Individual difference-oriented bilingual story understanding adaptive teaching system and method

Abstract

The invention relates to the technical field of bilingual teaching, in particular to an individual difference-oriented bilingual story understanding adaptive teaching system and method, and the system comprises a structure adaptation analysis module, an understanding feature induction module, a material matching evaluation module, an expression trend tracking module and a strategy configuration adjustment module. According to the invention, through dynamic identification based on multi-dimensional features of roles, time sequences, syntax, keywords, semantics, vision and the like, unique differences of students in story understanding, expression content, language migration and cognitive reaction are comprehensively analyzed, continuous tracking of material content, an understanding process and an expression track is linked, individual feature parameters are automatically generated, and the learning efficiency is improved. According to the method, fine screening and task difficulty adjustment of story fragments are realized, the evolution trend is concluded and expressed in real time, intervention configuration is dynamically adjusted, language structure application, cognitive migration and continuous improvement of multivariate expression ability are promoted, high adaptation of teaching content and student ability is promoted, and individuation of a learning path and timeliness of feedback regulation are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bilingual teaching, and particularly relates to an individual difference-oriented bilingual story understanding adaptive teaching system and method. BACKGROUND

[0002] The bilingual teaching field involves using two or more languages as a medium to promote the comprehensive improvement of learners' language ability and cross-cultural communication ability through diversified language activities such as listening, speaking, reading and writing, including teaching content design based on the law of language ontology, multi-modal language input and output, language cognitive process analysis and teaching evaluation methods, and overall covering multiple disciplines such as cognitive science, linguistics, education informatization, emphasizing optimizing learning paths through scientific methods and data analysis to adapt to the needs of different learners. Among them, the traditional bilingual story understanding adaptive teaching system refers to the language ability differences of learners in the story understanding process. Generally, teachers judge by experience, grade teaching materials, and take after-school tests to stratify learners, and adjust the difficulty and progress of the classroom according to the preset teaching process and content. It is generally completed by means of artificial stratification, manual evaluation and static teaching content selection.

[0003] The existing technology relies on artificial stratification and teaching material grading, uses static content and after-school tests, lacks tracking of multi-dimensional expression characteristics and stage changes in the learning process, and is difficult to realize dynamic monitoring of the actual language development process of learners. In the actual classroom where the situation is diverse and the expression method is constantly changing, the teaching content and the students' ability often do not match, the teaching adjustment lags behind, the feedback is single, and some students are in an unsuitable teaching level for a long time, resulting in understanding deviation, reduced interest and slow ability improvement. SUMMARY

[0004] The purpose of the present application is to solve the shortcomings in the prior art, and an individual difference-oriented bilingual story understanding adaptive teaching system and method are provided.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme: an individual difference-oriented bilingual story understanding adaptive teaching system, the system comprising: A structure adaptation analysis module analyzes the number of roles, the order of events and the student's repetition content based on the role conversion type bilingual text, judges the actual distribution of the time sequence adverb and the character appellation, compares the order of the repetition and the original text, combines the perception level grading data, induces the language structure application difference, and obtains the structure adaptation performance parameter; An understanding feature induction module calculates the number of keyword reactions in bilingual plot questions and answers based on the structure adaptation performance parameter, analyzes the integrity of the syntactic structure, compares the semantic retention when bilingual interpretation, combines the type standard and the migration strength, induces the language understanding change, and obtains the understanding feature offset; The material matching evaluation module filters out the structural difference segments in the material library based on the understanding feature offset, compares the sentence type recognition performance and the visual understanding lag, judges the semantic connection interruption distribution, analyzes the story structure difference, and obtains a structural matching distribution index; The expression trend tracking module optimizes the expression of the language segment in the bilingual story repetition detection based on the structural matching distribution index, counts the vocabulary repetition frequency and the sentence type transformation number, analyzes the sentence group length and density change, combines the expression specification parameters, and obtains an expression trend trajectory group.

[0006] The application improves that the structural adaptation performance parameters include expression accuracy, language structure complexity and information integration level, the understanding feature offset includes semantic extraction sensitivity, knowledge transfer amplitude and understanding stability, the structural matching distribution index includes material diversity distribution, structure fitting degree and content correlation strength, and the expression trend trajectory group includes expression mode variation, content presentation trend and repetition coherence.

[0007] The application improves that the structural adaptation analysis module comprises: The character distribution extraction submodule extracts the number of characters and the order of the first appearance of the characters in the text based on the role conversion type bilingual text, identifies the character names one by one in combination with the student repetition text, compares the order of the appearance of the character names in the original text and the repetition text, counts the frequency of the order error of the first appearance of the characters between the two texts, and obtains the character order difference; The time sequence distribution comparison submodule calls the character order difference, detects all time sequence adverbs in the student repetition text, distinguishes the distribution range and the use frequency of the adverbs according to the paragraphs, arranges the adverbs used in the order of the events in the original text, compares the paragraph coverage and mutual exclusion distribution of the two texts according to the time mark, and obtains the event order dislocation rate; The structural adaptation induction submodule analyzes the connection word concatenation frequency, the syntax nesting level, the subject replacement frequency and the sentence group switching density in the student repetition content according to the event order dislocation rate, combines the scenario perception level grading data, matches the features in the type interval, and obtains the structural adaptation performance parameters.

[0008] The application improves that the understanding feature induction module comprises: The keyword response submodule counts the response of all scenario keywords of the student in the bilingual scenario question and answer task based on the structural adaptation performance parameters, distinguishes the difference between the task stages, classifies the position and the repetition number of the keywords in each stage, accumulates and totals the number of keyword responses, and obtains the total number of keyword responses. The syntax structure detection submodule decomposes the sentence of the answer content based on the total amount of keyword reaction quantity, searches the integrity of the subject, predicate, object and modification component, discriminates the syntax structure coverage performance according to the sentence structure type standard, and obtains a syntax structure integrity grade coefficient; The semantic migration comparison submodule extracts bilingual mutual translation content based on the syntax structure integrity grade coefficient, decomposes the target language vocabulary and sentence meaning information, compares the source language information retention performance, combines the migration strength and sentence change index, and obtains an understanding feature offset.

[0009] The material matching evaluation module comprises: The segment screening submodule obtains the segment structure features of the picture-text combined story material library based on the understanding feature offset, collects the number of sentence patterns, the total number of information levels and the frequency of role perspective changes of each segment, analyzes the structure span according to the differences of the sentence patterns, information levels and perspective frequencies, and obtains a structure difference segment set by screening the segments meeting the conditions. The expression recognition submodule collects the recognition records of the students on the segments in the bilingual listening and reading task according to the structure difference segment set, extracts the correct frequency of sentence pattern recognition, the task reaction time length and the error expression type code of each segment, compares the recognition frequency with the reference standard data, analyzes the influence of the combination change of the reaction time length and the error expression on the understanding rhythm, and obtains a sentence pattern understanding lag set. The structure comparison submodule compares the visual pause distribution, the total number of narrative jumps and the understanding lag performance of each segment according to the sentence pattern understanding lag set, obtains the matching deviation degree, analyzes the corresponding degree of the evolution path of the original structure style of the segment, and obtains a structure matching distribution index.

[0010] The expression trend tracking module comprises: The vocabulary repetition submodule extracts the key words in the detected content based on the structure matching distribution index, counts the appearance frequency of the vocabulary in each detection task sentence, distinguishes the distribution of the vocabulary in the difference time sequence sentences, obtains the vocabulary repetition density level according to the difference between the repetition times and the vocabulary coverage in the sentences. The sentence pattern transformation submodule counts the number of sentence pattern types of each detection sentence based on the vocabulary repetition density level, distinguishes the change categories of the sentence structure between adjacent sentences, compares the use situation of the difference sentence pattern combinations, induces the sentence difference type and the use frequency, and obtains a sentence pattern transformation distribution difference. The expression proportion analysis submodule extracts the sentence group length and the number of sentence patterns of each sentence in the detected content based on the sentence pattern transformation distribution difference, compares each structure according to the expression specification parameters, calculates the expression density offset amplitude of each sentence, integrates the associated expression density data, and obtains an expression trend trajectory group.

[0011] The system further comprises: The policy configuration adjustment module adjusts the task arrangement rhythm, judges the semantic prompt interval setting, analyzes the graph-text intervention and expression trajectory matching, compares the intervention configuration positioning, induces the parameter change, and obtains the intervention parameter offset amplitude based on the expression trend trajectory group. The intervention parameter offset amplitude includes task distribution adjustment amplitude, intervention response sensitivity, and strategy matching level.

[0012] The policy configuration adjustment module comprises: The prompt rhythm judgment sub-module analyzes the arrangement rhythm of each task in the teaching library, compares the difference between the graph-text intervention prompt distribution, judges the density change of the semantic intervention prompt in each task interval, screens the abnormal interval with the best correlation with the sentence turning point, and obtains the prompt interval offset amount based on the expression trend trajectory group. The intervention fusion evaluation sub-module analyzes the distribution of the graph-text intervention fusion density on the expression node of the sentence, judges the corresponding situation of the intervention configuration between the expression time period and the node type, optimizes the intervention distribution order, and obtains the intervention configuration positioning frequency based on the prompt interval offset amount. The configuration offset calculation sub-module analyzes the graph-text intervention fusion density, sentence type density ratio, vocabulary repetition frequency and sentence expression type difference, compares the change process of structure arrangement and expression method, and obtains the intervention parameter offset amplitude based on the intervention configuration positioning frequency.

[0013] The bilingual story understanding adaptive teaching method for individual differences is executed based on the bilingual story understanding adaptive teaching system for individual differences, and comprises the following steps: S1: Based on the role conversion type bilingual text, the number of roles and the order of events are analyzed, the actual distribution of the time sequence adverbs and the person's name used by the student in the restatement text is judged in combination with the student's bilingual story restatement content, the corresponding situation between the student's restatement order and the original text order is compared, the structure adaptation performance parameters are obtained according to the plot perception level grading data. S2: Based on the structure adaptation performance parameters, the reaction times of the keywords involved in the bilingual plot question and answer process of the student are calculated, the complete performance of the syntactic structure in the answer content is analyzed, the semantic retention of the bilingual interpretation link is compared, the language understanding feature change of the student is induced, and the understanding feature offset amount is obtained. S3: Based on the understanding feature offset amount, the story segments with obvious structure difference in the graph-text combined story material library are screened, the sentence type identification performance and visual understanding lag of the student in the bilingual listening and reading task are compared, the distribution characteristics of the semantic connection breakpoints are judged, and the difference between the story segment structure and the student's understanding performance is analyzed, and the structure matching distribution index is obtained. S4: Based on the structure matching distribution index, the expression of the student bilingual story repetition detection process is optimized, the vocabulary repetition frequency in the detection content is counted, the number of sentence type transformation is measured, the proportional change of the sentence group length and the sentence type density is analyzed, the expression mode distribution trend is induced, and an expression trend trajectory group is obtained; S5: Based on the expression trend trajectory group, the teaching library task arrangement rhythm is adjusted, the semantic intervention prompt setting in each interval is judged, the matching relationship between the picture-text intervention fusion density and the student expression trajectory is analyzed, the positioning of the intervention configuration in the expression process is compared, and an intervention parameter offset amplitude is obtained.

[0014] Compared with the prior art, the advantages and positive effects of the present application are that: In the present application, through dynamic recognition based on role, time sequence, syntax, keyword, semantics, vision and other multi-dimensional characteristics, the unique differences of students in story understanding, expression content, language transfer and cognitive response are comprehensively analyzed, the continuous tracking of material content, understanding process and expression trajectory is linked, individual characteristic parameters are automatically generated, the fine screening of story segments and task difficulty adjustment are realized, the expression evolution trend is induced in real time, the intervention configuration is dynamically adjusted, the continuous improvement of language structure application, cognitive transfer and multi-expression ability is promoted, the high adaptation of teaching content and student ability is promoted, and the individualization of learning path and the timeliness of feedback regulation are enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 The system flowchart of the present application is shown in the figure; Figure 2 The flowchart of the structure adaptation analysis module in the present application is shown in the figure; Figure 3 The flowchart of the understanding characteristic induction module in the present application is shown in the figure; Figure 4 The flowchart of the material matching evaluation module in the present application is shown in the figure; Figure 5 The flowchart of the expression trend tracking module in the present application is shown in the figure; Figure 6 The flowchart of the strategy configuration adjustment module in the present application is shown in the figure. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0017] In the description of the present application, it should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, in the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0018] Embodiment one, please refer to Figure 1 The present application provides a technical solution: a bilingual story understanding adaptive teaching system for individual differences includes: The structure adaptation analysis module analyzes the number of characters and the order of events based on the role conversion type bilingual text, judges the actual distribution of the student's use of the temporal adverb and the character's appellation in the restatement text, compares the corresponding situation between the student's restatement order and the original text order, induces the student's adaptive performance in the application of language structure according to the story plot perception level grading data, and obtains the structure adaptation performance parameters; The understanding feature induction module calculates the number of reactions of the student to the key words involved in the bilingual plot question and answer process based on the structure adaptation performance parameters, analyzes the complete performance of the syntactic structure in the answer content, compares the semantic retention of the bilingual interpretation link, and induces the student's language understanding feature changes according to the sentence construction type standard, the key word extraction set and the language transfer strength reference, and obtains the understanding feature offset; The material matching evaluation module filters the story segments with obvious structural differences in the picture-text combined story material library based on the understanding feature offset, compares the student's sentence pattern recognition performance and visual understanding lag in the bilingual reading task, judges the distribution characteristics of the semantic connection breakpoints, and analyzes the differences between the story segment structure and the student's understanding performance, and obtains the structure matching distribution index; The expression trend tracking module optimizes the expression of the student's bilingual story restatement detection process based on the structure matching distribution index, counts the frequency of vocabulary repetition in the detection content, measures the number of sentence pattern transformations, analyzes the proportional changes of sentence group length and sentence pattern density, induces the expression mode distribution trend according to the expression specification corresponding parameters, and obtains the expression trend trajectory group; The strategy configuration adjustment module adjusts the task arrangement rhythm in the teaching library based on the expression trend trajectory group, judges the setting of semantic intervention prompts in each interval, analyzes the matching relationship between the picture-text intervention fusion density and the student's expression trajectory, compares the positioning of the intervention configuration in the expression process, induces the change of the intervention parameter configuration, and obtains the intervention parameter offset amplitude.

[0019] The structure adaptation performance parameters include expression accuracy, language structure complexity, and information integration level. The understanding characteristic deviation amount includes semantic extraction sensitivity, knowledge transfer amplitude, and understanding stability. The structure matching distribution indexes include material diversity distribution, structure fitting degree, and content correlation strength. The expression trend trajectory groups include expression mode variation, content presentation trend, and repetition coherence. The intervention parameter deviation amplitude includes task distribution adjustment amplitude, intervention response sensitivity, and strategy matching level.

[0020] In module 1, the role conversion type bilingual text refers to a bilingual (such as Chinese and English) story material containing multiple roles and having role identity, language expression or perspective switching in the process of narration. The time sequence adverb refers to an adverb (such as “first”, “then”, “last”, “afterwards”, “meanwhile” and the like) used to express time sequence or event sequence in story narration or repetition. The actual distribution in the repetition text refers to the position, frequency and distribution rule of the time sequence adverb or the person's name in the student's narration content when he / she repeats the story. The corresponding situation refers to the specific one-to-one matching or inconsistency of the student's repetition content and the original story content in terms of role sequence, event sequence and the like. The perception level grading data refers to the data parameters formed by dividing the student's ability in story plot understanding and expression into several levels or intervals according to educational psychology, curriculum standards and the like. The adaptation performance refers to the adaptation and performance level of the student in the process of repetition or understanding in terms of story structure, language expression and role conversion and the like.

[0021] In module 2, the number of keyword reactions refers to the number of times the student accurately identifies, mentions or uses the target keyword (such as a person, an action, a plot point and the like) in bilingual plot question and answer or test. The complete performance refers to the completeness of the sentence or paragraph in terms of syntactic structure (such as complete subject-predicate-object, complete modification component and the like) in the student's answer, repetition and the like. The semantic retention situation refers to the degree of complete retention or loss of original meaning, information amount and logical relationship when the student translates information from one language to another language. The construction type standard refers to the authoritative standard or classification table used to distinguish sentence structure types (such as simple sentence, parallel sentence, complex sentence and the like) or expression modes in the field of language education. The language transfer strength reference refers to the reference standard or quantitative interval for evaluating the influence degree of the student's mother tongue knowledge on the expression of the second language in the process of bilingual conversion. The language understanding characteristic change refers to the change characteristics of the student's language understanding ability, mode, accuracy and the like under different tasks, different stages and different materials.

[0022] In module 3, the story fragment with obvious difference refers to the bilingual story fragment with a large structural difference in terms of sentence pattern, structure, information amount, expression difficulty, etc. from the current ability of students; the sentence pattern recognition performance refers to the category and quantity of sentence patterns that students can accurately identify and understand in listening and reading tasks, as well as the actual performance; the visual understanding lag refers to the lack of reaction speed or synchronization between image information and text information when students receive picture-text materials; the distribution characteristics of the interruption point refer to the specific location and distribution law of the semantic flow interruption or pause (i.e. understanding obstacles, skipping, missing information, etc.) in the process of understanding the materials.

[0023] In module 4, the expression of the language segment refers to the expression of the complete language segment content of the students in the detection or test link; the frequency of vocabulary repetition refers to the frequency of the same vocabulary being used multiple times in the students' repetition or expression, reflecting the richness or singularity of vocabulary application; the proportion change refers to the proportion and its change trend between the sentence group length and the sentence pattern density in different detection contents or task stages; the expression specification corresponding parameter refers to the standardized parameter related to language expression ability, such as the language, sentence pattern or structure specification in the textbook, curriculum standard, grade examination, etc.; the distribution trend of expression mode refers to the frequency and its time variation law of various expression modes (such as simple description, complex repetition, reasoning expression, etc.) in the students' expression process.

[0024] In module 5, the task arrangement rhythm refers to the sequence, switching speed and rhythm arrangement of the task link in the teaching activities, i.e. the distribution and promotion rhythm of each learning task; the setting in each interval refers to the interval division and specific arrangement of different teaching parameters or prompt content in different tasks, links, difficulty segments, etc.; the matching relationship refers to the adaptation degree and corresponding relationship between the teaching intervention measures (such as picture-text combination, task allocation, semantic prompt, etc.) and the actual learning process and performance of students; the positioning in the expression process refers to the specific location of different intervention configurations applied in which link and which stage in the students' expression activities; the change of intervention parameter configuration refers to the adjustment and change of the intervention means or parameter settings in teaching with the change of students' performance or task demand.

[0025] Please refer to Figure 2 , the structure adaptation analysis module includes: The role distribution extraction sub-module extracts the number of roles and their first appearance order in the text based on the role conversion type bilingual text, identifies the characters one by one in the students' repetition text, and compares the appearance order of the character names in the original text and the repetition text to obtain the frequency of the first appearance order of the characters in the two texts. The role conversion type bilingual text needs to extract all the role words with personal characteristics. First, read the original bilingual text and filter out the names or pronoun forms of each character, such as "mom", "Tom", "the boy", "he", "Xiaoming" and other titles in Chinese and English stories. Then arrange them in order according to their first appearance in the text to form the character appearance sequence. Then read the student's retelling text and identify and extract all the character names. Determine whether the student used the same character name or alternative representation as the original text, such as replacing "the girl" with "Xiaohong". Compare the first appearance of each character name in the student's retelling with the order in the original text. When the order of a character in the retelling does not match the order of its first appearance in the original text, record it as a misplacement. For example, if the original text has characters A, B, C, and D in order, and the student's retelling has the order B, A, D, C, then the positions of A, B, C, and D in the student's retelling do not match the original text, and the misplacement frequency is 4. By comparing the order of the first appearance of all characters in the two texts, the cumulative number of character misplacements is summarized as the character order difference. Assuming that the number of characters in the student's retelling that match the original text is 3, and the number that do not match is 2, then the character order difference is 2. This difference value will be used in the downstream module as a reference for analyzing the student's story structure.

[0026] The time sequence distribution comparison submodule calls the character order difference, detects all the time sequence adverbs in the student's retelling text, and divides their distribution range and frequency by paragraph. Arrange the adverbs used in the original text to match the order of events, and mark the time in the paragraph coverage and mutual exclusion distribution of the two texts. Compare and get the event order misplacement rate. The time sequence adverbs in the student's retelling of the text are extracted and counted word by word, including phrases such as "first", "then", "last", "afterwards", "meanwhile", etc. The distribution position and frequency of each adverb are recorded according to the paragraph in which the adverb is located, forming an adverb distribution table. The time sequence adverbs in the original text are also processed in the same way to form an original timeline adverb distribution table. The same adverbs in the two adverb distribution tables are compared segment by segment. For example, the original text uses "first" to describe event A in the first paragraph, and "then" to describe event B in the second paragraph. If the student uses "then" in the first paragraph and "first" in the second paragraph in the retelling, the adverb distribution of the two paragraphs is mutually exclusive, which is recorded as an event sequence dislocation. The same comparison is made for all paragraphs in the text. When the time sequence adverbs used in a retelling paragraph do not maintain the same order as the original text, the paragraph is recorded as a dislocation paragraph. The number of dislocation paragraphs is counted and compared with the total number of paragraphs to obtain the event sequence dislocation rate. For example, if the original text has 5 paragraphs and 2 paragraphs have dislocation in the student's retelling, the event sequence dislocation rate is 0.4. According to the set standard, the dislocation rate between 0 and 0.2 is low dislocation, between 0.2 and 0.5 is medium dislocation, and greater than 0.5 is high dislocation. This dislocation rate will be used as a basis for the accuracy of connection and transition in the downstream structure adaptation analysis.

[0027] The structure adaptation induction sub-module analyzes the frequency of connection word strings, the level of syntactic nesting, the frequency of subject replacement, and the density of sentence group transitions in the student's retelling based on the event sequence dislocation rate. It also combines the scenario perception level data and matches the features in the type interval to obtain the structure adaptation performance parameters. The use of conjunctions in the student's paraphrase content is extracted, the number of each type of conjunction such as "because", "but", "so", "and", "but" in the paraphrase text is counted, the total frequency is calculated, and then the average connection use density is divided according to the number of sentences in each paragraph. Then read the structure of each sentence, analyze its syntactic structure, identify the nesting level, for example, whether a sentence contains a modifying clause, a dative clause or an adverbial component, record the nesting level of each sentence, calculate the average nesting depth, analyze the use of subjects in the text, record whether the subject changes in consecutive sentences, such as "I→Tom→he→theboy" expression, determine whether the subject is replaced, and count the frequency of replacement in the entire text. Then divide the sentence group of the entire paraphrase text, for example, every 3 consecutive sentences as a sentence group, count whether the sentence group is connected by adverbs or conjunctions, record the proportion between the number of sentence groups connected by adverbs or conjunctions and the total number of sentence groups, compare the value with the existing perception level grading data, for example, set the average length of the sentence group of the standard student in the high grade of primary school to 15 words, the average density of the conjunction is 3 in each paragraph, and the nesting level of the clause is 2 layers. If the conjunction density of a student in the paraphrase is 2.5, the nesting level is 1.5 layers, it matches the low adaptive interval, and by classifying the conjunction frequency, nesting level, subject replacement frequency, and sentence group transition density into corresponding intervals according to the level, the expression accuracy, language structure complexity, and information integration level are output as structure adaptation performance parameters.

[0028] Please refer to Figure 3 , understand that the feature induction module includes: The keyword response sub-module counts the response of all scenario keywords of the student in the bilingual scenario question and answer task based on the structure adaptation performance parameters, distinguishes the difference between the task stages, classifies the position and repetition number of the keywords in each stage, and accumulates and totals the number of keyword responses to obtain the total number of keyword responses. The structural adaptation performance parameter is used as a front input to judge the keyword response behavior of students in a bilingual scenario question and answer task. First, the text content in the question and answer task is structurally disassembled to extract the core scenario keywords set in the stem and standard answer, such as marking "girl", "open door", "find gift", "feel surprised" and other words in Chinese and English question and answer materials as the target keyword list. Then, the student's answer content is called, and the keyword hit in the student's answer is extracted in each stage according to the stage-by-stage prompts or task steps set in the task, such as asking for the role identity in stage one, asking for the event cause in stage two, and asking for emotional feedback in stage three. In the extraction process, comparison is made through word matching, the hit position and the number of occurrences of the same keyword are recorded, such as the student answering "the girl opened the door and the girl was very surprised" in response to "Who opened the door?". "The girl" appears twice, "opened the door" hits once, and "surprised" hits once, which are respectively counted into the keyword response record. The number of keywords in different task stages is classified, such as 3 keywords hit in stage one, 2 keywords hit in stage two, and 1 keyword hit in stage three. The total number of keywords in all stages is 6 keyword responses, and the frequency of repeated use is also accumulated into the total number, such as "the girl" repeated 3 times, so the total number of responses is 3 instead of 1. The sum of the number of keyword responses in all task stages is the total number of keyword responses, which is output as 12 times in this round of task. According to the response intensity, the low response interval is 04, the medium response interval is 59, and the high response interval is 10 and above.

[0029] The syntax structure detection submodule is based on the total number of keyword responses to disassemble the sentence of the answer content, retrieve the integrity of the subject, predicate, object and modification components, and classify the syntax structure coverage performance according to the sentence structure type standard to obtain the syntax structure integrity level coefficient. The total number of keyword reactions is taken as the input basis to decompose and process the sentence structure used by the students in the question and answer link. First, read the student's answer content text piece by piece, and decompose the language components of each sentence. Check if it contains complete core syntactic structure components such as subject, predicate, and object. At the same time, check if there are modifying components such as adjectives, adverbs, or prepositional phrases. For example, the student's answer is "the boy quickly ran into the room". The sentence contains "the boy" as the subject, "ran" as the predicate, and "into the room" as the object structure. "Quickly" is a modifying component. The sentence is judged to be complete in structure. If the answer is "ran quickly", the subject is missing and the structure is incomplete. Perform this kind of decomposition and retrieval operation on all answers sentence by sentence, and mark the complete or not label result. Complete the structure classification and judgment of all sentences. According to the sentence structure type standard, assign a value to each structure label. For example, the complete subject-predicate-object structure is level 3, the subject-predicate structure is level 2, and only the predicate or subject is level 1. If there are parallel sentences or compound sentences, they are weighted to level 4 or 5 according to the nesting layer. After completing the structure level division of all sentences, the average or mode classification of all level values is performed to determine the syntactic structure complete level coefficient of the student in this answer. If 6 out of 10 sentences are subject-predicate-object structures, 3 are subject-predicate structures, and 1 is missing the subject, the level weights are 3, 2, and 1 respectively. The total weighted value is 25, and the average level is 2.5. According to the set interval, 2.5 to 3.0 is the medium-high structure complete degree interval. Finally, the syntactic structure complete level coefficient of the student's answer this time is determined to be medium-high level.

[0030] The semantic migration comparison submodule extracts bilingual mutual translation content based on the syntactic structure complete level coefficient, decomposes the target language vocabulary and sentence meaning information, compares the source language information retention performance, combines the migration strength and sentence change indicators, and uses the formula:

[0031] Get the understanding feature offset ΔLS, where Qs i represents the semantic encoding of the i-th keyword in the source language, Qt i represents the semantic encoding of the i-th corresponding translation unit in the target language, Qm i represents the content significance weight of the i-th keyword, Qc i represents the component coverage number of the i-th sentence syntactic structure, Qr j represents the number of missing migration semantics in the j-th translation sentence, k represents the total number of translation sentences, and n represents the total number of keywords.

[0032] The feature offset is used to describe the comprehensive offset of language understanding features caused by language transfer, expression conversion, structure change, etc. in the bilingual story understanding and retelling task. The higher the value, the weaker the semantic retention, the more the migration defects, and the lower the structure matching degree in the process of language understanding and expression. On the contrary, the language understanding process is more stable, accurate, and smooth.

[0033] The structure of the English-Chinese translation text completed by the student in the bilingual task is extracted, the key words and sentence meaning units are vectorized and coded, and the source language coding set Qs is constructed according to the semantic direction of each key word in the original language i, At the same time, the target language coding set Qt is constructed according to the corresponding expression in the student translation text i, Taking the key words "river", "build", and "dangerous" used by the sample student in the "natural disaster scenario" bilingual task as an example, the original semantic intensity scores are 0.92, 0.85, and 0.89 respectively, the semantic retention scores in the translation sentence are 0.87, 0.75, and 0.83 respectively, and the normalized coding after semantic domain standardization is:

[0034] At the same time, the task sets the key word saliency weight Qm1=0.85 for "river" as the main element of the environment, sets Qm2=0.95 for "build" as the action verb, and sets Qm3=0.75 for "dangerous" as the emotional adjective. Combined with the results of the syntax analysis module, the main structure integrity distribution in the corresponding sentence of the student is 3, 4, and 3, and the corresponding normalized structure coverage number is Qc1=0.88, Qc2=0.93, and Qc3=0.88. For the information defects in the translation results, the system records the migration semantic defects as Qr1=0.20, Qr2=0.10, and Qr3=0.30, k=3, n=3. Substitute the above data into the formula, first, calculate the absolute difference multiplied by the weight:

[0035] Total = 0.0255 + 0.0665 + 0.0375 = 0.1295; Second, calculate the square root of the denominator:

[0036] Third, calculate the first fraction:

[0037] Fourth, calculate the second mean value:

[0038] The combined calculation results are as follows:

[0039] The results show that when the calculated understanding feature offset ΔLS=0.2833 falls within the set intermediate offset interval [0.25, 0.40], it indicates that the student has moderate semantic retention deviation and unstable structure expression phenomenon in the process of bilingual translation, specifically, the frequency of partial semantic weakening, insufficient syntactic structure support and information deficiency in the translated sentence is relatively high; the formula realizes the numerical description of cross-language semantic alignment offset by introducing the cross-action mechanism among keyword semantic difference, significance weight, structure completeness factor and migration deficiency rate, providing continuous measurable variable input for subsequent automatic adjustment of material structure and strategy configuration.

[0040] Referring to Figure 4 , the material matching evaluation module comprises: The segment screening submodule obtains the segment structure features of the picture-text combined story material library based on the understanding feature offset, collects the number of sentence patterns, the total number of information levels and the frequency of role perspective changes of each segment, analyzes the structure span according to the differences in sentence patterns, information levels and perspective frequencies, and obtains a structure difference segment set by screening the segments meeting the conditions; The structural features of each segment are extracted one by one. First, the text of each segment is disassembled into sentences, and the number of sentence patterns appearing in it is counted. The declarative sentences, interrogative sentences, exclamatory sentences, passive sentences, and parallel sentences are grouped according to the frequency of occurrence, and the total number of each type of sentence pattern is calculated and the number of occurrences of each type of sentence pattern is recorded. For example, in a certain segment, there are 4 types of sentence patterns, 10 declarative sentences, 3 passive sentences, 2 parallel sentences, and 1 interrogative sentence. The number of sentence patterns is 4, and the total number of sentence patterns is 16. Then the information level of the text is evaluated. By comparing the paragraph structure with the image label, the information points in the text are divided into primary events, detailed descriptions, background information, or emotional descriptions according to whether they are primary events, detailed descriptions, background information, or emotional descriptions. The number of different information levels appearing in the segment is counted. For example, a certain segment contains 2 primary events, 3 secondary details, and 1 tertiary emotion, so the total number of information levels is 6. At the same time, the frequency of changes in the expression of the character's perspective in the segment is scanned. When the narrative angle of a character switches from first person to third person within a paragraph, or from a single character description to a multi-character interlaced expression, it is considered as a change in perspective. The number of perspective changes in each segment is counted. For example, in a certain segment, there are 3 perspective transitions such as "I think that…" and "He actually…", so the frequency of perspective changes is 3. Then the three types of parameters are compared and the maximum and minimum values of the sentence pattern difference, information level span, and perspective switching span among all segments in the same batch of materials are calculated. Then, whether the span exceeds the structural selection standard is determined by setting a threshold value. The sentence pattern difference threshold is set to 3 types, the information level span threshold is set to 4 layers, and the perspective change frequency threshold is set to 2 times. When a certain segment exceeds any of these thresholds, it is determined that the structural span is large, and it is selected into the structural difference segment set. For example, segment A has 5 types of sentence patterns, 7 layers of information levels, and 5 times of perspective switching, all of which exceed the corresponding thresholds, so segment A is selected into the structural difference segment set.

[0041] The performance recognition sub-module collects the recognition records of students in the bilingual listening and reading task according to the structural difference segment set, extracts the correct frequency of sentence pattern recognition, task reaction time, and error expression type code of each segment, compares the recognition frequency with the reference standard data, analyzes the influence of the combination of reaction time and error expression on understanding rhythm, and obtains the sentence pattern understanding lag set. First, the student's recognition record of each difference segment is extracted, and the recorded listening and reading behavior log is disassembled according to the timestamp. After marking whether the student correctly identifies the main sentence type or structure style after playing each segment, the total number of sentence patterns correctly identified in the segment is counted, and the correct recognition frequency is calculated. For example, if segment A contains 9 main sentences and the student correctly identifies 6 of them, the correct frequency is 6. Then read the answering time after listening and reading, and record the time the student takes from listening to starting to answer as the task reaction time, which is counted in seconds. For example, if the student starts answering within 6 seconds after listening to the segment, the reaction time is recorded as 6 seconds. Sort all paragraph reaction times in order, compare the time trend of each paragraph, and then encode the language errors in the student's answers. Set up an error expression type coding dictionary, such as E01 for sentence inversion error, E02 for word collocation error, and E03 for time usage error. Mark the error code and its quantity in each answer, such as 1 E01 and 2 E03 in segment B, which means there are 3 error expressions, recorded as E01 x 1, E03 x 2. Then compare the correct recognition frequency of each segment with the reference standard data, which is set as the average number of correctly recognized sentences in the grade, which is 7 sentences. If the student only recognizes 5 sentences in the paragraph, it is lower than the benchmark, indicating that the understanding is insufficient. Jointly analyze the reaction time and error expression type when the reaction time is greater than the average value (such as 8 seconds) and the total number of error types exceeds 2, and determine that there is a lag in understanding. For example, segment C has a reaction time of 12 seconds and 3 types of errors, which is marked as a lagging segment. Collect all paragraph numbers that meet the conditions and their recognition frequency, reaction time, and error code to form a sentence pattern understanding lag set.

[0042] The structure comparison submodule compares the visual pause distribution, narrative jump total, and understanding delay performance of each segment according to the sentence pattern understanding lag set, using the formula:

[0043] Get the matching deviation degree, analyze the corresponding degree of the evolution path of the original structure style of the segment, and get the structure matching distribution index, where DU s represents the matching deviation degree of the structure difference segment under multi-feature comparison, n DU represents the total number of structure difference segments, VU z represents the number of visual pauses in the zth segment, c represents the total number of narrative jumps in the zth segment, and c represents the difference between the correct frequency and the reference standard in the zth segment. TU z represents the number of error expression types in the zth segment.

[0044] The matching deviation degree refers to the difference between the actual understanding performance of students in the bilingual listening and reading task and the original structural characteristics (such as visual pauses and narrative jumps) of each structural difference segment, reflecting the adaptation of students to different structural segments: the larger the value, the greater the difference between the structural understanding performance of students in the segment and the structural characteristics of the segment itself, and the more obvious the deviation of the pause, jump and error expression in the understanding process from the original structure; the smaller the value, the more consistent the understanding process of students with the original structure of the segment, and the smaller the deviation.

[0045] The original data of segment A is set as 3 times of visual pause, 2 items of narrative jump, 1 time less than the reference standard of correct recognition frequency, and 3 items of error expression, and the normalized values are VU z =0.60, JU z =0.40 、 RU z =0.20 、 TU z =0.60 ， Substitute the data into the formula, and the calculation process of segment A is as follows:

[0046] The original data of segment C is set as 5 times of visual pause, 3 items of narrative jump, 2 times of recognition frequency difference, and 2 items of error expression, and the normalized values are VU z =1.00, JU z =0.60 、 RU z =0.40 、 TU z =0.40, then:

[0047] The data of segment E is set as 2 times of visual pause, 1 item of narrative jump, 0 of recognition frequency difference, and 1 item of error expression, and the normalized values are VU z =0.40, JU z =0.20 、 RU z =0.00 、 TU z =0.20, and the calculation is as follows:

[0048] The original data of segment H is set as 6 times of visual pause, 4 items of narrative jump, 3 times of recognition frequency difference, and 4 items of error expression, and the normalized values are VU z =1.20, JU z =0.80 、 RU z =0.60 、 TUz = 0.80, calculated as follows:

[0049] The matching deviation results of the above four segments are substituted into the formula to calculate the average deviation degree:

[0050] The result shows that when the calculated matching deviation degree DU s = 0.5109 is compared with the set structure matching reference interval [0.00, 0.30], it can be judged that the current student's structure understanding performance on the selected structure difference segment has exceeded the upper limit of the basic fitting interval, and is in the moderate deviation section [0.30, 0.70], indicating that the student has a certain degree of inconsistency and fracture tendency in visual recognition, sentence restoration and structure maintenance ability. The formula integrates the visual and narrative jump load of the student through the square root term, reflects the actual comprehensive degree of structure pressure, and quantifies the difference with the weighted influence of recognition gap and expression error, ensuring that the evaluation of matching degree has a balanced and stable numerical basis.

[0051] Please refer to Figure 5 , the expression trend tracking module includes: The lexical repetition sub-module extracts key words in the detected content based on the structure matching distribution index, counts the frequency of words in each detected task segment, distinguishes the distribution of words in the difference time sequence segment, and obtains the lexical repetition density level according to the difference between the number of repetitions and the coverage of word classes in the segment. The original text content of each detection sentence is read from the task setting, and all meaningful words are extracted according to the part-of-speech classification tool, excluding function words such as "the", "a", "is", etc. to form a preliminary word list. For example, the detection sentence is "Tom ran to the door and opened it happily", the extracted word items are "Tom", "ran", "door", "opened", "happily", and the rest of the word items are not counted as key words. Then the frequency of each word item in the current sentence is counted one by one. For example, "Tom" appears once in this sentence, "ran" appears once, and the word item frequency is recorded as 1 respectively. After the statistics is completed, repeated detection is performed on all sentences to determine whether a key word appears repeatedly in other sentences of the same detection text. If a repeated record appears, mark the time sequence position of its appearance and construct a distribution correspondence table of word items and sentence positions. For example, "door" appears in both sentences 1 and 4, which are recorded as the 3rd position in the 1st sentence and the 6th position in the 4th sentence. Then, according to the task time sequence set in the detection text, it is determined whether such repeated appearance crosses a large time span between different segments. For example, sentence 1 is the beginning scene and sentence 4 is the ending scene, so this word is a time difference distribution word. Then, the number of repetitions of each key word and the span of its distributed sentences are combined to determine whether a word is a high-repetition word. If a word is repeated more than 3 times and appears in 3 different time logic segments, it is recorded as a high-repetition word. The frequency of all high-repetition words and the total amount of low-frequency words are compared to obtain the difference degree of word repetition distribution. Then, combined with the coverage range of the word class to which each word item in each sentence belongs, such as whether it is only a verb and a noun, or it covers adverbs, adjectives, preposition phrases, etc., the word class diversity level of each segment is determined. The segment with more than 5 types of word classes is defined as a high-coverage segment, and the segment with less than or equal to 3 types of word classes is defined as a low-coverage segment. Finally, the word repetition density is constructed according to the repetition frequency and the number of word classes in each segment. The repetition density level is divided according to the density value. For example, 0.3-0.5 is medium density, and more than 0.5 is high density. The output word repetition density level.

[0052] The sentence pattern transformation submodule is based on the word repetition density level, and counts the number of sentence pattern types in each detection sentence. The change category of the sentence pattern structure between adjacent sentences is distinguished, the difference in sentence pattern combination is compared, the difference type and usage frequency of the sentence pattern are summarized, and the sentence pattern transformation distribution difference is obtained. Detect the use of sentence structure in each section, read each sentence in the detected section, identify and mark its sentence type, judge whether it is a simple sentence, a parallel sentence, a passive sentence, a conditional sentence, an imperative sentence, an exclamatory sentence, etc., and count the number of each type in the current paragraph, such as the 3rd paragraph contains 4 sentences, 2 of which are simple sentences, 1 passive sentence and 1 conditional sentence, then record the number of sentence types as 3, then compare the adjacent sections horizontally to judge the change type of the sentence type used in the two sections before and after, for example, there are 2 new sentence types between the 3 sentence types of section 3 and the 5 sentence types of section 4, which is judged as an increase type, if the sentence types used in the two sections are exactly the same, it is recorded as a repeated type, if all the sentence types in section 4 are different from those in section 3, it is recorded as a completely different type, after recording the sentence type change type between all adjacent sections, continue to count the total number of each type of sentence type change in the detected text, for example, "increase type" change appears 4 times, "decrease type" 2 times, "repeat type" 3 times, and "completely different type" 1 time, then compare the number of changes of each type in the high and low density sections to determine whether a certain type of change is concentrated in a certain density interval, if the completely different type of sentence change mainly occurs in the low density section, it can be classified as a low density sentence difference type, otherwise it is classified as a high density corresponding, finally, according to the difference type, distribution interval and frequency of sentence type use, generate a sentence type change distribution table, count the proportion of each type of change in the detection task, and output the sentence type change distribution difference.

[0053] The expression ratio analysis submodule extracts the sentence group length and the number of sentence types in each section of the detected content based on the sentence type change distribution difference, compares each structure according to the expression specification parameters, and uses the formula:

[0054] Calculate the expression density offset amplitude of each section, integrate the associated expression density data, and get the expression trend trajectory group EY T , wherein LY represents the sentence group length of the detected section, DY represents the number of sentence types of the detected section, MY represents the standard number of sentence types in the expression specification parameters, and YK represents the structure reference constant under the corresponding detection link.

[0055] The expression trend trajectory group is obtained by quantifying the proportional relationship between the sentence group length (LY), the number of sentence types (DY) and the expression specification standard (MY) in each section of the detected content, and correcting it with the structure reference constant (YK). It is a comprehensive distribution type parameter that reflects the change of the degree of consistency between students' expression method, sentence group structure and expression specification in the whole process of the detection task. It reflects the overall change trajectory of the expression density, sentence type use and specification matching status between different sections in the bilingual story retelling detection link, and provides a basis for analyzing expression coherence, complexity, specification and trend changes.

[0056] The length LY of the sentence group and the number DY of the sentence pattern of each segment in the detection content are extracted. The length of the sentence group refers to the number of complete sentences in a single segment. In this embodiment, the sampled segment contains 12 sentences, and LY is set to 12. The number of sentence patterns refers to the number of sentence pattern types with obvious structural differences in the segment. After annotation and confirmation, it is confirmed to be 4 types, and DY is set to 4. The number of standard sentence patterns in the expression specification parameter is set to 5 types according to the requirements of the language course, and MY is set to 5. The structure reference constant YK is the expected value of the expression structure reference, which is calculated based on expert evaluation and the average deviation of the structure balanced paragraph in the same type of text, and YK is set to 2.2. Because there are dimensional differences between different parameters, normalization processing is required. After normalization, LY=0.75, DY=0.80, MY=1.00, and YK=0.55. The normalized parameters are substituted into the formula:

[0057] If the set expression density deviation tolerance interval is [0, 0.10], the result shows that the number of expression trend trajectory groups EY T =0.05 is within the set expression density deviation tolerance interval, which means that the change range of the expression structure density of the current segment is relatively close to the standard structure, and there is no sharp fluctuation, and the expression structure remains relatively stable. This numerical result is one of the basic data in the expression trend trajectory group, which participates in the construction of the structure fluctuation distribution sequence in the multi-segment sequence. The expression deviation values obtained by each segment can be sorted in time sequence, and the change direction, change section and fluctuation range of the expression trend can be further deduced, and the expression trend trajectory group can be integrated accordingly, thereby supporting the positioning and configuration of subsequent intervention strategies and the adjustment of task rhythm, Please refer to Figure 6 The strategy configuration adjustment module includes: The prompt rhythm judgment sub-module compares the difference between the arrangement rhythm of each task in the teaching library and the distribution of the graphic intervention prompt based on the expression trend trajectory group, judges the density change of the semantic intervention prompt in each task interval, selects the abnormal interval with the best correlation with the segment turning point, and obtains the prompt interval deviation; The expression trend trajectory group is taken as an analysis benchmark, all task items in the teaching library are extracted, and their setting order and time points are recorded in sequence according to task numbers to form a teaching task rhythm sequence, for example, the arrangement time intervals of task 1 to task 5 are 3, 4, 2, 5 and 3 minutes respectively, the change amplitude of the time interval between adjacent tasks is calculated, the appearance scene and position of the graphic-text intervention prompt according to the graphic-text intervention prompt density distribution, such as the number of prompts inserted in each paragraph, the position of the prompt in the sentence, and whether it is a graphic-text joint prompt, are recorded, a prompt density sequence is constructed, the difference between the prompt density and the task rhythm sequence is calculated after pairing them according to the task numbers, if the intervention prompt is concentrated in the rhythm dense segment, it is recorded as a prompt aggregation type, if it is concentrated in the task sparse segment, it is marked as a prompt dispersion type, the time distance between each prompt and the task turning point is calculated, whether it is aligned with the sentence turning node is recorded, a strong correlation interval is set within 5 seconds, a weak correlation interval is more than 10 seconds, and an ordinary correlation interval is the middle interval, the area with a prompt offset distance more than 10 seconds in all task paragraphs is marked as an abnormal prompt area, the task number and the corresponding prompt offset time of the abnormal interval are output as the offset interval, and the task segment with the least coincidence with the sentence turning point and the prompt density mutation is selected and output as the prompt interval offset.

[0058] The intervention fusion evaluation submodule analyzes the distribution of the graphic-text intervention fusion density on the expression node of the sentence segment based on the prompt interval offset, judges the correspondence between the intervention configuration and the expression time segment and the node type, optimizes the intervention distribution order, and obtains the positioning frequency of the intervention configuration; The graphic-text intervention fusion density is processed by sentence level labeling, all prompt information covered sentences are read, the number of interventions in each sentence is counted, and the sentence is marked as a text, image or graphic-text composite intervention type, the start and end nodes of the intervention information and the sentence expression are arranged according to the time axis, whether the intervention content falls on the key expression node such as the beginning of the sentence group, the sentence type conversion place and the emotion turning point is judged, the logical node type of the sentence expression is extracted, the node type such as parallel conversion, progressive expression and reason explanation is coded, the frequency of the intervention configuration in the node type is matched, when the frequency of the intervention configuration in a certain type of node is higher than 3 times and the frequency of the intervention configuration in other nodes is less than or equal to 1 time, the configuration is marked as a concentrated type, the intervention order of the high-frequency concentrated paragraph is adjusted, the graphic-text intervention is preferentially moved to the expression change point or the beginning of the sentence type conversion, the order difference between the original position and the adjusted position is recorded as the intervention migration span, the final position of all adjusted intervention points in the sentence and the node type covered by the intervention points are counted, a positioning matrix of the intervention configuration and the node is formed, and the positioning frequency of the intervention configuration is output according to the proportion of the number of interventions in each type of node to the total number of intervention configurations.

[0059] The configuration offset calculation submodule is based on the intervention configuration positioning frequency, analyzes the graphic-text intervention fusion density, sentence type density ratio, word repetition frequency and speech expression type difference, compares the change process of structure arrangement and expression method, and obtains the intervention parameter offset amplitude; The joint analysis of various structural indicators in the expression process is as follows: first, the corresponding graphic-text intervention fusion density in each expression speech is extracted, the ratio value of the intervention number and the expression paragraph length is recorded, and it is judged whether there is an abnormal graphic-text prompt density in each paragraph, for example, if the paragraph length is 30 words and the intervention number is 5 times, the density is 0.167, and if it exceeds 0.2, it is considered as a high-density paragraph, and the prompt density threshold is set to 0.15 for distinction, then the sentence type density ratio is extracted, the proportion of different sentence type structures in each paragraph is counted, for example, if a paragraph contains 5 sentence type structures and a total of 10 sentences, the sentence type density is 0.5, and the reference sentence type density threshold is set to 0.4, the third item is the word repetition frequency, the ratio of the number of repeated key words in the paragraph to the total number of words is counted, if 10 key words are repeated in a paragraph of 40 words, the repetition frequency is 0.25, the fourth item is the speech expression type difference, it is judged whether the current paragraph is narrative, description, explanation, comment or dialogue, and the expression method is classified and numbered, each type is marked as an expression method code, the change trajectory of the expression method in all speeches is calculated, if the expression changes from “explanation” to “dialogue” and then to “description”, it is recorded as 2 times of continuous switching, then the combination changes of the above four indicators in each speech are compared horizontally, it is judged whether there is a structure mismatching segment with high prompt density, low sentence type density, high word repetition and sudden change of expression method, if there is a change trajectory that is not coordinated between structure and expression, the paragraph is marked as a parameter offset paragraph, the number of all offset paragraphs is counted and the proportion of the number of offset paragraphs in the total number of paragraphs is calculated, for example, if there are 7 offset paragraphs in a total of 30 paragraphs, the intervention parameter offset amplitude is 0.233, and the offset amplitude value is output.

[0060] The bilingual story understanding self-adaptive teaching method facing individual differences includes the following steps: S1: Based on the role conversion type bilingual text, the number of roles and the order of events are analyzed, the actual distribution of time sequence adverbs and character appellations in the restatement text is judged according to the student bilingual story restatement content, the correspondence between the student restatement order and the original text order is compared, the structure adaptation performance parameters are obtained according to the plot perception level grading data; S2: Based on the structure adaptation performance parameters, the reaction times of the key words involved in the student's bilingual plot question and answer process are calculated, the complete performance of the syntactic structure in the answer content is analyzed, the semantic retention of the bilingual interpretation link is compared, and the language understanding characteristic changes of the student are summarized to obtain the understanding characteristic offset amount; S3: Based on the understanding of the feature offset, filter the story segments with obvious structural differences in the picture-text combined story material library, compare the sentence pattern recognition performance and visual understanding lag in the bilingual listening and reading task, judge the distribution characteristics of the semantic connection breakpoints, analyze the differences between the story segment structure and the students' understanding performance, and obtain the structure matching distribution index; S4: Based on the structure matching distribution index, optimize the expression of the student bilingual story retelling detection process, count the frequency of vocabulary repetition in the detection content, measure the number of sentence pattern transformations, analyze the proportional changes of sentence group length and sentence pattern density, summarize the expression mode distribution trend, and obtain the expression trend trajectory group; S5: Based on the expression trend trajectory group, adjust the task arrangement rhythm in the teaching library, judge the setting of semantic intervention prompts in each interval, analyze the matching relationship between the picture-text intervention fusion density and the student expression trajectory, compare the positioning of the intervention configuration in the expression process, and obtain the intervention parameter offset amplitude.

[0061] The above is only a preferred embodiment of the present application, and is not intended to limit the present application in other forms. Any skilled person in the art can use the disclosed technical content to make changes or modifications to equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made to the above embodiments without departing from the technical solution content of the present application, in accordance with the technical essence of the present application, still belongs to the protection scope of the technical solution of the present application.

Claims

1. An individual difference oriented bilingual story comprehension adaptive teaching system, characterized in that, The system comprises: The structure adaptation analysis module analyzes the number of roles, the order of events and the student's retelling content based on the role conversion bilingual text, judges the actual distribution of the time sequence adverb and the character name, compares the order of the retelling and the original text, combines the perception level grading data, induces the language structure application difference, and obtains the structure adaptation performance parameters; The understanding feature induction module calculates the keyword response frequency in bilingual plot question and answer based on the structure adaptation performance parameters, analyzes the integrity of the syntactic structure, compares the semantic retention during bilingual interpretation, combines the type standard and the migration strength, induces the language understanding change, and obtains the understanding feature offset; The material matching evaluation module filters the fragments with structural differences in the material library based on the understanding feature offset, compares the sentence type identification performance and the visual understanding lag, judges the distribution of semantic connection interruption, analyzes the story structure difference, and obtains the structure matching distribution index; The expression trend tracking module optimizes the expression of the language segment in bilingual story retelling detection based on the structure matching distribution index, counts the frequency of vocabulary repetition and the number of sentence type conversion, analyzes the length and density changes of the sentence group, combines the expression specification parameters, and obtains the expression trend trajectory group. 2.The individual difference oriented bilingual story comprehension self-adaptive teaching system according to claim 1, characterized in that, The structure adaptation performance parameters include expression accuracy, language structure complexity and information integration level, the understanding feature offset includes semantic extraction sensitivity, knowledge migration amplitude and understanding stability, the structure matching distribution index includes material diversity distribution, structure fit degree and content correlation strength, and the expression trend trajectory group includes expression mode variation, content presentation trend and retelling coherence. 3.The individual difference oriented bilingual story comprehension self-adaptive teaching system according to claim 1, characterized in that, The structure adaptation analysis module comprises: The role distribution extraction submodule extracts the number of roles and their first appearance order in the text based on the role conversion bilingual text, identifies the character names one by one in combination with the student's retelling text, compares the appearance order of the character names in the original text and the retelling text, counts the frequency of the order dislocation of the first appearance of the roles between the two texts, and obtains the role order difference; The time sequence distribution comparison submodule calls the role order difference, detects all the time sequence adverbs in the student's retelling text, distinguishes their distribution range and usage frequency according to the paragraphs, arranges the adverbs used in the order of events in the original text, compares the paragraph coverage and mutual exclusion distribution of the two texts according to the time mark, and obtains the event order dislocation rate; The structure adaptation induction submodule analyzes the connection word concatenation frequency, syntactic nesting level, subject replacement frequency and sentence group switching density in the student's retelling content according to the event order dislocation rate, combines the plot perception level grading data, matches the features in the type interval, and obtains the structure adaptation performance parameters. 4.The individual difference oriented bilingual story comprehension self-adaptive teaching system according to claim 1, wherein, The understanding feature induction module comprises: The keyword response submodule counts the response of all plot keywords of the student in the bilingual plot question and answer task based on the structure adaptation performance parameters, distinguishes the difference between the task stages, classifies the position and repetition frequency of the keywords in each stage, and accumulates and totals the keyword reaction quantity, and obtains the total amount of keyword reaction quantity. The syntax structure detection submodule decomposes the sentence of the answer content based on the total amount of keyword reaction quantity, retrieves the integrity of the subject, predicate, object and modification components, discriminates the syntax structure coverage performance in stages according to the sentence structure type standard, and obtains a syntax structure integrity level coefficient; The semantic migration comparison submodule extracts bilingual mutual translation content based on the syntax structure integrity level coefficient, decomposes the target language vocabulary and sentence meaning information, compares the source language information retention performance, combines the migration strength and sentence change index, and obtains an understanding feature offset amount. 5.The individual difference oriented bilingual story comprehension self-adaptive teaching system according to claim 1, wherein, The material matching evaluation module includes: The segment screening submodule obtains the segment structure features of the picture-text combined story material library based on the understanding feature offset amount, collects the number of sentence pattern styles, the total number of information levels and the frequency of role perspective changes of each segment, analyzes the structure span in view of the differences in sentence pattern styles, information levels and perspective frequencies, and obtains a structure difference segment set by screening the segments meeting the conditions; The expression recognition submodule collects the recognition records of the students on the segments in the bilingual listening and reading task according to the structure difference segment set, extracts the correct frequency of sentence pattern recognition, the task reaction time length and the error expression type code of each segment, compares the recognition frequency with the reference standard data, analyzes the influence of the combination of reaction time length and error expression on understanding rhythm, and obtains a sentence pattern understanding lag set; The structure comparison submodule compares the visual pause distribution, the total number of narrative jumps and the understanding lag performance of each segment based on the sentence pattern understanding lag set, obtains the matching deviation degree, analyzes the corresponding degree with the evolution path of the original structure style of the segment, and obtains a structure matching distribution index. 6.The individual difference oriented bilingual story comprehension self-adaptive teaching system according to claim 1, wherein, The expression trend tracking module includes: The vocabulary repetition submodule extracts the key words in the detected content based on the structure matching distribution index, counts the appearance frequency of the vocabulary in each detected task sentence, distinguishes the distribution of the vocabulary in the different time sequence sentences, obtains the vocabulary repetition density level according to the differences in repetition times and vocabulary coverage in the sentences, and obtains the vocabulary repetition density level; The sentence pattern transformation submodule counts the number of sentence pattern types in each detected sentence based on the vocabulary repetition density level, distinguishes the change categories of the sentence structure between adjacent sentences, compares the use situation of the difference sentence pattern combinations, induces the sentence difference type and the use frequency, and obtains the sentence pattern transformation distribution difference; The expression proportion analysis submodule extracts the sentence group length and the number of sentence patterns of each sentence in the detected content based on the sentence pattern transformation distribution difference, compares each structure according to the expression specification parameters, calculates the expression density offset amplitude of each sentence, integrates the associated expression density data, and obtains an expression trend trajectory group. 7.The individual difference oriented bilingual story comprehension self-adaptive teaching system according to claim 1, wherein, The system further includes: The strategy configuration adjustment module adjusts the task arrangement rhythm, judges the semantic prompt interval setting, analyzes the matching between the picture-text intervention and the expression trajectory, compares the intervention configuration positioning, induces the parameter changes, and obtains an intervention parameter offset amplitude; The intervention parameter offset amplitude includes a task distribution adjustment amplitude, an intervention response sensitivity and a strategy matching level. 8.The individual difference oriented bilingual story comprehension self-adaptive teaching system according to claim 7, characterized in that, The strategy configuration adjustment module includes: The prompt rhythm judgment submodule analyzes the arrangement rhythm of each task in the teaching library based on the expression trend trajectory group, compares the difference between the distribution of the graphic-text intervention prompt, judges the density change of the semantic intervention prompt in each task interval, screens the abnormal interval with the best correlation with the semantic transition point, and obtains the prompt interval offset; The intervention fusion evaluation submodule analyzes the distribution of the graphic-text intervention fusion density on the expression node based on the prompt interval offset, judges the corresponding situation of the intervention configuration between the expression time period and the node type, optimizes the intervention distribution order, and obtains the intervention configuration positioning frequency; The configuration offset calculation submodule analyzes the graphic-text intervention fusion density, the sentence type density ratio, the word repetition frequency, and the expression type difference based on the intervention configuration positioning frequency, compares the change process of the structure arrangement and the expression method, and obtains the intervention parameter offset amplitude.

9. A bilingual story comprehension self-adaptive teaching method oriented to individual differences, characterized in that, The method is used to realize the individual difference-oriented bilingual story understanding adaptive teaching system of any one of claims 1-8, comprising the following steps: S1: Based on the role conversion type bilingual text, the number of roles and the order of events are analyzed, the actual distribution of the time sequence adverbs and the character appellations used by the students in the restatement text is judged in combination with the student bilingual story restatement content, the corresponding situation between the student restatement order and the original text order is compared, the structure adaptation performance parameters are obtained according to the plot perception level grading data; S2: Based on the structure adaptation performance parameters, the reaction times of the key words involved in the student's bilingual plot question and answer process are calculated, the complete performance of the syntactic structure in the answer content is analyzed, the semantic retention of the bilingual interpretation link is compared, the changes in the student's language understanding characteristics are summarized, and the understanding characteristic offset is obtained; S3: Based on the understanding characteristic offset, the story segments with obvious structure differences in the graphic-text combined story material library are screened, the sentence type recognition performance and visual understanding lag of the students in the bilingual listening and reading task are compared, the distribution characteristics of the semantic connection breakpoints are judged, and the differences between the story segment structure and the student understanding performance are analyzed, and the structure matching distribution index is obtained; S4: Based on the structure matching distribution index, the expression of the student bilingual story restatement in the detection process is optimized, the word repetition frequency in the detection content is counted, the number of sentence type transformations is measured, the ratio change of the sentence group length and the sentence type density is analyzed, the expression method distribution trend is summarized, and the expression trend trajectory group is obtained; S5: Based on the expression trend trajectory group, the arrangement rhythm of the tasks in the teaching library is adjusted, the setting of the semantic intervention prompt in each interval is judged, the matching relationship between the graphic-text intervention fusion density and the student expression trajectory is analyzed, the positioning of the intervention configuration in the expression process is compared, and the intervention parameter offset amplitude is obtained.

Images