Chinese composition correcting method and system based on intelligent algorithm

Abstract

The invention relates to the technical field of Chinese composition correction, and discloses a Chinese composition correction method and system based on an intelligent algorithm. The Chinese composition correction system comprises a data annotation and preprocessing module, a feature extraction module, a multi-mode learning module, a model optimization and integration module, a real-time feedback and interaction module, a personalized guidance module and a security and privacy module, wherein the multi-mode learning module comprises image and text combination and audio and text combination. According to the Chinese composition correction method and system based on the intelligent algorithm, the naturalness and emotion expression of speech expression evaluation are improved by arranging a multi-modal learning module and a model optimization and integration module of the Chinese composition correction system and utilizing image and text combination and audio and text combination; the model needs to be continuously optimized to improve the accuracy and credibility of evaluation, and the effect of improving the accuracy and the use effect of the Chinese composition correcting method and system is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of Chinese composition correction, and specifically to a Chinese composition correction method and system based on intelligent algorithms. Background Technique

[0002] Natural language processing technology has gradually penetrated into fields such as Chinese compositions. In some work on diagnosing basic dimensions of compositions and statistical analysis, computers can share some of the more trivial work of teachers. Therefore, it is necessary to use a Chinese composition correction system based on intelligent algorithms to correct Chinese compositions.

[0003] According to the publication number CN113360608A, a Chinese composition correction system and method combining humans and machines are disclosed. The system includes a composition acquisition system, a preprocessing system, a correction system, and a material recommendation system.

[0004] Adopting the above technical solution can solve the problems existing in the prior art, can present intuitive correction results, and provide more functions. However, in the above technical solution, the correction model of the composition correction system cannot train an accurate model, and the composition evaluation has a certain degree of subjectivity. The model cannot be optimized to improve the accuracy and credibility of the evaluation, thus affecting the evaluation of complex Chinese compositions by the Chinese composition correction method and system, and reducing the accuracy and usage effect of the Chinese composition correction method and system. Summary of the Invention

[0005] The purpose of the present invention is to provide a Chinese composition correction method and system based on intelligent algorithms to solve the problems raised in the above background technique.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A Chinese composition correction method and system based on intelligent algorithms, including a Chinese composition correction system. The Chinese composition correction system includes a data annotation and preprocessing module, a feature extraction module, a multimodal learning module, a model optimization and integration module, a real-time feedback and interaction module, a personalized guidance module, and a security and privacy module. The multimodal learning module includes the combination of image and text and the combination of audio and text. Based on the combination of image and text, for compositions involving illustrations or pictures, image recognition technology can be introduced to evaluate the degree of combination of pictures and text. Based on the combination of audio and text, for compositions of recitation or oral expression, speech recognition technology can be introduced to evaluate the naturalness of speech expression and emotional expression.

[0007] The model optimization and integration module includes a model optimization unit and a model integration unit. The model optimization unit includes a deep learning model and multi-task learning. Based on the deep learning model, the text understanding and generation capabilities are improved. Based on the multi-task learning, the comprehensive performance of the model is enhanced. The model integration unit includes ensemble learning and an adaptive model. Based on the ensemble learning, the prediction results of multiple models can be combined to improve the accuracy and stability of the overall score. Based on the adaptive model, the weights and parameters of the model can be dynamically adjusted according to the historical performance and characteristics of the students to achieve personalized evaluation;

[0008] The real-time feedback and interaction module includes real-time feedback and an interaction platform. Based on the real-time feedback and interaction platform, it is convenient for students and teachers to conduct feedback and interactive communication. The personalized guidance module includes a personalized model and dynamic adjustment. Based on the personalized model and dynamic adjustment, the learning history, writing characteristics, and common mistakes of each student can be recorded, and the evaluation criteria and feedback content of the model can be dynamically adjusted according to the progress of the students.

[0009] Preferably, the data annotation and preprocessing module includes a data annotation unit and a data preprocessing unit. Based on the data annotation unit and the data preprocessing unit, the data collected by the Chinese composition correction system can be annotated and preprocessed.

[0010] Through the data annotation unit, the data can be annotated in multiple dimensions and experts can be invited to annotate it. Also, through the data preprocessing unit, the data can be preprocessed in many aspects to facilitate the use of the Chinese composition correction method and system to correct Chinese compositions.

[0011] Preferably, the data annotation unit includes multi-dimensional annotation and expert annotation. Based on the multi-dimensional annotation, in addition to traditional sentiment analysis and part-of-speech tagging, multi-dimensional annotation of the composition can be increased. Based on the expert annotation, Chinese teachers and literature experts can be invited to annotate some compositions in detail to provide high-quality annotation data.

[0012] Through multi-dimensional annotation and expert annotation, it is convenient to disassemble and analyze the composition, thereby improving the accuracy of the Chinese composition correction method and system.

[0013] Preferably, the data preprocessing unit includes context awareness, sentiment analysis, and splicing processing. Based on the context awareness, context-aware word segmentation and part-of-speech tagging technologies can be introduced to improve the understanding and processing capabilities of specific domains. Based on the sentiment analysis, more advanced sentiment analysis algorithms are used to improve the accuracy of sentiment recognition. Based on the splicing processing, for long texts, text fragment splicing technology can be used to ensure that the model can process complete compositions.

[0014] Utilize the context awareness, sentiment analysis, and splicing processing of the data preprocessing unit to precisely preprocess the data, process the data from multiple aspects, and thereby improve the data processing ability of the Chinese composition correction method and system.

[0015] Preferably, the feature extraction module includes a language feature unit, a content feature unit, and an emotion feature unit. Based on the language feature unit, content feature unit, and emotion feature unit of the feature extraction module, the features of Chinese compositions can be extracted.

[0016] The language feature unit can be used to extract the language features in the composition, the content feature unit can be used to extract the content features in the composition, and the emotion feature unit can be used to extract the emotion features in the composition.

[0017] Preferably, the language feature unit includes advanced vocabulary recognition, syntactic analysis, and text coherence. Based on the advanced vocabulary recognition, a word vector model can be used to identify the use of advanced vocabulary and idioms. Based on the syntactic analysis, dependency syntactic analysis can be introduced to more precisely identify the structure and logical relationships of sentences. Based on the text coherence, a text coherence model can be used to evaluate whether the logical relationships and transitions between sentences are natural.

[0018] Identify the use of advanced vocabulary and idioms through a word vector model, identify the structure and logical relationships of sentences using syntactic analysis, and evaluate whether the logical relationships and transitions between sentences are natural using text coherence.

[0019] Preferably, the content feature unit includes topic relevance, content depth, and innovation. Based on the topic relevance, a topic model and keyword extraction technology can be used to evaluate whether the composition closely adheres to the topic. Based on the content depth, semantic similarity analysis can be used to evaluate the depth of exploration of the topic content in the composition. Based on the innovation, a generative adversarial network can be used to generate a benchmark for innovative content to evaluate the innovation of the composition.

[0020] Evaluate whether the composition closely adheres to the topic through topic relevance, evaluate the depth of exploration of the topic content in the composition using content depth, and evaluate the innovation of the composition using innovation performance.

[0021] Preferably, the emotion feature unit includes emotion clues and emotional fluctuations. Based on the emotion clues, the emotion clues can be identified and analyzed. Based on the emotional fluctuations, the reasonableness of the overall emotional fluctuations can be evaluated.

[0022] Identify and analyze the emotion clues in the composition through emotion clues, and evaluate the overall emotional fluctuations using emotional fluctuations.

[0023] Preferably, the security and privacy module includes data protection and transparency. Based on the data protection, the security of student data can be ensured, relevant laws and regulations can be complied with, and student privacy can be protected. Based on the transparency, a transparent evaluation process and results can be provided, enabling students and teachers to understand how the model works.

[0024] By utilizing the data protection and transparency of the security and privacy module, the data security of this Chinese composition grading system can be improved, thereby protecting the personal privacy of students' compositions.

[0025] A grading method for a Chinese composition grading system based on intelligent algorithms includes the following steps:

[0026] S1: Define the requirements and goals of the system and collect user requirements;

[0027] S2: Collect and annotate high-quality composition data to establish a dataset;

[0028] S3: Select a suitable model, train and optimize the model;

[0029] S4: Develop the front-end and back-end systems to implement data processing, model deployment, and user interface;

[0030] S5: Conduct system testing to evaluate the performance of the model and user experience;

[0031] S6: Deploy the system and perform iterative optimization based on user feedback.

[0032] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0033] First, the present invention is provided with a multi-modal learning module and a model optimization and integration module in the Chinese composition correction system. By combining images and text and audio and text in the multi-modal learning module, for compositions involving illustrations or pictures, image recognition technology can be introduced to evaluate the combination degree of pictures and text through the combination of images and text. For compositions of recitation or oral expression, speech recognition technology can be introduced through the combination of audio and text to evaluate the naturalness of speech expression and emotional expression. Moreover, by using the model optimization unit and the model integration unit in the model optimization and integration module, through the deep learning model and multi-task learning in the model optimization unit, the text understanding and generation ability can be conveniently improved by using the deep learning model, and the comprehensive performance of the model can be improved by using multi-task learning. The integrated learning and adaptive model in the model integration unit can combine the prediction results of multiple models through integrated learning to improve the accuracy and stability of the overall score, and use the adaptive model to dynamically adjust the weights and parameters of the model according to the historical performance and characteristics of students to achieve personalized evaluation. Thus, an accurate model can be trained. Since composition evaluation has a certain degree of subjectivity, the model needs to be continuously optimized to improve the accuracy and credibility of the evaluation, which plays a role in improving the accuracy and usage effect of the Chinese composition correction method and system.

[0034] Second, the present invention is provided with a real-time feedback and interaction module and a personalized guidance module in the Chinese composition correction system. By using the real-time feedback and interaction platform in the real-time feedback and interaction module, students and teachers can conveniently conduct feedback and interactive communication through the real-time feedback and interaction platform. By using the personalized model and dynamic adjustment in the personalized guidance module, the learning history, writing characteristics, and common mistakes of each student can be recorded through the personalized model and dynamic adjustment, and according to the progress of the students, the evaluation criteria and feedback content of the model can be dynamically adjusted. As a result, students can immediately obtain the correction results and improvement suggestions, adjust and improve their writing in a timely manner, and the system can also provide personalized guidance according to the specific situation of each student to improve the quality of writing teaching, which plays a role in further improving the accuracy and usage effect of the Chinese composition correction method and system. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a system module diagram of the Chinese composition correction system of the present invention;

[0036] Figure 2 It is a system module diagram of the data annotation unit of the present invention;

[0037] Figure 3 It is a system module diagram of the data preprocessing unit of the present invention;

[0038] Figure 4 It is a system module diagram of the language feature unit of the present invention;

[0039] Figure 5 System module diagram of the content feature unit of the present invention;

[0040] Figure 6 System module diagram of the emotion feature unit of the present invention;

[0041] Figure 7 System module diagram of the model optimization unit of the present invention;

[0042] Figure 8 System module diagram of the model integration unit of the present invention;

[0043] Figure 9 Flow chart tree diagram of the correction method of the Chinese composition correction system based on intelligent algorithms of the present invention. Specific implementation manners

[0044] 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.

[0045] Embodiment 1

[0046] Please refer to Figures 1-9 , including a Chinese composition correction system, the Chinese composition correction system includes a data annotation and preprocessing module, a feature extraction module, a multimodal learning module, a model optimization and integration module, a real-time feedback and interaction module, a personalized guidance module, and a security and privacy module. The multimodal learning module includes the combination of image and text and the combination of audio and text. Based on the combination of image and text, for compositions involving illustrations or pictures, image recognition technology can be introduced to evaluate the combination degree of the picture and the text. Based on the combination of audio and text, for compositions of recitation or oral expression, speech recognition technology can be introduced to evaluate the naturalness of the speech expression and the emotional expression. By setting the multimodal learning module of the Chinese composition correction system, the combination of image and text and the combination of audio and text of the multimodal learning module can be used to conveniently introduce image recognition technology for compositions involving illustrations or pictures through the combination of image and text to evaluate the combination degree of the picture and the text, and for compositions of recitation or oral expression through the combination of audio and text, speech recognition technology can be introduced to evaluate the naturalness of the speech expression and the emotional expression, which plays a role in improving the accuracy and use effect of the Chinese composition correction method and system.

[0047] The data annotation and preprocessing module includes a data annotation unit and a data preprocessing unit. Based on the data annotation unit and the data preprocessing unit, the data collected by the Chinese composition correction system can be annotated and preprocessed. Through the data annotation unit, multi-dimensional annotation of the data can be carried out and experts can be invited for annotation. Also, through the data preprocessing unit, multi-faceted preprocessing of the data can be carried out to facilitate the use of the Chinese composition correction method and system to correct Chinese compositions.

[0048] The data annotation unit includes multi-dimensional annotation and expert annotation. Based on multi-dimensional annotation, in addition to traditional sentiment analysis and part-of-speech tagging, multi-dimensional annotation of the composition can be increased. Based on expert annotation, Chinese teachers and literature experts can be invited to annotate some compositions in detail to provide high-quality annotation data. Through multi-dimensional annotation and expert annotation, it is convenient to disassemble and analyze the composition, thereby improving the accuracy of the Chinese composition correction method and system.

[0049] The data preprocessing unit includes context awareness, sentiment analysis, and splicing processing. Based on context awareness, the word segmentation and part-of-speech tagging technology with context awareness can be introduced to improve the understanding and processing ability of specific fields. Based on sentiment analysis, more advanced sentiment analysis algorithms are used to improve the accuracy of sentiment recognition. Based on splicing processing, for long texts, text fragment splicing technology can be used to ensure that the model can process complete compositions. By using the context awareness, sentiment analysis, and splicing processing of the data preprocessing unit, the data can be accurately preprocessed, processed from multiple aspects, and thus the data processing ability of the Chinese composition correction method and system can be improved.

[0050] The specific implementation manner of this embodiment is as follows: When in use, when it is necessary to use the Chinese composition correction system based on intelligent algorithms to correct Chinese compositions, first use the combination of image and text and the combination of audio and text of the multi-modal learning module. It is convenient to introduce image recognition technology through the combination of image and text for compositions involving illustrations or pictures to evaluate the degree of combination of pictures and text. And through the combination of audio and text, for compositions of recitation or oral expression, speech recognition technology can be introduced to evaluate the naturalness of speech expression and emotional expression. Through the data annotation unit, multi-dimensional annotation of the data can be carried out and experts can be invited for annotation. Also, through the data preprocessing unit, multi-faceted preprocessing of the data can be carried out to facilitate the use of the Chinese composition correction method and system to correct Chinese compositions. Through multi-dimensional annotation and expert annotation, it is convenient to disassemble and analyze the composition, thereby improving the accuracy of the Chinese composition correction method and system. By using the context awareness, sentiment analysis, and splicing processing of the data preprocessing unit, the data can be accurately preprocessed, processed from multiple aspects, and thus the data processing ability of the Chinese composition correction method and system can be improved.

[0051] Embodiment Two

[0052] Please refer to Figures 1-9 . The model optimization and integration module includes a model optimization unit and a model integration unit. The model optimization unit includes a deep learning model and multi-task learning. Based on the deep learning model, the text understanding and generation capabilities are improved. Based on multi-task learning, the overall performance of the model is enhanced. The model integration unit includes ensemble learning and an adaptive model. Based on ensemble learning, the prediction results of multiple models can be combined to improve the accuracy and stability of the overall score. Based on the adaptive model, the weights and parameters of the model can be dynamically adjusted according to the student's historical performance and characteristics to achieve personalized evaluation. By setting up the model optimization and integration module of the Chinese composition correction system and using the model optimization unit and model integration unit of the module, the deep learning model and multi-task learning of the model optimization unit can be used to conveniently improve the text understanding and generation capabilities with the deep learning model, and the overall performance of the model can be improved with multi-task learning. The ensemble learning and adaptive model of the model integration unit can use ensemble learning to combine the prediction results of multiple models to improve the accuracy and stability of the overall score, and use the adaptive model to dynamically adjust the weights and parameters of the model according to the student's historical performance and characteristics to achieve personalized evaluation, so as to train an accurate model. Composition evaluation has a certain degree of subjectivity, and the model needs to be continuously optimized to improve the accuracy and credibility of the evaluation, which plays a role in improving the accuracy and usage effect of the Chinese composition correction method and system.

[0053] The feature extraction module includes a language feature unit, a content feature unit, and an emotion feature unit. Based on the language feature unit, content feature unit, and emotion feature unit of the feature extraction module, the features of Chinese compositions can be extracted. The language feature unit can be used to extract the language features in the composition, the content feature unit can be used to extract the content features in the composition, and the emotion feature unit can be used to extract the emotion features in the composition.

[0054] The language feature unit includes advanced vocabulary recognition, syntactic analysis, and text coherence. Based on advanced vocabulary recognition, a word vector model can be used to identify the use of advanced vocabulary and idioms. Based on syntactic analysis, dependency syntactic analysis can be introduced to more accurately identify the structure and logical relationships of sentences. Based on text coherence, a text coherence model can be used to evaluate whether the logical relationships and transitions between sentences are natural. The use of advanced vocabulary and idioms is identified through the word vector model, the structure and logical relationships of sentences are identified using syntactic analysis, and whether the logical relationships and transitions between sentences are natural is evaluated using text coherence.

[0055] The content feature unit includes topic relevance, content depth, and innovation. Based on topic relevance, topic models and keyword extraction techniques can be used to evaluate whether a composition closely adheres to the topic. Based on content depth, semantic similarity analysis can be used to evaluate the depth of exploration of the topic content in the composition. Based on innovation, a benchmark for generating innovative content using a generative adversarial network can be used to evaluate the innovation of the composition. Whether the composition closely adheres to the topic is evaluated through topic relevance, the depth of exploration of the topic content in the composition is evaluated using content depth, and the innovation of the composition is evaluated using innovation performance.

[0056] The emotional feature unit includes emotional clues and emotional fluctuations. Based on emotional clues, emotional clues can be identified and analyzed. Based on emotional fluctuations, the reasonableness of the emotional fluctuations throughout the text can be evaluated. The emotional clues in the composition are identified and analyzed through emotional clues, and the emotional fluctuations throughout the text are evaluated using emotional fluctuations.

[0057] The specific implementation method of this embodiment is as follows: When in use, when it is necessary to correct a Chinese composition using this Chinese composition correction system based on intelligent algorithms, first use the model optimization unit and model integration unit of the model optimization and integration module. Through the deep learning model and multi-task learning of the model optimization unit, it is convenient to use the deep learning model to improve text understanding and generation capabilities, and use multi-task learning to improve the comprehensive performance of the model. The integrated learning and adaptive model of the model integration unit can use integrated learning to combine the prediction results of multiple models to improve the accuracy and stability of the overall score, and use the adaptive model to dynamically adjust the weights and parameters of the model according to the historical performance and characteristics of the students to achieve personalized evaluation, so as to be able to train an accurate model. Composition evaluation has a certain degree of subjectivity, and the model needs to be continuously optimized to improve the accuracy and credibility of the evaluation. The language feature unit can be used to extract the language features in the composition, the content feature unit can be used to extract the content features in the composition, and the emotional feature unit can be used to extract the emotional features in the composition. Advanced vocabulary and idiom usage are identified through the word vector model, the structure and logical relationship of sentences are identified using syntactic analysis, and the logical relationship and transition between sentences are evaluated using text coherence. Whether the composition closely adheres to the topic is evaluated through topic relevance, the depth of exploration of the topic content in the composition is evaluated using content depth, and the innovation of the composition is evaluated using innovation performance. The emotional clues in the composition are identified and analyzed through emotional clues, and the emotional fluctuations throughout the text are evaluated using emotional fluctuations.

[0058] Embodiment III

[0059] Please refer to Figures 1-9, the real-time feedback and interaction module includes real-time feedback and an interaction platform. Based on the real-time feedback and interaction platform, it is convenient for students and teachers to conduct feedback and interaction. The personalized guidance module includes a personalized model and dynamic adjustment. Based on the personalized model and dynamic adjustment, it can record the learning history, writing characteristics, and common mistakes of each student, and according to the progress of the students, dynamically adjust the evaluation criteria and feedback content of the model. By setting up the real-time feedback and interaction module and the personalized guidance module of the Chinese composition correction system, the real-time feedback and interaction platform of the real-time feedback and interaction module can be utilized. Through the real-time feedback and interaction platform, it is convenient for students and teachers to conduct feedback and interaction. By using the personalized model and dynamic adjustment of the personalized guidance module, it can record the learning history, writing characteristics, and common mistakes of each student through the personalized model and dynamic adjustment, and according to the progress of the students, dynamically adjust the evaluation criteria and feedback content of the model. As a result, students can immediately obtain the correction results and improvement suggestions, timely adjust and improve their writing, and the system can also provide personalized guidance according to the specific situation of each student, improving the quality of writing teaching, and further improving the accuracy and usage effect of the Chinese composition correction method and system.

[0060] The security and privacy module includes data protection and transparency. Based on data protection, it can ensure the security of students' data, comply with relevant laws and regulations, and protect students' privacy. Based on transparency, it can provide a transparent evaluation process and results, enabling students and teachers to understand the working principle of the model. By using the data protection and transparency of the security and privacy module, the data security of the Chinese composition correction system can be improved, thereby protecting the personal privacy of students' compositions.

[0061] The specific implementation method of this embodiment is as follows: When in use, when it is necessary to correct a Chinese composition using the Chinese composition correction system based on intelligent algorithms, first utilize the real-time feedback and interaction platform of the real-time feedback and interaction module. Through the real-time feedback and interaction platform, it is convenient for students and teachers to conduct feedback and interaction. By using the personalized model and dynamic adjustment of the personalized guidance module, it can record the learning history, writing characteristics, and common mistakes of each student through the personalized model and dynamic adjustment, and according to the progress of the students, dynamically adjust the evaluation criteria and feedback content of the model. As a result, students can immediately obtain the correction results and improvement suggestions, timely adjust and improve their writing, and the system can also provide personalized guidance according to the specific situation of each student, improving the quality of writing teaching. By using the data protection and transparency of the security and privacy module, the data security of the Chinese composition correction system can be improved, thereby protecting the personal privacy of students' compositions.

[0062] A correction method for a Chinese composition correction system based on intelligent algorithms includes the following steps:

[0063] S1: Define the system requirements and goals, and collect user requirements;

[0064] S2: Collect and annotate high-quality composition data to establish a dataset;

[0065] S3: Select a suitable model, train and optimize the model;

[0066] S4: Develop the front-end and back-end systems to implement data processing, model deployment, and user interface;

[0067] S5: Conduct system testing to evaluate the model's performance and user experience;

[0068] S6: Deploy the system and perform iterative optimization based on user feedback.

[0069] The working principle of the present invention is as follows: When in use, when it is necessary to correct a Chinese composition using the Chinese composition correction system based on intelligent algorithms, first, the image and text combination and audio and text combination of the multi-modal learning module are utilized. For a composition involving illustrations or pictures, image recognition technology can be introduced through the combination of image and text to evaluate the degree of combination of the picture and the text. For a composition of recitation or oral expression, speech recognition technology can be introduced through the combination of audio and text to evaluate the naturalness of speech expression and emotional expression. The data annotation unit can perform multi-dimensional annotation on the data and invite experts for annotation. The data preprocessing unit can also preprocess the data in various aspects to facilitate the correction of Chinese compositions using this Chinese composition correction method and system. Through multi-dimensional annotation and expert annotation, the composition can be disassembled and analyzed conveniently, thereby improving the accuracy of this Chinese composition correction method and system. By using the context awareness, sentiment analysis, and splicing processing of the data preprocessing unit, the data can be preprocessed accurately, and the data can be processed from multiple aspects, thereby improving the data processing ability of this Chinese composition correction method and system. When in use, when it is necessary to correct a Chinese composition using the Chinese composition correction system based on intelligent algorithms, first, the model optimization unit and the model integration unit of the model optimization and integration module are utilized. Through the deep learning model and multi-task learning of the model optimization unit, the text understanding and generation ability can be improved conveniently using the deep learning model, and the comprehensive performance of the model can be improved using multi-task learning. The integrated learning and adaptive model of the model integration unit can combine the prediction results of multiple models using integrated learning to improve the accuracy and stability of the overall score, and the weights and parameters of the model can be dynamically adjusted according to the historical performance and characteristics of the students using the adaptive model to achieve personalized evaluation, thereby enabling an accurate model to be trained. Composition evaluation has a certain degree of subjectivity, and the model needs to be continuously optimized to improve the accuracy and credibility of the evaluation. The language feature unit can extract the language features in the composition, the content feature unit can extract the content features in the composition, and the emotion feature unit can extract the emotion features in the composition. Advanced vocabulary and idiom usage are identified through the word vector model, the structure and logical relationship of sentences are identified using syntactic analysis, and the logical relationship and transition between sentences are evaluated using text coherence to determine whether they are natural. The relevance of the theme is evaluated to determine whether the composition closely adheres to the topic, and the depth of content is evaluated to determine the depth of exploration of the topic content in the composition. The innovation performance is also evaluated to determine the innovation of the composition. The emotional clues in the composition are identified and analyzed through the emotional clues, and the emotional fluctuations throughout the text are evaluated using the emotional fluctuations. When in use, when it is necessary to correct a Chinese composition using the Chinese composition correction system based on intelligent algorithms, first, the real-time feedback and interaction platform of the real-time feedback and interaction module is utilized. Through the real-time feedback and interaction platform, students and teachers can conveniently conduct feedback and interactive communication, and the personalized model and dynamic adjustment of the personalized guidance module are utilized.It can record the learning history, writing characteristics, and common mistakes of each student through a personalized model and dynamic adjustment, and dynamically adjust the evaluation criteria and feedback content of the model according to the progress of the students, so that students can immediately obtain the correction results and improvement suggestions, adjust and improve their writing in a timely manner, and enable the system to provide personalized guidance according to the specific situation of each student, improving the quality of writing teaching. By using the data protection and transparency of the security and privacy module, the data security of this Chinese composition correction system can be improved, thereby protecting the personal privacy of students' compositions.

[0070] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0071] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art 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 Chinese composition correction system based on intelligent algorithms, including a Chinese composition correction system, characterized in that: The Chinese composition correction system includes a data annotation and preprocessing module, a feature extraction module, a multi-modal learning module, a model optimization and integration module, a real-time feedback and interaction module, a personalized guidance module, and a security and privacy module. The multi-modal learning module includes the combination of image and text and the combination of audio and text. Based on the combination of image and text, for compositions involving illustrations or pictures, image recognition technology can be introduced to evaluate the degree of combination of pictures and text. Based on the combination of audio and text, for compositions of recitation or oral expression, speech recognition technology can be introduced to evaluate the naturalness of speech expression and emotional expression; The model optimization and integration module includes a model optimization unit and a model integration unit. The model optimization unit includes a deep learning model and multi-task learning. Based on the deep learning model, the text understanding and generation ability can be improved. Based on the multi-task learning, the comprehensive performance of the model can be improved. The model integration unit includes ensemble learning and an adaptive model. Based on the ensemble learning, the prediction results of multiple models can be combined to improve the accuracy and stability of the overall score. Based on the adaptive model, the weights and parameters of the model can be dynamically adjusted according to the historical performance and characteristics of students to achieve personalized evaluation; The real-time feedback and interaction module includes real-time feedback and an interaction platform. Based on the real-time feedback and the interaction platform, it is convenient for students and teachers to conduct feedback and interactive communication. The personalized guidance module includes a personalized model and dynamic adjustment. Based on the personalized model and dynamic adjustment, the learning history, writing characteristics, and common mistakes of each student can be recorded, and the evaluation criteria and feedback content of the model can be dynamically adjusted according to the progress of the students.

2. The Chinese composition marking system based on intelligent algorithm according to claim 1, characterized in that: The data annotation and preprocessing module includes a data annotation unit and a data preprocessing unit. Based on the data annotation unit and the data preprocessing unit, the data collected by the Chinese composition correction system can be annotated and preprocessed.

3. The Chinese composition correction system based on intelligent algorithm according to claim 2, characterized in that: The data annotation unit includes multi-dimensional annotation and expert annotation. Based on the multi-dimensional annotation, in addition to traditional sentiment analysis and part-of-speech tagging, multi-dimensional annotation of compositions can be increased. Based on the expert annotation, Chinese language teachers and literary experts can be invited to conduct detailed annotation on some compositions to provide high-quality annotation data.

4. The Chinese composition correction system based on intelligent algorithm according to claim 2, characterized in that: The data preprocessing unit includes context awareness, sentiment analysis, and splicing processing. Based on the context awareness, context-aware word segmentation and part-of-speech tagging technology can be introduced to improve the understanding and processing ability of specific domains. Based on the sentiment analysis, more advanced sentiment analysis algorithms are used to improve the accuracy of sentiment recognition. Based on the splicing processing, for long texts, text segment splicing technology can be used to ensure that the model can process complete compositions.

5. The Chinese composition correction system based on intelligent algorithm according to claim 1, characterized in that: The feature extraction module includes a language feature unit, a content feature unit, and an emotion feature unit. Based on the language feature unit, content feature unit, and emotion feature unit of the feature extraction module, the features of Chinese compositions can be extracted.

6. The Chinese composition marking system based on intelligent algorithm according to claim 5, characterized in that: The language feature unit includes advanced vocabulary recognition, syntactic analysis, and text coherence. Based on the advanced vocabulary recognition, a word vector model can be used to identify the use of advanced vocabulary and idioms. Based on the syntactic analysis, dependency syntactic analysis can be introduced to more precisely identify the structure and logical relationships of sentences. Based on the text coherence, a text coherence model can be used to evaluate whether the logical relationships and transitions between sentences are natural.

7. An intelligent algorithm-based Chinese composition correction system according to claim 5, characterized in that: The content feature unit includes topic relevance, content depth, and innovation. Based on the topic relevance, a topic model and keyword extraction technology can be used to evaluate whether the composition closely adheres to the topic. Based on the content depth, semantic similarity analysis can be used to evaluate the depth of exploration of the topic content in the composition. Based on the innovation, a generative adversarial network can be used to generate a benchmark for innovative content to evaluate the innovation of the composition.

8. An intelligent algorithm-based Chinese composition marking system according to claim 5, characterized in that: The emotional feature unit includes emotional cues and emotional fluctuations. Based on the emotional cues, emotional cues can be identified and analyzed. Based on the emotional fluctuations, the reasonableness of the emotional fluctuations throughout the text can be evaluated.

9. The Chinese composition correction system based on intelligent algorithm according to claim 1, characterized in that: The security and privacy module includes data protection and transparency. Based on the data protection, the security of students' data can be ensured, relevant laws and regulations can be complied with, and students' privacy can be protected. Based on the transparency, a transparent evaluation process and results can be provided to enable students and teachers to understand the working principle of the model.

10. A marking method for a Chinese composition marking system based on an intelligent algorithm according to any one of claims 1-9, characterized in that: Specifically, it includes the following steps: S1: Define the requirements and goals of the system and collect user requirements; S2: Collect and annotate high-quality composition data to establish a dataset; S3: Select a suitable model, train and optimize the model; S4: Develop the front-end and back-end systems to implement data processing, model deployment, and user interface; S5: Conduct system testing to evaluate the performance of the model and the user experience; S6: Deploy the system and perform iterative optimization based on user feedback.

Images