate a first data set and a second data set; The model correction module is used to perform integrated calculations on the first data set to generate a model correction judgment coefficient PDX and a model correction reference coefficient T, and analyze the model correction judgment coefficient PDX to determine whether to use the model correction reference coefficient T; The ability evaluation module is used to perform integrated analysis on the second data set and the model correction reference coefficient T to generate a spoken language ability analysis coefficient KYX, and analyze the spoken language ability analysis coefficient KYX to determine whether there are spoken language ability problems; The weakness tracing module is used to perform integrated analysis on the first data set, calculate and obtain a speech rate problem tracing coefficient CF, a pitch problem tracing coefficient CG, and an amplitude problem tracing coefficient CH, and analyze the three to perform source tracing; The feedback module is used to display various data.

[0007] Preferably, the data acquisition module includes a first acquisition unit, a second acquisition unit, and a data preprocessing unit; The first acquisition unit is used to collect the signal parameters of the instrument to obtain the time-domain signal A of the instrument, the sampling rate B, the bit rate C, the microphone response rate D, and the signal-to-noise ratio E; The second acquisition unit collects the data of the user to obtain the speech rate F, the pitch G, the syllable amplitude H, the pause time I, the vocabulary J, and the vocabulary evaluation rate K of the user: The data preprocessing unit is used to organize the collected multi-source data to generate a first data set and a second data set; The first data set includes the time-domain signal A, the sampling rate B, the bit rate C, the microphone response rate D, and the signal-to-noise ratio E; The second data set includes the speech rate F, the pitch G, the syllable amplitude H, the pause time I, the vocabulary J, and the vocabulary evaluation rate K.

[0008] Preferably, the model correction module includes a correction coefficient calculation unit and a correction coefficient analysis unit; The correction coefficient calculation unit is used to perform integrated analysis on the first data set to calculate and generate a model correction judgment coefficient PDX and a model correction reference coefficient T; The correction coefficient analysis unit is used to perform data analysis on the model correction judgment coefficient PDX to determine whether to activate the model correction reference coefficient T.

[0009] Preferably, the correction coefficient calculation unit calculates and obtains the model correction judgment coefficient PDX and the model correction reference coefficient T through the following formula: ; ;

[0010] Where: A is the time-domain signal, B is the sampling rate, C is the bit rate, D is the microphone response rate, E is the signal-to-noise ratio, a1, a2, a3, b1, b2, and b3 are weight values, and the values of a1, a2, a3, b1, b2, and b3 are adjusted and set by the user.

[0011] Preferably, the specific manner in which the correction coefficient analysis unit performs data analysis on the model correction judgment coefficient PDX is as follows: When it means that the subsequent calculation does not need to start the model correction reference coefficient T; When it means that the subsequent calculation needs to start the model correction reference coefficient T.

[0012] Preferably, the ability evaluation module includes an ability evaluation coefficient calculation unit and an ability evaluation coefficient analysis unit. The ability evaluation coefficient calculation unit is used to calculate and obtain the spoken language ability analysis coefficient KYX; The ability evaluation coefficient analysis unit is used to analyze the calculated and obtained spoken language ability analysis coefficient KYX.

[0013] Preferably, the ability evaluation coefficient calculation unit calculates and obtains the spoken language ability analysis coefficient KYX through the following formula: ; ;

[0014] Where: c1, c2, c3, c4, and b5 are weight values, and the values of c1, c2, c3, c4, and b5 are adjusted and set by the user, and T is the model correction reference coefficient; F is the speaking speed, H is the syllable amplitude, G is the pitch, J is the vocabulary, I is the pause time, and K is the vocabulary evaluation rate.

[0015] Preferably, the specific analysis manner of the ability evaluation coefficient analysis unit for the spoken language ability analysis coefficient KYX is as follows: When it means that the user has a problem of insufficient spoken language ability; When it means that the user does not have a problem of insufficient spoken language ability; Where KYX1 is the spoken language ability analysis coefficient calculated without using the model correction reference coefficient T, and KYX2 is the spoken language ability analysis coefficient calculated using the model correction reference coefficient T.

[0016] Preferably, the weakness tracing module includes a weakness coefficient calculation unit and a weakness coefficient analysis unit. The weakness coefficient calculation unit is used to calculate and obtain the speaking speed problem tracing coefficient CF, the pitch problem tracing coefficient CG, and the amplitude problem tracing coefficient CH; The weakness coefficient analysis unit is used to analyze the calculated speech rate problem traceability coefficient CF, pitch problem traceability coefficient CG, and amplitude problem traceability coefficient CH. The specific method is as follows: When it indicates that the user has any problem in speech rate, pitch, or amplitude; When it indicates that the user does not have any problem in speech rate, pitch, or amplitude.

[0017] Preferably, the weakness coefficient calculation unit calculates and obtains the speech rate problem traceability coefficient CF, pitch problem traceability coefficient CG, and amplitude problem traceability coefficient CH through the following formulas respectively: ; ; ;

[0018] In the formula: d1, d2, and d3 are weight values, and the values of d1, d2, and d3 are adjusted and set by the user; F is the speech rate, H is the syllable amplitude, G is the pitch, J is the vocabulary, I is the pause time, and K is the vocabulary evaluation rate.

[0019] Compared with the prior art, the beneficial effects of the present invention are: 1. Through the collaborative action of multiple modules, this system breaks through the limitations of traditional oral assessment methods and provides a more comprehensive, accurate, and intelligent assessment solution. Through multi-dimensional data collection and refined oral ability analysis, this system can more accurately identify the advantages and weaknesses of learners, avoiding the disadvantages of relying too much on manual scoring or simple speech recognition in traditional assessment methods. Through dynamic adjustment of the model and detailed weakness traceability analysis, this system can provide customized improvement plans for learners, significantly improving the accuracy and practicality of oral ability assessment, and promoting the scientific and personalized oral ability training.

[0020] 2. The weakness coefficient calculation unit uses mathematical formulas to calculate the speech rate problem traceability coefficient CF, pitch problem traceability coefficient CG, and amplitude problem traceability coefficient CH. The calculation of each coefficient comprehensively considers different oral performance characteristics, including speech rate F, pitch G, syllable amplitude H, pause time I, vocabulary J, and vocabulary evaluation rate K. Through such a design, the system can conduct refined analysis for different oral characteristics and accurately trace the specific problems existing in the learners' oral performance. This quantitative analysis method is more scientific and accurate than traditional oral assessment methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is the overall three-dimensional structure schematic diagram of the present invention.

[0022] In the figure: 1. Data acquisition module; 11. First acquisition unit; 12. Second acquisition unit; 13. Data preprocessing unit; 2. Model correction module; 21. Correction coefficient calculation unit; 22. Correction coefficient analysis unit; 3. Ability evaluation module; 31. Ability evaluation coefficient calculation unit; 32. Ability evaluation coefficient analysis unit; 4. Weakness tracing module; 41. Weakness coefficient calculation unit; 42. Weakness coefficient analysis unit; 5. Feedback module. Specific implementation manner

[0023] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0024] Embodiment 1: Please refer to Figure 1 , a multi-dimensional oral ability quantitative evaluation and weakness tracing system, including a data acquisition module 1, a model correction module 2, an ability evaluation module 3, a weakness tracing module 4, and a feedback module 5; The data acquisition module 1 is used to obtain multi-source data of the user and the detection device, and preprocess the obtained data to generate a first data set and a second data set; The model correction module 2 is used to perform integration calculation on the first data set to generate a model correction judgment coefficient PDX and a model correction reference coefficient T, and analyze the model correction judgment coefficient PDX to determine whether to use the model correction reference coefficient T; The ability evaluation module 3 is used to perform integration analysis on the second data set and the model correction reference coefficient T to generate an oral ability analysis coefficient KYX, and analyze the oral ability analysis coefficient KYX to determine whether there is an oral ability problem; The weakness tracing module 4 is used to perform integration analysis on the first data set, calculate and obtain a speech rate problem tracing coefficient CF, a pitch problem tracing coefficient CG, and an amplitude problem tracing coefficient CH, and analyze the three to perform tracing; The feedback module 5 is used to display each item of data.

[0025] In this embodiment: The role of the data acquisition module 1 in the system is to provide basic data for subsequent analysis and evaluation through multi-source data acquisition, provide accurate and comprehensive input data for subsequent analysis, and thus ensure the effectiveness of subsequent processing steps.

[0026] The key role of the model correction module 2 is to generate the model correction judgment coefficient PDX and the model correction reference coefficient T through the integrated calculation of the first data set, and analyze PDX. The core task of this process is to deeply analyze the spoken language data to determine whether the model needs to be adjusted according to the current spoken language performance. During the calculation process, the model correction module not only considers various quality parameters of audio signals, but also decides whether to activate the model correction reference coefficient T based on the analysis results. The purpose of this is to dynamically adjust the model to cope with different speech characteristics, and further improve the accuracy and adaptability of the evaluation system.

[0027] The ability evaluation module 3 is the core part of the system for calculating and analyzing the spoken language ability analysis coefficient KYX. This module integrates the spoken language feature parameters of the second data set and the model correction reference coefficient T, and generates a quantitative spoken language ability evaluation result through the analysis of these data. If the spoken language ability analysis coefficient KYX reaches the preset standard, the system indicates that there is no problem with the user's spoken language ability, otherwise potential spoken language problems can be identified. Through this precise quantitative evaluation, the ability evaluation module can provide an intuitive and objective spoken language ability assessment, helping learners and educators effectively identify and improve deficiencies in spoken language.

[0028] The main function of the weakness tracing module 4 is to conduct an integrated analysis of the first data set, thereby calculating the speech rate problem tracing coefficient CF, the pitch problem tracing coefficient CG, and the amplitude problem tracing coefficient CH. This module can deeply explore specific problems that may exist in spoken language, such as too fast speech rate, inappropriate pitch, or insufficient amplitude, through the analysis of these coefficients, and then help users clarify specific language weaknesses. Through this tracing process, the system not only identifies the overall problems of spoken language ability, but can also provide specific feedback, thus providing targeted improvement plans for users.

[0029] The feedback module 5 provides users with comprehensive and clear feedback results to help users understand the various data in the evaluation and analysis process. Through this module, users can intuitively view all relevant data, analysis results, and correction suggestions. The feedback module can present the data, analysis results, and system suggestions to users to ensure the transparency and effectiveness of information. In this way, learners can not only obtain an overall evaluation of their spoken language ability, but also receive specific guidance on how to improve their spoken language ability, promoting the continuous improvement and growth of learners.

[0030] Through the collaborative efforts of multiple modules, this system breaks through the limitations of traditional oral assessment methods and provides a more comprehensive, accurate, and intelligent assessment solution. Through multi-dimensional data collection and refined oral ability analysis, this system can more accurately identify the strengths and weaknesses of learners, avoiding the drawbacks of over-reliance on manual scoring or simple speech recognition in traditional assessment methods. Through dynamic adjustment of the model and detailed weakness traceability analysis, this system can provide learners with customized improvement plans, significantly enhancing the accuracy and practicality of oral ability assessment and promoting the scientific and personalized training of oral ability.

[0031] Example 2: Please refer to Figure 1 , the data collection module 1 includes a first collection unit 11, a second collection unit 12, and a data preprocessing unit 13; The first collection unit 11 is used to collect the signal parameters of the instrument, thereby obtaining the time-domain signal A of the instrument, the sampling rate B, the bit rate C, the microphone response rate D, and the signal-to-noise ratio E; The second collection unit 12 collects data of the user, thereby obtaining the speaking speed F, the pitch G, the syllable amplitude H, the pause time I, the vocabulary J, and the vocabulary rating K of the user: The data preprocessing unit 13 is used to organize the multi-source data collected, thereby generating a first data set and a second data set; The first data set includes the time-domain signal A, the sampling rate B, the bit rate C, the microphone response rate D, and the signal-to-noise ratio E; The second data set includes the speaking speed F, the pitch G, the syllable amplitude H, the pause time I, the vocabulary J, and the vocabulary rating K.

[0032] In this embodiment: The data collection module 1 includes a first collection unit 11, a second collection unit 12, and a data preprocessing unit 13. The design of this module aims to comprehensively capture the user's oral performance and related speech quality parameters through multi-source data collection, providing accurate data support for subsequent analysis and evaluation.

[0033] The first collection unit 11 is responsible for collecting signal parameters from the instrument, obtaining the relevant time-domain signal A, sampling rate B, bit rate C, microphone response rate D, and signal-to-noise ratio E. These parameters help evaluate the quality and clarity of the speech signal and provide basic data at the audio signal level for subsequent analysis.

[0034] The second collection unit 12 is responsible for collecting relevant data from the user's oral expression, including the speaking speed F, the pitch G, the syllable amplitude H, the pause time I, the vocabulary J, and the vocabulary rating K. These data reflect the language ability characteristics of the user and can help accurately evaluate the fluency and richness of oral expression.

[0035] The data preprocessing unit 13 sorts and processes the collected multi-source data to generate two data sets - the first data set and the second data set. The first data set includes the time-domain signal A, sampling rate B, bit rate C, microphone response rate D, and signal-to-noise ratio E, which mainly reflect the quality of the speech signal; the second data set includes the speech rate F, pitch G, syllable amplitude H, pause time I, vocabulary size J, and vocabulary rating K, which mainly reflect the expression characteristics of the spoken language. Through this preprocessing process, the system can effectively clean noise and standardize data, providing consistent and high-quality input for subsequent analysis.

[0036] This system respectively collects the signal parameters from the instrument and the spoken language performance data of the user through the first acquisition unit 11 and the second acquisition unit 12, comprehensively covering multiple dimensions such as speech quality and expression characteristics. This multi-source data acquisition method breaks through the limitation of traditional spoken language assessment relying only on a single data source, and can more comprehensively and accurately evaluate the user's spoken language ability. Through the integration and construction of the first data set and the second data set, the system can comprehensively grasp the user's spoken language performance, including both the quality of the speech signal such as the signal-to-noise ratio and sampling rate, and the user's spoken language characteristics such as the speech rate and vocabulary size. This provides a more accurate data basis for subsequent spoken language ability analysis, problem diagnosis, and ability improvement.

[0037] The data preprocessing unit 13 cleans, standardizes, and sorts the collected multi-source data, eliminating noise and inconsistencies in the data. This process ensures the quality and consistency of the data, helps improve the accuracy of subsequent model correction, ability assessment, and weakness tracing, and ensures the stable operation of the entire system.

[0038] Example three: Please refer to Figure 1 , the model correction module 2 includes a correction coefficient calculation unit 21 and a correction coefficient analysis unit 22; The correction coefficient calculation unit 21 is used to perform integrated analysis on the first data set to calculate and generate a model correction judgment coefficient PDX and a model correction reference coefficient T; The correction coefficient analysis unit 22 is used to perform data analysis on the model correction judgment coefficient PDX to determine whether to activate the model correction reference coefficient T.

[0039] The correction coefficient calculation unit 21 calculates and obtains the model correction judgment coefficient PDX and the model correction reference coefficient T through the following formula: ; ;

[0040] Where: A is the time-domain signal, B is the sampling rate, C is the bit rate, D is the microphone response rate, E is the signal-to-noise ratio, a1, a2, a3, b1, b2, and b3 are weight values, and the values of a1, a2, a3, b1, b2, and b3 are adjusted and set by the user.

[0041] The specific way for the correction coefficient analysis unit 22 to perform data analysis on the model correction judgment coefficient PDX is as follows: When it means that the subsequent calculation does not need to start the model correction reference coefficient T; When it means that the subsequent calculation needs to start the model correction reference coefficient T.

[0042] In this embodiment: The correction coefficient calculation unit 21 uses an accurate mathematical formula to calculate and generate the model correction judgment coefficient PDX and the model correction reference coefficient T. This calculation integrates multiple factors such as the time-domain signal A, sampling rate B, bit rate C, microphone response rate D, and signal-to-noise ratio E in the first dataset, considering multi-dimensional speech quality parameters, making the model correction more refined and accurate. Each weight value a1, a2, a3, b1, b2, b3 in the formula can be adjusted by the user according to actual needs, making this module highly flexible and adaptable, and capable of dynamically adjusting according to the specific spoken language characteristics of different users.

[0043] By weighting and integrating multiple speech quality parameters through the formula, the calculation results of PDX and T are more accurate and reliable. Specifically, the formula introduces weight settings for each parameter, which can reflect various quality differences in the speech signal. For example, through combination terms such as A×E and C², various key features that may affect the spoken language ability in the speech signal can be accurately evaluated, and thus a more reasonable model correction basis can be provided for subsequent ability evaluation. Users can adjust the weight values according to the actual situation to ensure that the model correction can better adapt to the spoken language performance of different users.

[0044] Through the analysis of PDX, the correction coefficient analysis unit 22 can automatically determine whether to start the model correction reference coefficient T. This automated mechanism based on threshold judgment enables the system to dynamically adjust the model according to real-time data during the spoken language ability evaluation without manual intervention. Specifically, when PDX≤0.4, the system believes that the spoken language performance meets the expectations and no correction is needed; while when PDX>0.4, it means that the system detects potential problems and needs to optimize the evaluation model by adjusting the model correction reference coefficient T to ensure the accuracy of the evaluation results. This intelligent adjustment mechanism significantly improves the automation level of the system and reduces human errors.

[0045] Embodiment 4: Please refer to Figure 1, the ability evaluation module 3 includes an ability evaluation coefficient calculation unit 31 and an ability evaluation coefficient analysis unit 32. The ability evaluation coefficient calculation unit 31 is used to calculate and obtain the spoken language ability analysis coefficient KYX; The ability evaluation coefficient analysis unit 32 is used to analyze the calculated and obtained spoken language ability analysis coefficient KYX.

[0046] The ability evaluation coefficient calculation unit 31 calculates and obtains the spoken language ability analysis coefficient KYX through the following formula: ; ;

[0047] In the formula: c1, c2, c3, c4, and b5 are weight values, and the values of c1, c2, c3, c4, and b5 are adjusted and set by the user, and T is the model correction reference coefficient; F is the speech rate, H is the syllable amplitude, G is the pitch, J is the vocabulary, I is the pause time, and K is the vocabulary evaluation rate.

[0048] The specific analysis method of the ability evaluation coefficient analysis unit 32 for the spoken language ability analysis coefficient KYX is as follows: When , it represents that the user has a problem of insufficient spoken language ability; When , it represents that the user does not have a problem of insufficient spoken language ability; Where KYX1 is the spoken language ability analysis coefficient calculated without using the model correction reference coefficient T, and KYX2 is the spoken language ability analysis coefficient calculated using the model correction reference coefficient T.

[0049] In this embodiment: The ability evaluation coefficient calculation unit 31 calculates the spoken language ability analysis coefficient KYX through a complex mathematical formula. This coefficient combines multi-dimensional spoken language performance parameters, including the speech rate F, syllable amplitude H, pitch G, vocabulary J, pause time I, and vocabulary evaluation rate K, as well as the correction coefficient T. In the specific calculation formula, the formula can quantify the spoken language ability level of the learner through weighted integration of these spoken language feature parameters, ensuring a more comprehensive and refined evaluation. The weight values c1, c2, c3, c4, and c5 can be adjusted by the user, so as to be flexibly configured according to different spoken language evaluation requirements, enhancing the adaptability and personalization of the system.

[0050] Through the calculation of KYX1 and KYX2, the system can provide the evaluation results without using the model correction reference coefficient T and with the corrected coefficient at the same time. This design endows the system with the ability of dynamic model correction, enabling it to automatically adjust the model according to the complexity of spoken language performance during evaluation. For example, when the values of KYX1 and KYX2 change, the system can determine whether it is necessary to correct the spoken language ability evaluation model, further improving the accuracy and reliability of the evaluation results.

[0051] During the analysis of KYX1 and KYX2 by the ability evaluation coefficient analysis unit 32, clear threshold judgment criteria are set: when KYX1 or KYX2 ≤ 0.55, it indicates that the user has problems with insufficient spoken language ability; while when KYX1 or KYX2 > 0.55, it indicates that there are no obvious problems with the spoken language ability. Through this intelligent analysis, the system can quickly and accurately identify the user's spoken language ability level, further guiding the user to improve their spoken language expression. This standardized evaluation mechanism makes the spoken language ability evaluation more objective and scientific, avoiding the common subjective biases in manual scoring.

[0052] Example Five: Please refer to Figure 1 , the weakness tracing module 4 includes a weakness coefficient calculation unit 41 and a weakness coefficient analysis unit 42. The weakness coefficient calculation unit 41 is used to calculate and obtain the speech rate problem tracing coefficient CF, the pitch problem tracing coefficient CG, and the amplitude problem tracing coefficient CH; The weakness coefficient analysis unit 42 is used to analyze the calculated speech rate problem tracing coefficient CF, the pitch problem tracing coefficient CG, and the amplitude problem tracing coefficient CH, and the specific method is as follows: When , it represents that the user has any one of the problems of speech rate, pitch, or amplitude; When , it represents that the user does not have any one of the problems of speech rate, pitch, or amplitude.

[0053] The weakness coefficient calculation unit 41 calculates and obtains the speech rate problem tracing coefficient CF, the pitch problem tracing coefficient CG, and the amplitude problem tracing coefficient CH respectively through the following formulas: ; ; ;

[0054] In the formula: d1, d2, and d3 are weight values, and the values of d1, d2, and d3 are adjusted and set by the user; F is the speech rate, H is the syllable amplitude, G is the pitch, J is the vocabulary size, I is the pause time, and K is the vocabulary evaluation rate.

[0055] In this embodiment: The weakness coefficient calculation unit 41 uses mathematical formulas to calculate the speech rate problem traceability coefficient CF, the pitch problem traceability coefficient CG, and the amplitude problem traceability coefficient CH. The calculation of each coefficient comprehensively considers different oral performance characteristics, including speech rate F, pitch G, syllable amplitude H, pause time I, vocabulary J, and vocabulary rating K. Through such a design, the system can conduct refined analysis for different oral characteristics and accurately trace the specific problems existing in the learner's oral performance. This quantitative analysis method is more scientific and accurate than traditional oral assessment methods.

[0056] By calculating CF, CG, and CH, this system can clearly identify the specific problems of the user in terms of speech rate, pitch, or amplitude. CF can accurately identify the situation of too fast or too slow speech rate by analyzing the ratio of speech rate to pause time; CG can identify the situation where the pitch deviates from the normal range by weighing pitch and pause time; while CH can help discover the mismatch between syllable amplitude and vocabulary. Such weakness traceability analysis can provide clear and accurate basis for subsequent personalized training, ensuring the pertinence of the training plan.

[0057] The weight values d1, d2, d3 in the formula are adjusted and set by the user, which enables the system to perform personalized optimization according to the needs or specific situations of different users. For example, for some learners, they may be more concerned about the optimization of speech rate, while for other users, they may need to pay more attention to pitch or amplitude. The adjustable weight makes the system more flexible and can better adapt to the oral ability assessment needs of different users.

[0058] The weakness coefficient analysis unit 42 analyzes the calculated CF, CG, and CH and makes a judgment according to the preset threshold: when any one of the coefficients is less than or equal to 0.45, the system will automatically determine that the user has problems in terms of speech rate, pitch, or amplitude; when the coefficient is greater than 0.45, it indicates that the user has no related problems. This analysis mechanism not only has strong intelligent judgment ability, but also can automatically identify and feedback specific oral weaknesses according to real-time assessment. This intelligent analysis function greatly improves the accuracy and real-time performance of the system assessment, helping learners identify and correct the problems existing in their oral language faster.

[0059] The content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0060] Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A multi-dimensional oral ability quantitative assessment and weakness tracing system, characterized by: It includes a data collection module (1), a model modification module (2), a capability assessment module (3), a weakness tracing module (4) and a feedback module (5); The data acquisition module (1) is used to acquire multi-source data from the user and the detection device, and pre-process the acquired data, thereby generating a first data set and a second data set; The model correction module (2) is used to perform integrated calculation on the first data set to generate a model correction judgment coefficient PDX and a model correction reference coefficient T, and to analyze the model correction judgment coefficient PDX to determine whether to use the model correction reference coefficient T; The ability assessment module (3) integrates and analyzes the second data set and the model correction reference coefficient T to generate a speaking ability analysis coefficient KYX, and analyzes the speaking ability analysis coefficient KYX to determine whether a speaking ability problem occurs; The weakness tracing module (4) is used to integrate and analyze the first data set, calculate and obtain the speech rate problem tracing coefficient CF, the pitch problem tracing coefficient CG and the amplitude problem tracing coefficient CH, and analyze the three to perform source tracing; The feedback module (5) is used to display various data.

2. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 1, characterized in that: The data acquisition module (1) comprises a first acquisition unit (11), a second acquisition unit (12) and a data preprocessing unit (13); The first acquisition unit (11) is used to acquire signal parameters of the instrument, thereby obtaining the time domain signal A, sampling rate B, bit rate C, microphone response rate D and signal-to-noise ratio E of the instrument; The second collection unit (12) collects the user's data, thereby obtaining the user's speech speed F, pitch G, syllable amplitude H, pause time I, vocabulary J and vocabulary evaluation rate K: The data preprocessing unit (13) is used to sort the collected multi-source data, thereby generating a first data set and a second data set; The first data set includes a time domain signal A, a sampling rate B, a bit rate C, a microphone response rate D, and a signal-to-noise ratio E; The second data set includes speech rate F, pitch G, syllable amplitude H, pause time I, vocabulary size J and vocabulary rating rate K.

3. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 2, characterized in that: The model correction module (2) comprises a correction coefficient calculation unit (21) and a correction coefficient analysis unit (22); The correction coefficient calculation unit (21) is used to perform integrated analysis on the first data set to calculate and generate a model correction judgment coefficient PDX and a model correction reference coefficient T; The correction coefficient analysis unit (22) is used to perform data analysis on the model correction judgment coefficient PDX to determine whether the model correction reference coefficient T needs to be started.

4. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 3, characterized in that: The correction coefficient calculation unit (21) calculates and obtains the model correction judgment coefficient PDX and the model correction reference coefficient T through the following formula: ; ; Wherein: A is the time domain signal, B is the sampling rate, C is the bit rate, D is the microphone response rate, E is the signal-to-noise ratio, a1, a2, a3, b1, b2 and b3 are weight values, and the values ​​of a1, a2, a3, b1, b2 and b3 are adjusted and set by the user.

5. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 4, characterized in that: The specific method of the correction coefficient analysis unit (22) for performing data analysis on the model correction judgment coefficient PDX is as follows: when When , it means that the subsequent calculation does not need to start the model correction reference coefficient T; when , it means that the subsequent calculation needs to start the model correction reference coefficient T.

6. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 5, characterized in that: The ability assessment module (3) comprises an ability assessment coefficient calculation unit (31) and an ability assessment coefficient analysis unit (32), wherein the ability assessment coefficient calculation unit (31) is used to calculate and obtain the oral ability analysis coefficient KYX; The ability evaluation coefficient analysis unit (32) is used to analyze the calculated oral ability analysis coefficient KYX.

7. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 6, characterized in that: The ability evaluation coefficient calculation unit (31) calculates and obtains the oral ability analysis coefficient KYX by the following formula: ; ; Wherein: c1, c2, c3, c4 and b5 are weight values, and the values ​​of c1, c2, c3, c4 and b5 are adjusted and set by the user, and T is the model correction reference coefficient; F is speaking rate, H is syllable amplitude, G is pitch, J is vocabulary, I is pause time, and K is vocabulary rating rate.

8. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 7, characterized in that: The specific analysis method of the ability evaluation coefficient analysis unit (32) for the oral ability analysis coefficient KYX is as follows: when When the user has insufficient oral skills, when When the user has no problem of insufficient oral ability; KYX1 is the oral proficiency analysis coefficient obtained by calculation without using the model-corrected reference coefficient T, and KYX2 is the oral proficiency analysis coefficient obtained by calculation using the model-corrected reference coefficient T.

9. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 8, characterized in that: The weakness tracing module (4) comprises a weakness coefficient calculation unit (41) and a weakness coefficient analysis unit (42), wherein the weakness coefficient calculation unit (41) is used to calculate and obtain a speech rate problem tracing coefficient CF, a pitch problem tracing coefficient CG and an amplitude problem tracing coefficient CH; The weakness coefficient analysis unit (42) is used to analyze the calculated speech rate problem tracing coefficient CF, the pitch problem tracing coefficient CG and the amplitude problem tracing coefficient CH, in the following specific manner: when When , it means that the user has any problem with speech speed, pitch or amplitude; when , it means the user has no problems with speaking speed, pitch or amplitude.

10. A multi-dimensional oral ability quantitative assessment and weakness tracing system according to claim 9, characterized in that: The weakness coefficient calculation unit (41) calculates and obtains the speech rate problem tracing coefficient CF, the pitch problem tracing coefficient CG and the amplitude problem tracing coefficient CH respectively through the following formulas: ; ; ; Wherein: d1, d2 and d3 are weight values, and the values ​​of d1, d2 and d3 are adjusted and set by the user; F is speaking rate, H is syllable amplitude, G is pitch, J is vocabulary, I is pause time, and K is vocabulary rating rate.

Images