English composition evaluation method and device

Abstract

The invention discloses an English composition evaluation method and device, belongs to the technical field of artificial intelligence in the education industry, and aims to solve the problem that a scoring result of an existing English composition automatic scoring system is greatly different from a manual scoring result. Comprising the following steps: obtaining a to-be-scored composition; semantic analysis is carried out on the to-be-scored composition, wherein the semantic analysis comprises high-level vocabulary analysis, semantic coherence analysis, logic hierarchy analysis, sentence pattern analysis and theme consistency analysis; obtaining a high-level vocabulary score, a semantic coherence score, a logic hierarchy score, a theme consistency score and a high-level sentence pattern score; based on a preset weight coefficient, performing weighted summation on the high-level vocabulary score, the semantic coherence score, the logic hierarchy score, the high-level sentence pattern score and the theme consistency score to obtain a final score of the to-be-scored composition; according to the invention, the accuracy and reliability of English composition scoring can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology in the education industry, and in particular to an English composition evaluation method and device. Background Art

[0002] English writing ability is not only an important indicator to measure students' comprehensive language application ability, but also an indispensable skill in academic exchanges, career development and other fields. Therefore, English composition evaluation occupies a pivotal position in the education system. Traditional English composition evaluation mainly relies on manual scoring by teachers. Although this method can provide relatively accurate scoring results, it also has some inherent problems. First, manual scoring is inefficient, especially in large-scale examinations or homework corrections. Teachers need to invest a lot of time and energy, which not only increases the workload of teachers, but also may lead to delays in scoring results. Secondly, manual scoring is subjective, and different teachers may give different scores to the same essay. This difference may be due to factors such as personal experience, educational background, and cultural differences, thus affecting the fairness and consistency of scoring.

[0003] In recent years, although some algorithm-based automatic scoring systems have emerged, these systems attempt to simulate the manual scoring process through natural language processing and machine learning technology, but they are still insufficient in capturing the deep semantics and logical structure of the composition. Automatic scoring systems often focus on checking grammar and spelling, while ignoring key factors such as the content quality, logical coherence and innovation of the composition, resulting in a large gap between the scoring results and the manual scoring results.

[0004] The disclosure of the above background technology content is only used to assist in understanding the concept and technical solution of the present invention. It does not necessarily belong to the prior art of this patent application. In the absence of clear evidence that the above content has been disclosed on the filing date of this patent application, the above background technology should not be used to evaluate the novelty and creativity of the present application. Summary of the invention

[0005] The present application provides an English composition evaluation method and device, which are used to solve the problem that there is a large gap between the scoring results of the existing English composition automatic scoring system and the manual scoring results.

[0006] To achieve the above objectives, the present application discloses the following technical solutions:

[0007] In a first aspect, the present application provides an English composition evaluation method, comprising the following steps:

[0008] Get the essays to be graded;

[0009] Conduct semantic analysis on the essays to be graded, including advanced vocabulary analysis, semantic coherence analysis, logical hierarchy analysis, sentence structure analysis, and theme consistency analysis; obtain advanced vocabulary scores, semantic coherence scores, logical hierarchy scores, theme consistency scores, and advanced sentence structure scores;

[0010] Based on the preset weight coefficients, the advanced vocabulary score, semantic coherence score, logical hierarchy score, advanced sentence structure score and topic consistency score are weighted and summed to obtain the final score of the essay to be scored.

[0011] In the embodiment of the present application, a multi-dimensional semantic analysis is adopted. The method first performs a high-level vocabulary analysis to evaluate the vocabulary level used in the composition, especially the frequency and diversity of high-level vocabulary. This analysis can help the scoring system identify the writer's language ability and reflect his or her writing level, thereby improving the accuracy of the scoring.

[0012] Next, the semantic coherence analysis focuses on the logical relationships and coherence between paragraphs and sentences, and evaluates the overall coherence of the essay by analyzing cohesive words and thematic continuation. This analysis helps identify logical flow, thereby improving the reliability of scoring and ensuring that the scoring system can understand the author's main ideas.

[0013] Logical hierarchical analysis further evaluates the organization and hierarchical structure of the essay to ensure that the logical relationship between paragraphs and sentences is reasonable. Through this analysis, the scoring system can accurately identify high-quality arguments and structures, enhancing the objectivity of the scoring.

[0014] Sentence analysis focuses on the diversity and complexity of sentences and evaluates the sentence structures used. This analysis helps the scoring system identify the writer's expressiveness and language richness, further improving the accuracy of the scoring.

[0015] Finally, thematic consistency analysis is used to ensure that the essay topic is unified and consistent and avoids deviations from the topic. Through this analysis, the scoring system can better assess the degree to which the essay grasps the topic, thereby improving the reliability of the overall scoring.

[0016] In summary, by introducing multi-dimensional semantic analysis, the accuracy and reliability of English composition scoring can be effectively improved, the shortcomings of the existing automatic scoring system in deep semantic understanding and logical structure evaluation can be solved, and the error between the scoring results of the existing English composition automatic scoring system and the manual scoring results can be narrowed.

[0017] In some possible implementations of the first aspect, the step of performing advanced vocabulary analysis on the composition to be scored and obtaining an advanced vocabulary score includes:

[0018] Tag each word in the essay to be graded with a part-of-speech tag;

[0019] Create a first matching set, which is an empty set;

[0020] Comparing each word and its corresponding part-of-speech tag with a preset advanced vocabulary list, the preset advanced vocabulary list includes a set of advanced words and their parts of speech and corresponding preset scores;

[0021] When the word and its part of speech match the preset high-level vocabulary list, and the high-level vocabulary does not appear in the first matching set, adding the high-level vocabulary and the preset score of the corresponding high-level vocabulary to the first matching set;

[0022] The preset scores of all advanced vocabulary in the first matching set are accumulated to calculate the total score of the advanced vocabulary of the composition to be graded.

[0023] Among them, by tagging the part of speech of each word and comparing it with the preset advanced vocabulary list, the advanced vocabulary used in the composition can be accurately identified.

[0024] Using a preset list of advanced vocabulary words ensures that all essays to be graded are evaluated under the same standards. This standardized process reduces the subjective factors that may be introduced by manual scoring and improves the consistency of scoring, which can effectively avoid the problem of scoring differences, especially in large-scale evaluations.

[0025] When adding advanced vocabulary and its corresponding score, it is required that the advanced vocabulary does not appear in the first matching set, which can effectively avoid repeated calculation of the score of the same vocabulary. This mechanism ensures that the contribution of each advanced vocabulary is evaluated separately, making the final advanced vocabulary score more realistic and accurate.

[0026] In some possible implementations of the first aspect, the step of performing sentence analysis on the composition to be scored and obtaining a high-level sentence score includes:

[0027] Get a preset advanced sentence table, which contains multiple predefined advanced sentence types and their corresponding scores;

[0028] Perform syntactic analysis on each sentence in the essay to be graded and identify the sentence structure in the sentence;

[0029] Match the identified sentence patterns with the advanced sentence pattern table, determine whether each sentence pattern belongs to an advanced sentence pattern, and record the occurrence frequency of each advanced sentence pattern;

[0030] According to the frequency of occurrence of each advanced sentence pattern in the essay to be graded and the predefined score in the advanced sentence pattern table, the contribution score of each advanced sentence pattern is calculated; the specific calculation formula is:

[0031] Contribution score of advanced sentence A = ∑(number of occurrences of advanced sentence A × preset score of advanced sentence A × penalty factor) where:

[0032] Advanced sentence pattern A represents any advanced sentence pattern;

[0033] The penalty factor is used to adjust the contribution score of high-level sentences that appear too frequently; the penalty factor calculation formula is:

[0034]

[0035] Among them, k is a constant greater than 0, which is used to control the intensity of the penalty;

[0036] The contribution scores of all advanced sentences are accumulated to obtain the preliminary advanced sentence score of the essay to be graded;

[0037] Get the total number of sentences in the essay to be graded, divide the preliminary advanced sentence score by the total number of sentences to calculate the advanced sentence score; the specific calculation formula is:

[0038]

[0039] Among them, through comprehensive consideration of the frequency of occurrence and contribution scores of advanced sentences, combined with the adjustment of penalty factors, it is ensured that the score is not affected by the frequent occurrence of a single advanced sentence, making the final score more reasonable and objective.

[0040] Dividing the preliminary advanced sentence score by the total number of sentences to calculate the advanced sentence score can more accurately reflect the effectiveness of the use of advanced sentences. Specifically, simply counting the number of advanced sentences may result in longer essays getting higher scores, while ignoring that short and concise essays may also use a wealth of advanced sentences. By combining it with the total number of sentences, it can more truly reflect the effectiveness of writers using advanced sentences in essays of different lengths.

[0041] In some possible implementations of the first aspect, the English composition evaluation method further includes the following steps:

[0042] determining whether the advanced vocabulary score of the composition to be graded is lower than a first preset value;

[0043] When the advanced vocabulary score of the composition to be graded is lower than a first preset value, a first sentence set without advanced vocabulary is selected; the first sentence set includes the first sentence;

[0044] At least a set of second sentences whose semantic similarity with the first sentence reaches a second preset value and whose high-level vocabulary score exceeds the first preset value are selected from the preset model text library;

[0045] calculating a high-level vocabulary score for each sentence in the second set of sentences;

[0046] sorting each sentence in the second sentence set according to the high-level vocabulary score of each sentence in the second sentence set; and generating a first recommendation set according to the sorting result, the first recommendation set at least including the sentence ranked first by the high-level vocabulary score;

[0047] Based on the first recommendation set, modification suggestions are generated.

[0048] Among them, the recommended advanced vocabulary can be used as a reference to help writers accumulate more advanced vocabulary to improve their writing skills.

[0049] In addition, the recommendation process places special emphasis on semantic similarity with the original sentence, ensuring that the recommended sentence is not only rich in vocabulary, but also consistent with the logic and theme of the original text. This approach of focusing on semantic coherence can effectively improve the fluency and readability of the overall composition, avoiding the problem of incoherence or abrupt combination caused by replacing single words or recommending sentences that are irrelevant to the content, thus making the writer's rewriting smoother.

[0050] In some possible implementations of the first aspect, when the final score of the composition to be scored is greater than or equal to a third preset value, the composition to be scored is stored in a preset model essay library.

[0051] In this way, more excellent model essays can be accumulated for reference by subsequent writers, thereby promoting the improvement of overall writing standards and forming a virtuous circle.

[0052] In some possible implementations of the first aspect, the English composition evaluation method further includes the following steps:

[0053] Conduct sentence structure analysis on the second sentence of the essay to be graded to determine whether the second sentence uses advanced sentence structures;

[0054] When the second sentence does not use advanced sentence patterns, sentences whose semantic similarity with the second sentence reaches a fourth preset value and which use advanced sentence patterns are retrieved from the model text library to form a third sentence set;

[0055] Sorting each sentence in the third sentence set according to the semantic similarity, and generating a second recommendation set according to the sorting result, wherein the second recommendation set includes at least the sentence ranked first in the semantic similarity;

[0056] Based on the second recommendation set, modification suggestions are generated.

[0057] This can help writers better understand and apply advanced sentence patterns, thereby improving the language level and expression effect of their writing.

[0058] In a second aspect, the present application embodiment provides an English composition evaluation device, comprising:

[0059] The first acquisition module is used to acquire the composition to be graded;

[0060] The first scoring module is used to conduct semantic analysis on the essays to be scored, including advanced vocabulary analysis, semantic coherence analysis, logical hierarchy analysis, sentence structure analysis and theme consistency analysis; obtaining advanced vocabulary scores, semantic coherence scores, logical hierarchy scores, theme consistency scores and advanced sentence structure scores;

[0061] The scoring synthesis module is used to perform weighted summation of the advanced vocabulary score, semantic coherence score, logical hierarchy score, advanced sentence structure score and topic consistency score based on preset weight coefficients to obtain the final score of the composition to be scored.

[0062] In some possible implementations of the second aspect, the first scoring module includes:

[0063] The first tagging unit is used to tag the part of speech of each word in the composition to be scored;

[0064] A first creating unit, used to create a first matching set, where the first matching set is an empty set;

[0065] A first comparison unit is used to compare each word and its corresponding part-of-speech tag with a preset high-level vocabulary list, wherein the preset high-level vocabulary list includes a group of high-level words and their parts of speech and corresponding preset scores;

[0066] A first adding unit, configured to add the high-level word and the preset score of the corresponding high-level word to the first matching set when the word and its part of speech match a preset high-level word list and the high-level word does not appear in the first matching set;

[0067] The first calculation unit is used to accumulate the preset scores of all the advanced vocabulary in the first matching set to calculate the total score of the advanced vocabulary of the composition to be scored.

[0068] In some possible implementations of the second aspect, the first scoring module further includes:

[0069] A second acquisition unit is used to acquire a preset high-level sentence table, where the high-level sentence table includes a plurality of predefined high-level sentence types and their corresponding scores;

[0070] The first recognition unit is used to perform syntactic analysis on each sentence in the composition to be graded and recognize the sentence structure in the sentence;

[0071] A first matching unit is used to match the identified sentence pattern with the high-level sentence pattern table, determine whether each sentence pattern belongs to a high-level sentence pattern, and record the occurrence frequency of each high-level sentence pattern;

[0072] The second calculation unit is used to calculate the contribution score of each advanced sentence pattern according to the frequency of occurrence of each advanced sentence pattern in the composition to be scored and the predefined score in the advanced sentence pattern table; the specific calculation formula is:

[0073] Contribution score of advanced sentence A = ∑(number of occurrences of advanced sentence A × preset score of advanced sentence A × penalty factor) where:

[0074] Advanced sentence pattern A represents any advanced sentence pattern;

[0075] The penalty factor is used to adjust the contribution score of high-level sentences that appear too frequently. The penalty factor calculation formula is:

[0076]

[0077] Among them, k is a constant greater than 0, which is used to control the intensity of the penalty;

[0078] The third calculation unit is used to accumulate the contribution scores of all advanced sentences to obtain a preliminary advanced sentence score of the composition to be scored;

[0079] The fourth calculation unit is used to obtain the total number of sentences in the composition to be scored, and divide the preliminary advanced sentence score by the total number of sentences to calculate the advanced sentence score; the specific calculation formula is:

[0080]

[0081] In some possible implementations of the second aspect, the English composition evaluation device further includes:

[0082] A first judgment module is used to judge whether the advanced vocabulary score of the composition to be graded is lower than a first preset value;

[0083] A first screening module is used to screen out a first sentence set without advanced vocabulary when the advanced vocabulary score of the composition to be graded is lower than a first preset value; the first sentence set includes the first sentence;

[0084] A first matching module is used to select a set of second sentences from at least a preset model text library, the semantic similarity with the first sentence reaching a second preset value and the high-level vocabulary score exceeding the first preset value;

[0085] A first calculation module, configured to calculate a high-level vocabulary score for each sentence in the second sentence set;

[0086] A first sorting module is used to sort each sentence in the second sentence set according to the high-level vocabulary score of each sentence in the second sentence set; and generate a first recommendation set according to the sorting result, the first recommendation set at least including the sentence ranked first by the high-level vocabulary score;

[0087] The first suggestion generating module is used to generate a modification suggestion based on the first recommendation set.

[0088] In some possible implementations of the second aspect, the English composition evaluation device further includes a first adding module, which is used to store the composition to be scored in a preset model essay library when the final score of the composition to be scored is greater than or equal to a third preset value.

[0089] In some possible implementations of the second aspect, the English composition evaluation device further includes:

[0090] The second judgment module is used to perform sentence analysis on the second sentence of the composition to be graded, and judge whether the second sentence uses a high-level sentence pattern;

[0091] The second screening module is used to retrieve sentences from the model text library whose semantic similarity with the second sentence reaches a fourth preset value and which use advanced sentence patterns when the second sentence does not use advanced sentence patterns, to form a third sentence set;

[0092] A second sorting module is used to sort each sentence in the third sentence set according to semantic similarity, and generate a second recommendation set according to the sorting result, wherein the second recommendation set includes at least the sentence ranked first in semantic similarity;

[0093] The second suggestion generating module is used to generate modification suggestions based on the second recommendation set.

[0094] In a third aspect, an embodiment of the present application provides a terminal device, including a processing device and a memory;

[0095] The memory is used to store one or more program instructions;

[0096] A processing device, used to run one or more program instructions to execute the method of any technical solution of the first aspect.

[0097] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processing device, a method for implementing any technical solution of the first aspect is implemented.

[0098] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which is a method according to any one of the technical solutions of the first aspect when executed by a processing device.

[0099] Among them, the technical effects brought about by any design method in the second to fifth aspects can refer to the technical effects brought about by different design methods in the first aspect, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0101] Figure 1 A schematic diagram of a method for evaluating English composition provided in some embodiments of the present application Figure 1 ;

[0102] Figure 2 A schematic diagram of a method for evaluating English composition provided in some embodiments of the present application Figure 2 ;

[0103] Figure 3 A schematic diagram of a method for evaluating English composition provided in some embodiments of the present application Figure 3 ;

[0104] Figure 4 A schematic diagram of the structure of an English composition evaluation device provided in some embodiments of the present application;

[0105] Figure 5 for Figure 4 The structural diagram of the first scoring module in the English composition evaluation device shown;

[0106] Figure 6 It is a structural diagram of a terminal device suitable for implementing some embodiments of the present application. DETAILED DESCRIPTION

[0107] Specific embodiments of the present invention will now be mentioned in detail. Although the present invention is described in conjunction with these specific embodiments, it should be appreciated that it is not intended to limit the present invention to these specific embodiments. On the contrary, these embodiments are intended to cover substitutions, changes or equivalent embodiments that may be included in the spirit and scope of the invention defined by the claims. In the following description, a large number of specific details are set forth in order to provide a comprehensive understanding of the present invention. The present invention may be implemented without some or all of these specific details.

[0108] When used in conjunction with "including," "methods comprising," or similar language in this specification and the appended claims, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0109] English writing ability is not only an important indicator to measure students' comprehensive language application ability, but also an indispensable skill in academic exchanges, career development and other fields. Therefore, English composition evaluation occupies a pivotal position in the education system. Traditional English composition evaluation mainly relies on manual scoring by teachers. Although this method can provide relatively accurate scoring results, it also has some inherent problems. First, manual scoring is inefficient, especially in large-scale examinations or homework corrections. Teachers need to invest a lot of time and energy, which not only increases the workload of teachers, but also may lead to delays in scoring results. Secondly, manual scoring is subjective, and different teachers may give different scores to the same essay. This difference may be due to factors such as personal experience, educational background, and cultural differences, thus affecting the fairness and consistency of scoring.

[0110] In recent years, although some algorithm-based automatic scoring systems have emerged, these systems attempt to simulate the manual scoring process through natural language processing and machine learning technology, but they are still insufficient in capturing the deep semantics and logical structure of the composition. Automatic scoring systems often focus on checking grammar and spelling, while ignoring key factors such as the content quality, logical coherence and innovation of the composition, resulting in a large gap between the scoring results and the manual scoring results.

[0111] In response to the above technical problems, the overall idea of ​​the technical solution provided by the present application is as follows: The embodiment of the present application provides an English composition evaluation method, comprising the following steps: obtaining a composition to be scored; performing semantic analysis on the composition to be scored, the semantic analysis including advanced vocabulary analysis, semantic coherence analysis, logical hierarchy analysis, sentence analysis and theme consistency analysis; obtaining an advanced vocabulary score, a semantic coherence score, a logical hierarchy score, a theme consistency score and an advanced sentence score; based on a preset weight coefficient, performing weighted summation on the advanced vocabulary score, the semantic coherence score, the logical hierarchy score, the advanced sentence score and the theme consistency score to obtain a final score for the composition to be scored.

[0112] The method uses a multi-dimensional semantic analysis. The method first conducts high-level vocabulary analysis to evaluate the vocabulary level used in the composition, especially the frequency and diversity of high-level vocabulary. This analysis can help the scoring system identify the writer's language ability and reflect his or her writing level, thereby improving the accuracy of the scoring. Next, the semantic coherence analysis focuses on the logical relationship and coherence between paragraphs and sentences, and evaluates the overall coherence of the composition by analyzing cohesive words and topic continuation. This analysis helps to identify logical fluency, thereby improving the reliability of the scoring and ensuring that the scoring system can understand the author's main ideas. Logical hierarchy analysis further evaluates the argument organization and hierarchy of the composition to ensure that the logical relationship between paragraphs and sentences is reasonable. Through this analysis, the scoring system can accurately identify high-quality arguments and structures and enhance the objectivity of the scoring. Sentence analysis focuses on the diversity and complexity of sentences and evaluates the sentence structure used. This analysis helps the scoring system identify the writer's expression ability and language richness, further improving the accuracy of the scoring. Finally, the theme consistency analysis is used to ensure the unity and consistency of the composition theme and avoid deviation from the theme. Through this analysis, the scoring system can better evaluate the degree of grasp of the theme of the article, thereby improving the reliability of the overall scoring.

[0113] In summary, by introducing multi-dimensional semantic analysis, the accuracy and reliability of English composition scoring can be effectively improved, the shortcomings of the existing automatic scoring system in deep semantic understanding and logical structure evaluation can be solved, and the error between the scoring results of the existing English composition automatic scoring system and the manual scoring results can be narrowed.

[0114] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0115] See also Figure 1 The present application embodiment provides an English composition evaluation method, comprising the following steps:

[0116] S101: Obtaining the composition to be graded; specifically, the composition to be graded may be obtained by the user uploading a file, and common file formats such as .txt, .docx, and .pdf are supported. Of course, the present application is not limited thereto. In other embodiments, the composition to be graded may be obtained by the user scanning a paper document using OCR (optical character recognition).

[0117] S102: Conduct semantic analysis on the essay to be scored, including advanced vocabulary analysis, semantic coherence analysis, logical hierarchy analysis, sentence structure analysis and theme consistency analysis; obtain advanced vocabulary score, semantic coherence score, logical hierarchy score, theme consistency score and advanced sentence structure score;

[0118] Specifically, advanced vocabulary refers to words or phrases that are used less frequently in a specific language environment, are more semantically rich, and are more precise in expression. These words usually include, but are not limited to: polysyllabic words: such as "incredible", "unprecedented", etc. Vocabulary with a specific domain: such as academic, literary or professional vocabulary, such as "analyze", "synthesize", etc. More expressive words that replace common words: for example, "utilize" instead of "use", "assist" instead of "help".

[0119] Advanced sentences refer to complex and diverse sentence structures that can effectively convey complex ideas and enhance the diversity and depth of expression. These sentences usually include but are not limited to: compound and complex sentences: such as "Although it was raining, we decided to go for a walk." Sentences containing clauses: such as attributive clauses and adverbial clauses: "The book that you lent me was fascinating." Sentences using a variety of conjunctions and transition words: such as "Furthermore", "In contrast to", etc.

[0120] S103: Based on the preset weight coefficient, the advanced vocabulary score, semantic coherence score, logical hierarchy score, advanced sentence structure score and topic consistency score are weighted and summed to obtain the final score of the composition to be scored. The weight coefficient is determined according to actual needs and is not limited here.

[0121] See also Figure 2 In step S102, the steps of performing advanced vocabulary analysis on the composition to be scored and obtaining advanced vocabulary scores include:

[0122] S102a1: Tag each word in the essay to be graded for part of speech;

[0123] S102a2: Create a first matching set, which is an empty set;

[0124] S102a3: Compare each word and its corresponding part-of-speech tag with a preset advanced vocabulary list, where the preset advanced vocabulary list includes a group of advanced vocabulary and its part-of-speech and corresponding preset scores;

[0125] S102a4: When the word and its part of speech match the preset high-level vocabulary list, and the high-level vocabulary does not appear in the first matching set, add the high-level vocabulary and the preset score of the corresponding high-level vocabulary to the first matching set;

[0126] S102a5: Accumulate the preset scores of all advanced vocabulary in the first matching set to calculate the total score of the advanced vocabulary of the composition to be scored.

[0127] Among them, by tagging the part of speech of each word and comparing it with the preset advanced vocabulary list, the advanced vocabulary used in the composition can be accurately identified.

[0128] Using a preset list of advanced vocabulary words ensures that all essays to be graded are evaluated under the same standards. This standardized process reduces the subjective factors that may be introduced by manual scoring and improves the consistency of scoring, which can effectively avoid the problem of scoring differences, especially in large-scale evaluations.

[0129] When adding advanced vocabulary and its corresponding score, it is required that the advanced vocabulary does not appear in the first matching set, which can effectively avoid repeated calculation of the score of the same vocabulary. This mechanism ensures that the contribution of each advanced vocabulary is evaluated separately, making the final advanced vocabulary score more realistic and accurate.

[0130] Preferably, in step S102, a semantic coherence analysis is performed on the composition to be scored, and a method for obtaining a semantic coherence score may be: using BERT (Bidirectional Encoder Representations from Transformers) to obtain a contextual representation of a sentence, and evaluating coherence by calculating the similarity between sentences to obtain a semantic coherence score.

[0131] Exemplarily, the following steps may be included:

[0132] The composition text to be scored is divided into an independent sentence set S = {s1, s2, s3, ..., sn}, where each si is an independent sentence.

[0133] The BERT model is used to encode each sentence in the sentence set S and generate the sentence context representation matrix V, where V[i] represents the vector representation of sentence si.

[0134] The coherence between sentences is evaluated by calculating the cosine similarity between each pair of sentences. The specific steps include:

[0135] For each pair of sentences si and sj, calculate their cosine similarity:

[0136]

[0137] in:

[0138] V[i]: sentence s i The vector representation of ;

[0139] V[j]: sentence sj The vector representation of ;

[0140] ·: represents the dot product operation of vectors;

[0141] V[i] represents the magnitude of vector V[i];

[0142] V[j] represents the magnitude of vector V[j];

[0143] Output similarity score set S_similarity = {similarity(s1,s2),similarity(s1,s3),...,similarity(sn-1,sn)}.

[0144] The similarity score set S_similarity is processed to calculate the semantic coherence score C_total, specifically:

[0145]

[0146] Where n: the total number of sentences in the sentence set S.

[0147] Preferably, in step S102, the logical hierarchy of the essay to be scored can be analyzed by ArgumentMiningTools (logic analysis software) to obtain a logical hierarchy score. Exemplarily, the following steps can be included:

[0148] Use the ArgumentMining tool to parse the text of the essay to be graded and identify the main arguments, supporting arguments, and rebuttal arguments in the text. The tool analyzes sentence structure and semantics to construct an argument map;

[0149] The tool uses natural language processing technology to identify the logical relationship between sentences, including causal relationships, contrast relationships, and illustrative relationships. This process usually includes:

[0150] Syntactic analysis: identifying the subject, predicate, and object of a sentence.

[0151] Semantic analysis: understanding the specific meaning of a sentence and its role in the overall argument;

[0152] Based on the identified arguments and logical relationships, a logical hierarchy is constructed to form a tree-like or network-like structure diagram showing the hierarchy and relationship between the arguments;

[0153] The ArgumentMining tool evaluates the logical hierarchy of the article based on the constructed argument structure and logical relationships, and generates a logical hierarchy score. The scoring criteria may include: the clarity of the argument, the coherence of the logical relationship, and the sufficiency of the supporting argument.

[0154] See also Figure 3 In step S102, the steps of performing sentence analysis on the composition to be scored and obtaining a high-level sentence score include:

[0155] S102b1: Obtain a preset advanced sentence table, where the advanced sentence table includes a plurality of predefined advanced sentence types and their corresponding scores;

[0156] S102b2: Perform syntactic analysis on each sentence in the essay to be graded and identify the sentence structure in the sentence;

[0157] S102b3: matching the identified sentence pattern with the advanced sentence pattern table, determining whether each sentence pattern belongs to an advanced sentence pattern, and recording the occurrence frequency of each advanced sentence pattern;

[0158] S102b4: Calculate the contribution score of each advanced sentence pattern according to the frequency of occurrence of each advanced sentence pattern in the composition to be graded and the predefined score in the advanced sentence pattern table; the specific calculation formula is:

[0159] Contribution score of advanced sentence A = ∑(number of occurrences of advanced sentence A × preset score of advanced sentence A × penalty factor) where:

[0160] Advanced sentence pattern A represents any advanced sentence pattern;

[0161] The penalty factor is used to adjust the contribution score of high-level sentences that appear too frequently; the penalty factor calculation formula is:

[0162]

[0163] Among them, k is a constant greater than 0, which is used to control the intensity of the penalty;

[0164] S102b5: Accumulate the contribution scores of all advanced sentences to obtain a preliminary advanced sentence score of the composition to be graded;

[0165] S102b6: Obtain the total number of sentences in the composition to be graded, and divide the preliminary advanced sentence score by the total number of sentences to calculate the advanced sentence score; the specific calculation formula is:

[0166]

[0167] Among them, through comprehensive consideration of the frequency of occurrence and contribution scores of advanced sentences, combined with the adjustment of penalty factors, it is ensured that the score is not affected by the frequent occurrence of a single advanced sentence, making the final score more reasonable and objective.

[0168] Dividing the preliminary advanced sentence score by the total number of sentences to calculate the advanced sentence score can more accurately reflect the effectiveness of the use of advanced sentences. Specifically, simply counting the number of advanced sentences may result in longer essays getting higher scores, while ignoring that short and concise essays may also use a wealth of advanced sentences. By combining it with the total number of sentences, it can more truly reflect the effectiveness of writers using advanced sentences in essays of different lengths.

[0169] In step S102, the steps of performing a theme consistency analysis on the composition to be scored and obtaining a theme consistency score include:

[0170] Get the scoring criteria;

[0171] Extracting the subject keywords in the scoring criteria to form a subject keyword set;

[0172] Perform keyword matching analysis on the composition to be rated according to the theme keyword set to obtain a keyword matching analysis result;

[0173] Using semantic analysis technology to calculate the semantic similarity between the essay to be rated and the model essay, and generate a semantic similarity score;

[0174] Based on the keyword matching analysis results and semantic similarity scores, the topic consistency score is calculated.

[0175] Preferably, when the final score of the composition to be scored is greater than or equal to the third preset value, the composition to be scored is stored in a preset model essay library.

[0176] In this way, more excellent model essays can be accumulated for reference by subsequent writers, thereby promoting the improvement of overall writing standards and forming a virtuous circle.

[0177] Preferably, the English composition evaluation method further comprises the following steps:

[0178] determining whether the advanced vocabulary score of the composition to be graded is lower than a first preset value;

[0179] When the advanced vocabulary score of the composition to be graded is lower than a first preset value, a first sentence set without advanced vocabulary is selected; the first sentence set includes the first sentence;

[0180] At least a set of second sentences whose semantic similarity with the first sentence reaches a second preset value and whose high-level vocabulary score exceeds the first preset value are selected from the preset model text library;

[0181] calculating a high-level vocabulary score for each sentence in the second set of sentences;

[0182] sorting each sentence in the second sentence set according to the high-level vocabulary score of each sentence in the second sentence set; and generating a first recommendation set according to the sorting result, the first recommendation set at least including the sentence ranked first by the high-level vocabulary score;

[0183] Based on the first recommendation set, modification suggestions are generated.

[0184] Among them, the recommended advanced vocabulary can be used as a reference to help writers accumulate more advanced vocabulary to improve their writing skills.

[0185] In addition, the recommendation process places special emphasis on semantic similarity with the original sentence, ensuring that the recommended sentence is not only rich in vocabulary, but also consistent with the logic and theme of the original text. This approach of focusing on semantic coherence can effectively improve the fluency and readability of the overall composition, avoiding the problem of incoherence or abrupt combination caused by replacing single words or recommending sentences that are irrelevant to the content, thus making the writer's rewriting smoother.

[0186] Preferably, the English composition evaluation method further comprises the following steps:

[0187] Conduct sentence structure analysis on the second sentence of the essay to be graded to determine whether the second sentence uses advanced sentence structures;

[0188] When the second sentence does not use advanced sentence patterns, sentences whose semantic similarity with the second sentence reaches a fourth preset value and which use advanced sentence patterns are retrieved from the model text library to form a third sentence set;

[0189] Sorting each sentence in the third sentence set according to the semantic similarity, and generating a second recommendation set according to the sorting result, wherein the second recommendation set includes at least the sentence ranked first in the semantic similarity;

[0190] Based on the second recommendation set, modification suggestions are generated.

[0191] This can help writers better understand and apply advanced sentence patterns, thereby improving the language level and expression effect of their writing.

[0192] See also Figure 4 Based on the same inventive concept as the English composition evaluation method in the aforementioned embodiment, the present application embodiment provides an English composition evaluation device, comprising:

[0193] The first acquisition module 201 is used to acquire the composition to be graded;

[0194] The first scoring module 202 is used to perform semantic analysis on the essay to be scored, wherein the semantic analysis includes advanced vocabulary analysis, semantic coherence analysis, logical hierarchy analysis, sentence analysis and theme consistency analysis; and obtain advanced vocabulary scores, semantic coherence scores, logical hierarchy scores, theme consistency scores and advanced sentence scores;

[0195] The scoring synthesis module 203 is used to perform weighted summation of the high-level vocabulary score, semantic coherence score, logical hierarchy score, high-level sentence structure score and topic consistency score based on preset weight coefficients to obtain the final score of the composition to be scored.

[0196] See also Figure 5 In some embodiments, the first scoring module 202 includes:

[0197] The first tagging unit 202a1 is used to tag each word in the composition to be graded with a part of speech;

[0198] A first creating unit 202a2, configured to create a first matching set, where the first matching set is an empty set;

[0199] A first comparison unit 202a3, used to compare each word and its corresponding part-of-speech tag with a preset high-level vocabulary list, the preset high-level vocabulary list including a group of high-level words and their parts of speech and corresponding preset scores;

[0200] A first adding unit 202a4, configured to add the high-level word and the preset score of the corresponding high-level word to the first matching set when the word and its part of speech match the preset high-level word list and the high-level word does not appear in the first matching set;

[0201] The first calculation unit 202a5 is used to accumulate the preset scores of all high-level vocabulary in the first matching set to calculate the total score of the high-level vocabulary of the composition to be scored.

[0202] In some embodiments, the first scoring module 202 further includes a third acquisition unit, which is used to acquire the context representation of the sentence through BERT (Bidirectional Encoder Representations from Transformers), and evaluate the coherence by calculating the similarity between the sentences to obtain a semantic coherence score. For the specific method, please refer to the aforementioned English composition evaluation method, which will not be repeated here.

[0203] In some embodiments, the first scoring module 202 further includes a fourth acquisition unit for analyzing the logical hierarchy of the essay to be scored through Argument Mining Tools (logic analysis software) to obtain a logical hierarchy score. For specific methods, please refer to the aforementioned English essay evaluation method, which will not be repeated here.

[0204] See also Figure 5 In some embodiments, the first scoring module 202 further includes:

[0205] The second acquisition unit 202b1 is used to acquire a preset high-level sentence table, which includes a plurality of predefined high-level sentence types and their corresponding scores;

[0206] The first recognition unit 202b2 is used to perform syntactic analysis on each sentence in the composition to be graded and recognize the sentence structure in the sentence;

[0207] The first matching unit 202b3 is used to match the identified sentence pattern with the high-level sentence pattern table, determine whether each sentence pattern belongs to a high-level sentence pattern, and record the occurrence frequency of each high-level sentence pattern;

[0208] The second calculation unit 202b4 is used to calculate the contribution score of each advanced sentence pattern according to the frequency of occurrence of each advanced sentence pattern in the composition to be scored and the predefined score in the advanced sentence pattern table; the specific calculation formula is:

[0209] Contribution score of advanced sentence A = ∑(number of occurrences of advanced sentence A × preset score of advanced sentence A × penalty factor) where:

[0210] Advanced sentence pattern A represents any advanced sentence pattern;

[0211] The penalty factor is used to adjust the contribution score of high-level sentences that appear too frequently. The penalty factor calculation formula is:

[0212]

[0213] Among them, k is a constant greater than 0, which is used to control the intensity of the penalty;

[0214] The third calculation unit 202b5 is used to accumulate the contribution scores of all advanced sentences to obtain a preliminary advanced sentence score of the composition to be scored;

[0215] The fourth calculation unit 202b6 is used to obtain the total number of sentences in the composition to be scored, and divide the preliminary high-level sentence score by the total number of sentences to calculate the high-level sentence score; the specific calculation formula is:

[0216]

[0217] See also Figure 4 In some embodiments, the first scoring module 202 further includes: a fifth acquisition module, which is used to perform a theme consistency analysis on the composition to be scored to obtain a theme consistency score. For specific methods, please refer to the aforementioned English composition evaluation method, which will not be repeated here.

[0218] In some embodiments, the English composition evaluation device further comprises:

[0219] A first judgment module 204 is used to judge whether the advanced vocabulary score of the composition to be graded is lower than a first preset value;

[0220] The first screening module 205 is used to screen out a first sentence set without advanced vocabulary when the advanced vocabulary score of the composition to be graded is lower than a first preset value; the first sentence set includes the first sentence;

[0221] A first matching module 206, configured to select at least from a preset sample text library a set of second sentences whose semantic similarity with the first sentence reaches a second preset value and whose high-level vocabulary score exceeds the first preset value;

[0222] A first calculation module 207, configured to calculate a high-level vocabulary score for each sentence in the second sentence set;

[0223] A first sorting module 208 is used to sort each sentence in the second sentence set according to the high-level vocabulary score of each sentence in the second sentence set; and generate a first recommendation set according to the sorting result, the first recommendation set at least including the sentence with the highest high-level vocabulary score ranking;

[0224] The first suggestion generating module 209 is used to generate modification suggestions based on the first recommendation set. Exemplarily, the modification suggestions may be to highlight the parts that need to be modified in different colors in the document submitted by the user, and provide specific suggestions and explanations in the sidebar or comments.

[0225] See also Figure 4 In some embodiments, the English composition evaluation device further includes a first adding module 210, which is used to store the composition to be scored into a preset model essay library when the final score of the composition to be scored is greater than or equal to a third preset value.

[0226] See also Figure 4 In some embodiments, the English composition evaluation device further includes:

[0227] The second judgment module 211 is used to perform sentence analysis on the second sentence of the composition to be graded, and judge whether the second sentence uses a high-level sentence pattern;

[0228] The second screening module 212 is used to retrieve sentences from the model text library whose semantic similarity with the second sentence reaches a fourth preset value and uses the advanced sentence structure when the second sentence does not use the advanced sentence structure, to form a third sentence set;

[0229] A second sorting module 213 is used to sort each sentence in the third sentence set according to the semantic similarity, and generate a second recommendation set according to the sorting result, wherein the second recommendation set includes at least the sentence ranked first in the semantic similarity;

[0230] The second suggestion generating module 214 is configured to generate a modification suggestion based on the second recommendation set.

[0231] See also Figure 6 , based on the inventive concept of an English composition evaluation method in the aforementioned embodiment, the embodiment of the present application provides a terminal device. The terminal device may include but is not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), etc., and fixed terminals such as digital TVs, desktop computers, etc. The terminal device includes a processing device 301 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to the program stored in ROM302 (read-only memory) or the program loaded from the storage device 308 into RAM303 (random access memory). In RAM303, various programs and data required for the operation of the terminal device are also stored. The processing device 301, ROM302 and RAM303 are connected to each other via a bus 304. The input / output interface (i.e., I / O interface 305) is also connected to the bus 304.

[0232] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the terminal device to communicate with other devices wirelessly or by wire to exchange data.

[0233] In particular, according to some embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, some embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In some such embodiments, the computer program can be downloaded and installed from the network through the communication device 309, or installed from the storage device 308, or installed from the ROM 302. When the computer program is executed by the processing device 301, the above-mentioned functions defined in the method of some embodiments of the present application are executed.

[0234] It should be noted that the computer-readable medium recorded in some embodiments of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In some embodiments of the present application, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In some embodiments of the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0235] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an adhoc peer-to-peer network), as well as any currently known or future developed network.

[0236] The computer-readable medium may be included in the terminal device, or may exist independently without being installed in the terminal device. The computer-readable medium carries one or more programs, and when the one or more programs are executed by the terminal device, the method steps of any of the above technical solutions may be implemented.

[0237] Computer program code for performing the operations of some embodiments of the present application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages-such as Java, Smalltalk, C++, and also conventional procedural programming languages-such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on the remote computer, or completely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network-including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., using an Internet service provider to connect through the Internet).

[0238] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0239] The modules described in some embodiments of the present application may be implemented in software or hardware. The modules described may also be arranged in the processing device 301 .

[0240] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0241] Some embodiments of the present application further provide a computer program product, including a computer program, which implements any of the above-mentioned English composition evaluation methods when executed by the processing device 301.

[0242] The above description is only some preferred embodiments of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present application is not limited to the technical solutions formed by the specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above invention concept. For example, the above features are replaced with (but not limited to) technical features with similar functions applied for in the embodiments of the present application to form a technical solution.

[0243] Although the present invention has been described in detail above by general description and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements can be made to the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all belong to the scope of protection claimed by the present invention.

Claims

1. An English composition evaluation method, characterized in that: The following steps are involved: Get the essays to be graded; Performing semantic analysis on the composition to be scored, wherein the semantic analysis includes advanced vocabulary analysis, semantic coherence analysis, logical hierarchy analysis, sentence structure analysis and theme consistency analysis; obtaining an advanced vocabulary score, a semantic coherence score, a logical hierarchy score, a theme consistency score and an advanced sentence structure score; Based on the preset weight coefficients, the high-level vocabulary score, semantic coherence score, logical hierarchy score, high-level sentence structure score and topic consistency score are weighted and summed to obtain the final score of the composition to be scored.

2. The English composition evaluation method according to claim 1, characterized in that: The steps of performing advanced vocabulary analysis on the composition to be graded and obtaining an advanced vocabulary score include: Performing part-of-speech tagging on each word in the composition to be graded; Creating a first matching set, where the first matching set is an empty set; Comparing each word and its corresponding part-of-speech tag with a preset high-level vocabulary list, the preset high-level vocabulary list including a set of high-level words and their parts of speech and corresponding preset scores; When the word and its part of speech match a preset high-level vocabulary list, and the high-level vocabulary does not appear in the first matching set, adding the high-level vocabulary and the preset score of the high-level vocabulary corresponding to it to the first matching set; The preset scores of all the advanced vocabulary in the first matching set are accumulated to calculate the total score of the advanced vocabulary of the composition to be scored.

3. The English composition evaluation method according to claim 2, characterized in that: The steps of performing sentence analysis on the composition to be graded and obtaining a high-level sentence score include: Obtaining a preset high-level sentence table, wherein the high-level sentence table includes a plurality of predefined high-level sentence types and their corresponding scores; Performing syntactic analysis on each sentence in the composition to be graded to identify the sentence structure in the sentence; Matching the identified sentence patterns with the advanced sentence pattern table, determining whether each sentence pattern belongs to an advanced sentence pattern, and recording the occurrence frequency of each advanced sentence pattern; According to the frequency of occurrence of each advanced sentence pattern in the composition to be scored and the predefined score in the advanced sentence pattern table, the contribution score of each advanced sentence pattern is calculated; the specific calculation formula is: Contribution score of Advanced Sentence A ＝ ∑(Number of occurrences of advanced sentence A × Default score for Advanced Sentence A × Penalty factor) where: Advanced sentence pattern A represents any advanced sentence pattern; The penalty factor is used to adjust the contribution score of high-level sentences that appear too frequently; the penalty factor calculation formula is: Among them, k is a constant greater than 0, which is used to control the intensity of the penalty; Accumulating the contribution scores of all advanced sentences to obtain a preliminary advanced sentence score of the composition to be graded; The total number of sentences in the composition to be scored is obtained, and the preliminary high-level sentence score is divided by the total number of sentences to calculate the high-level sentence score; the specific calculation formula is:

4. The English composition evaluation method according to claim 1 or 3, characterized in that: The following steps are also included: Determining whether the advanced vocabulary score of the composition to be graded is lower than a first preset value; When the high-level vocabulary score of the composition to be graded is lower than a first preset value, a first sentence set without high-level vocabulary is selected; the first sentence set includes a first sentence; At least a second set of sentences whose semantic similarity with the first sentence reaches a second preset value and whose high-level vocabulary score exceeds the first preset value are selected from the preset model text library; calculating a high-level vocabulary score for each sentence in the second set of sentences; sorting each sentence in the second set of sentences according to a high-level vocabulary score of each sentence in the second set of sentences; and generating a first recommendation set according to the ranking result, wherein the first recommendation set at least includes the sentence ranked first by the high-level vocabulary score; Based on the first recommendation set, a modification suggestion is generated.

5. The English composition evaluation method according to claim 4, characterized in that: When the final score of the composition to be scored is greater than or equal to the third preset value, the composition to be scored is stored in a preset model essay library.

6. The English composition evaluation method according to claim 5, characterized in that: The following steps are also included: Performing sentence analysis on the second sentence of the composition to be graded to determine whether the second sentence uses an advanced sentence structure; When the second sentence does not use advanced sentence patterns, sentences whose semantic similarity with the second sentence reaches a fourth preset value and which use advanced sentence patterns are retrieved from the model text library to form a third sentence set; Sorting each sentence in the third sentence set according to semantic similarity, and generating a second recommendation set according to the sorting result, wherein the second recommendation set includes at least the sentence ranked first in semantic similarity; Based on the second recommendation set, the modification suggestion is generated.

7. An English composition evaluation device, characterized in that: include: The first acquisition module is used to acquire the composition to be graded; The first scoring module is used to perform semantic analysis on the composition to be scored, wherein the semantic analysis includes advanced vocabulary analysis, semantic coherence analysis, logical hierarchy analysis, sentence analysis and theme consistency analysis; and obtain an advanced vocabulary score, a semantic coherence score, a logical hierarchy score, a theme consistency score and an advanced sentence score; The scoring synthesis module is used to perform weighted summation of the high-level vocabulary score, semantic coherence score, logical hierarchy score, high-level sentence structure score and topic consistency score based on preset weight coefficients to obtain a final score for the composition to be scored.

8. A terminal device, characterized in that: including a processing device and a memory; The memory is used to store one or more program instructions; The processing device is used to run one or more program instructions to perform the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processing device, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processing device, the method according to any one of claims 1 to 6 is implemented.

Images