English graded reading text simplification method and system based on multi-modal data fusion

Abstract

The invention discloses an English graded reading text simplification method and system based on multi-modal data fusion, and relates to the technical field of text simplification, and the method comprises the steps: obtaining historical answer data from historical text reading records, evaluating the English knowledge mastering level of a reader in a current period, and obtaining English mastering data; acquiring historical English mastering data of other readers, acquiring English mastering data of the reader, and evaluating reading reference values of other readers to the reader to obtain a reference reader; acquiring historical text reading data and historical English mastering data of a reference reader, acquiring English mastering data, and evaluating text suitability between a reading text in the historical text reading data and the reader to obtain target text data; and obtaining target text data of the reader, simplifying a target reading text in the target text data in combination with the English mastering data, and pushing the simplified target reading text to the reader.

Description

Technical Field

[0001] The present invention relates to the technical field of text simplification, and in particular to a method and system for simplifying English graded reading text based on multimodal data fusion. Background Art

[0002] Multimodal data fusion technology refers to the technology of integrating and analyzing information from different sources and in different forms. These modalities can include text, audio, images, transmitter data, etc. Its core is to use the complementarity between different modal data to make up for the limitations of single modal information. The application of multimodal data fusion technology in the field of English graded text simplification includes but is not limited to the following benefits: 1. Increase learners' interest in reading. Multimodal fusion can transform boring text content into content of interest to learn, thereby increasing learners' interest in reading; 2. Better cultural and emotional transmission. Multimodal data can better convey text background and emotional information, which is more helpful for learners to fully understand English.

[0003] In the traditional process of simplifying English graded reading texts, the reader's specific English level is determined based on the reader's grade or English test level. However, this method actually has a big problem. Even if different readers have roughly the same vocabulary, the specific words they know and the degree of their mastery of long and difficult sentences are different. The traditional method cannot grade readers according to their actual English level, and different readers at different stages have different English texts suitable for them. The existing methods can neither conduct personalized analysis of the reader's reading level, nor select appropriate English texts for them, nor modify and simplify English texts according to the reader's level. Not only will it cause readers to read English texts that are inconsistent with their actual abilities, it may even cause readers to have serious aversion and lose their initiative to read. Summary of the invention

[0004] The object of the present invention is to provide a method and system for simplifying English graded reading text based on multimodal data fusion, so as to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for simplifying English graded reading text based on multimodal data fusion, the method comprising: Step S100: obtaining the reader's historical text reading record, obtaining historical question answering data from the historical text reading record, evaluating the reader's English knowledge mastery level in the current period, and obtaining English mastery data; Step S200: Obtaining historical English mastery data of other readers, obtaining the English mastery data of the reader, evaluating the reading reference value of other readers to the reader, and obtaining reference readers; Step S300: Acquire historical text reading data and historical English mastery data of a reference reader, acquire English mastery data, evaluate the text adaptability between the reading text in the historical text reading data and the reader, and obtain target text data; Step S400: Obtain the reader's target text data, simplify the target reading text in the target text data in combination with the English mastery data, and push the simplified target reading text to the reader.

[0006] Furthermore, step S100 includes: Step S101: Acquire each historical text reading record of the reader, and acquire historical answer data from the historical text reading record, wherein the historical answer data includes the correct rate of each question in the reading text in the historical text reading record, and acquires several knowledge points involved in each question; Step S102: obtaining the knowledge points involved in a certain question among the questions, and if the correct rate of a certain question is greater than a preset correct rate threshold, the question is recorded as a characteristic question, and the historical text reading record is recorded as a marked historical text reading record of the knowledge point; Step S103: Evaluate the reader's English knowledge level in the current cycle. The specific evaluation process is as follows: Acquire the text contents involved in several characteristic questions in the historical text reading records, collect the words in the text contents, obtain the word set of the reader in the historical text reading records, and take the union of the word sets of the reader in each historical text reading record to obtain the mastered word set of the reader; Step S104: Obtain the total number of words A in the reading text in the historical text reading record sum , get the total reading time t of the reader in the historical text reading record sum , calculate the marked reading efficiency E=A of the historical text reading records sum / t sum , get the average value E of the marked reading efficiency of each historical text reading record △ , calculate the characteristic reading efficiency of historical text reading records γ = E / E △ , and normalize the feature reading efficiency; Get the total number of historical text reading records of questions containing knowledge points C sum , accumulate the characteristic reading efficiency of each marked historical text reading record of the knowledge point and divide it by the total number C sum , get the knowledge mastery value of the knowledge point; When the knowledge mastery value is greater than the preset knowledge mastery threshold, it is determined that the reader has mastered the knowledge point, and the knowledge point is recorded as the reader's mastered knowledge point; Obtain and aggregate the various knowledge points mastered by the readers to obtain the knowledge set mastered by the readers; Step S105: Obtain the knowledge set and vocabulary set mastered by the reader in the current cycle, and collect them to obtain the reader's English mastery data.

[0007] Further, step S200 includes: Step S201: obtaining the English level of each other reader and the reader in the reading platform, wherein the English level of each other reader in the current period is higher than that of the reader, obtaining the historical English mastery data of other readers in the reading platform in each historical period, and obtaining the mastery knowledge set and mastery word set of other readers from the historical English mastery data; Step S202: Evaluate the reading reference value of other readers to the reader. The specific evaluation process is as follows: Calculate the vocabulary similarity between other readers in each historical period, where the vocabulary similarity between other readers in the αth historical period is r α =(G´ α ∩G) / G´ α ∪G,G´ α is the mastered word set of other readers in the αth historical period, and G is the mastered word set of the reader; Step S203: setting a value β, where β≠0, obtaining various knowledge points in the English text reading on the reading platform, and obtaining the knowledge set mastered by the reader; When a knowledge point among various knowledge points is a mastered knowledge point in the reader's mastered knowledge set, the characteristic value of a knowledge point is assigned to 0. Otherwise, the characteristic value of a knowledge point is assigned to β. Based on the reader's mastered knowledge set, the characteristic values ​​of various knowledge points are obtained and aggregated to obtain the reader's characteristic knowledge vector H; Step S204: Obtain the characteristic knowledge vector H' of other readers in the αth historical period α , calculate the knowledge similarity value W between other readers in the αth historical period α : , Calculate the reading reference value P of other readers to readers in the αth historical period α =λ1×W α +λ2×r α , where λ1 and λ2 are respectively the preset first similarity coefficient and the second similarity coefficient, where λ1>0, λ2>0; Step S205: When reading the reference value P αIf the reading reference value is greater than a preset reading reference threshold, it is determined that other readers have reading reference value to the reader, and other readers are recorded as the reader's reference readers. The α+1th historical period is recorded as the reference historical period between the reference reader and the reader, and each reference reader of the reader is obtained.

[0008] Furthermore, step S300 includes: Step S301: obtaining a reference reader of a reader, and obtaining historical English mastery data of the reference reader in a number of historical periods after a reference historical period; The change ratio of the reference readers in several historical periods is calculated, where the change ratio L in the δth historical period among several historical periods is δ =[(G △ （sum,δ） -G △ （sum,δ-1） ) / G △ （sum,δ-1） ] / 2×[(S △ （sum,δ） -S △ （sum,δ-1） ) / S △ （sum,δ-1） ] / 2, where G △ （sum,δ-1） , G △ （sum,δ） are the total number of words in the mastery word set of the reference reader in the δ-1th historical period and the δth historical period, respectively. △ （sum,δ-1） , S △ （sum,δ） are the total number of mastered knowledge points in the mastered knowledge set of the reference readers in the δ-1th historical period and the δth historical period respectively; Obtain the average value of the change ratio of the reference readers in several historical periods. When the average value is greater than the preset average threshold, the reference readers will be retained, otherwise, they will be eliminated. Step S302: obtaining the historical text reading data of several reference readers in the reference historical period, wherein the historical text reading data includes the reading texts of the reference readers in the reference historical period, wherein the contents of the reading texts are text contents that have been artificially simplified; Step S303: evaluating the text adaptability between the reading texts in the historical text reading data of several reference readers and the readers. The specific evaluation process is as follows: The reading texts of several reference readers in the reference history period are obtained respectively, sorted and collected according to the reading time sequence, and a reference reading text set is obtained; Acquire several knowledge points and several words contained in each reading text in the reference reading text set, and collect them to obtain a marked knowledge point set and a marked word set of each reading text; Calculate the text adaptation value between each reading text and the reader, where the text adaptation value X between the qth reading text in each reading text and the reader is q =1 / 2×[(S∩Y q ) / Y q +(G∩K q ) / K q ], where S is the reader's mastered knowledge set, G is the reader's mastered vocabulary set, and Y q is the set of labeled words of the qth reading text, K q is the set of labeled knowledge points for the qth reading text; When the text fits the value X q If the value is greater than a preset text adaptation threshold, the reader is determined to have text adaptation with the qth reading text, and the qth reading text is recorded as the target reading text; Step S304: Acquire and aggregate target reading texts of the reader to obtain the target text data of the reader; In the above steps, reference readers are further screened by grasping the change ratio. This is because reference readers have English levels similar to those of the readers within a certain historical period, but among the reference readers, there are those whose English levels are constantly improving and those that are stagnant. In order to effectively improve the readers' English level, by grasping the change ratio, reference readers who have improved their English level through reading can be retained. Subsequently, only referring to the reading texts read by these reference readers can effectively improve the readers' English level, which is also convenient for the formulation of subsequent reading plans.

[0009] Furthermore, step S400 includes: Step S401: acquiring each target reading text from the target text data, and randomly selecting a number of target reading texts from each target reading text as reading materials for the reader in the current cycle; Step S402: simplifying a plurality of target reading texts according to the English mastery data of the readers, wherein the specific simplification process for the vth target reading text among the plurality of target reading texts is as follows: Read the labeled word set Y´ in the text for the vth target v and the set of labeled knowledge points K´ v To obtain, when the vth target reads several words in a certain paragraph of the text, which are not in the marked word set Y´ v When reading, select synonyms from the word set corresponding to the reader's English level for replacement; When a certain paragraph in the vth target reading text contains only a few knowledge points, it is not in the marked knowledge point set K´ v In the code, split the text content corresponding to a certain paragraph; When a certain paragraph in the vth target reading text contains several knowledge points that are not in the marked knowledge point set K´ v If a certain paragraph contains more than a certain number of words than a preset threshold, annotate the words in the paragraph that are not in the word set corresponding to the reader's English level, and split the paragraph into several short sentences according to the text content of the paragraph to simplify the text content of the paragraph; Step S403: obtaining a plurality of simplified target reading texts, and pushing them to readers in the current cycle.

[0010] In order to better implement the above method, an English graded reading text simplification system based on multimodal data fusion is also proposed, which includes an English level assessment module, a reference value assessment module, a text adaptation assessment module, and an intelligent simplification module; English proficiency assessment module, used to assess the reader's English knowledge mastery level in the current cycle and obtain English proficiency data; The reference value evaluation module is used to obtain the historical English mastery data of other readers, and evaluate the reading reference value of other readers to the reader in combination with the reader's English mastery data, and obtain the reference reader; A text adaptation evaluation module is used to obtain historical text reading data of a reference reader, evaluate the text adaptability between the reading text in the historical text reading data and the reader, and obtain target text data; The intelligent simplification module is used to obtain the target reading text in the target text data and perform intelligent simplification on the target reading text.

[0011] Furthermore, the English proficiency assessment module includes a marked historical text reading record unit and an English proficiency assessment unit; A marked historical text reading record unit is used to obtain a marked historical text reading record of a knowledge point involved in a certain question according to the accuracy rate of a certain question in the historical text reading record; The English proficiency assessment unit is used to assess the reader's English knowledge level in the current cycle and obtain the reader's English proficiency data.

[0012] Further, the reference value assessment module includes a reading reference value unit and a reference value assessment unit; The reading reference value unit is used to calculate the reading reference value of other readers to readers in each historical period; The reference value evaluation unit is used to evaluate the reading reference value of other readers to the reader according to the reading reference value, and obtain the reference reader.

[0013] Further, the text adaptation evaluation module includes an adaptation degree analysis unit and a text adaptation evaluation unit; The adaptation degree analysis unit is used to analyze the text adaptation degree between each reading text in the historical text reading data and the reader, and calculate the text adaptation value between each reading text and the reader; The text adaptation evaluation unit is used to evaluate the text adaptability between the reading text in the historical text reading data and the reader according to the text adaptation value to obtain the target text data.

[0014] Further, the intelligent simplification module includes an intelligent simplification unit; The intelligent simplification unit is used to intelligently simplify several target reading texts in the target text data according to the readers' English mastery data, and push them to the readers in the current cycle.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention realizes intelligent simplification of reading texts read by readers. Taking into account the English proficiency of different readers, it is impossible to simplify the reading texts read by readers without accurately understanding the English proficiency of readers. Therefore, firstly, the English proficiency of readers in the current period is evaluated based on the readers' historical text reading records, that is, the English mastery data. After obtaining the readers' English mastery data, reference readers with reference value are found by evaluating the similarity of the English proficiency between other readers in different historical periods, and analyzing whether the reading texts of the reference readers in the historical period are adapted to the readers, so as to obtain reading texts suitable for readers to read, because these reading texts are the reading texts that have been artificially simplified and adjusted, which greatly saves the time of text selection and simplification. Finally, considering that readers may not understand the content in the text, the content of the reading text is further simplified, which not only enables readers to read what is consistent with their actual ability, but also increases the initiative of readers to read, which is helpful to improve the readers' English proficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a method flow chart of the English graded reading text simplification method and system based on multimodal data fusion of the present invention; Figure 2 It is a module schematic diagram of the English graded reading text simplification method and system based on multimodal data fusion of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a method for simplifying English graded reading text based on multimodal data fusion, the method comprising: Step S100: obtaining the reader's historical text reading record, obtaining historical question answering data from the historical text reading record, evaluating the reader's English knowledge mastery level in the current period, and obtaining English mastery data; Wherein, step S100 includes: Step S101: Acquire each historical text reading record of the reader, and acquire historical answer data from the historical text reading record, wherein the historical answer data includes the correct rate of each question in the reading text in the historical text reading record, and acquires several knowledge points involved in each question; For example, several knowledge points include attributive clauses, subjunctive mood, etc.; Step S102: obtaining the knowledge points involved in a certain question among the questions, and if the correct rate of a certain question is greater than a preset correct rate threshold, the question is recorded as a characteristic question, and the historical text reading record is recorded as a marked historical text reading record of the knowledge point; Step S103: Evaluate the reader's English knowledge level in the current cycle. The specific evaluation process is as follows: Acquire the text contents involved in several characteristic questions in the historical text reading records, collect the words in the text contents, obtain the word set of the reader in the historical text reading records, and take the union of the word sets of the reader in each historical text reading record to obtain the mastered word set of the reader; Step S104: Obtain the total number of words A in the reading text in the historical text reading record sum , get the total reading time t of the reader in the historical text reading record sum , calculate the marked reading efficiency E=A of the historical text reading records sum / t sum , get the average value E of the marked reading efficiency of each historical text reading record △ , calculate the characteristic reading efficiency of historical text reading records γ = E / E △ , and normalize the feature reading efficiency; Get the total number of historical text reading records of questions containing knowledge points Csum , accumulate the characteristic reading efficiency of each marked historical text reading record of the knowledge point and divide it by the total number C sum , get the knowledge mastery value of the knowledge point; When the knowledge mastery value is greater than the preset knowledge mastery threshold, it is determined that the reader has mastered the knowledge point, and the knowledge point is recorded as the reader's mastered knowledge point; Obtain and aggregate the various knowledge points mastered by the readers to obtain the knowledge set mastered by the readers; Step S105: obtaining the knowledge set and vocabulary set mastered by the reader in the current cycle, and collecting them to obtain the English mastery data of the reader; Step S200: Obtaining historical English mastery data of other readers, obtaining the English mastery data of the reader, evaluating the reading reference value of other readers to the reader, and obtaining reference readers; Wherein, step S200 includes: Step S201: obtaining the English level of each other reader and the reader in the reading platform, wherein the English level of each other reader in the current period is higher than that of the reader, obtaining the historical English mastery data of other readers in the reading platform in each historical period, and obtaining the mastery knowledge set and mastery word set of other readers from the historical English mastery data; For example, English level here can include CEFR, TOEFL, IELTS, etc.; Step S202: Evaluate the reading reference value of other readers to the reader. The specific evaluation process is as follows: Calculate the vocabulary similarity between other readers in each historical period, where the vocabulary similarity between other readers in the αth historical period is r α =(G´ α ∩G) / G´ α ∪G,G´ α is the mastered word set of other readers in the αth historical period, and G is the mastered word set of the reader; Step S203: setting a value β, where β≠0, obtaining various knowledge points in the English text reading on the reading platform, and obtaining the knowledge set mastered by the reader; When a knowledge point among various knowledge points is a mastered knowledge point in the reader's mastered knowledge set, the characteristic value of a knowledge point is assigned to 0. Otherwise, the characteristic value of a knowledge point is assigned to β. Based on the reader's mastered knowledge set, the characteristic values ​​of various knowledge points are obtained and aggregated to obtain the reader's characteristic knowledge vector H; Step S204: Obtain the characteristic knowledge vector H' of other readers in the αth historical period α , calculate the knowledge similarity value W between other readers in the αth historical period α : , Calculate the reading reference value P of other readers to readers in the αth historical period α =λ1×W α +λ2×r α , where λ1 and λ2 are respectively the preset first similarity coefficient and the second similarity coefficient, where λ1>0, λ2>0; For example, the first similarity coefficient λ1 is 0.7, the second similarity coefficient λ2 is 0.3, the knowledge similarity value W1 between other readers in the first historical period is 0.8, and the vocabulary similarity value r1 between other readers in the first historical period is 0.9; Calculate the reading reference value of other readers to readers in the first historical period: P1 = 0.7 × 0.8 + 0.3 × 0.9 = 0.83; Step S205: When reading the reference value P α If the reading reference value is greater than the preset reading reference threshold, it is determined that other readers have reading reference value to the reader, and other readers are recorded as the reader's reference readers, and the α+1th historical period is recorded as the reference reader and the reader's reference historical period, and each reference reader of the reader is obtained; Step S300: Acquire historical text reading data and historical English mastery data of a reference reader, acquire English mastery data, evaluate the text adaptability between the reading text in the historical text reading data and the reader, and obtain target text data; Wherein, step S300 includes: Step S301: obtaining a reference reader of a reader, and obtaining historical English mastery data of the reference reader in a number of historical periods after a reference historical period; The change ratio of the reference readers in several historical periods is calculated, where the change ratio L in the δth historical period among several historical periods is δ =[(G △ （sum,δ） -G △ （sum,δ-1） ) / G △ （sum,δ-1） ] / 2×[(S △ （sum,δ） -S △ （sum,δ-1） ) / S △ （sum,δ-1） ] / 2, where G △ （sum,δ-1） , G △ （sum,δ） are the total number of words in the mastery word set of the reference reader in the δ-1th historical period and the δth historical period, respectively. △（sum,δ-1） , S △ （sum,δ） are the total number of mastered knowledge points in the mastered knowledge set of the reference readers in the δ-1th historical period and the δth historical period respectively; Obtain the average value of the change ratio of the reference readers in several historical periods. When the average value is greater than the preset average threshold, the reference readers will be retained, otherwise, they will be eliminated. Step S302: obtaining the historical text reading data of several reference readers in the reference historical period, wherein the historical text reading data includes the reading texts of the reference readers in the reference historical period, wherein the contents of the reading texts are text contents that have been artificially simplified; Step S303: evaluating the text adaptability between the reading texts in the historical text reading data of several reference readers and the readers. The specific evaluation process is as follows: The reading texts of several reference readers in the reference history period are obtained respectively, sorted and collected according to the reading time sequence, and a reference reading text set is obtained; Acquire several knowledge points and several words contained in each reading text in the reference reading text set, and collect them to obtain a marked knowledge point set and a marked word set of each reading text; Calculate the text adaptation value between each reading text and the reader, where the text adaptation value X between the qth reading text in each reading text and the reader is q =1 / 2×[(S∩Y q ) / Y q +(G∩K q ) / K q ], where S is the reader's mastered knowledge set, G is the reader's mastered vocabulary set, and Y q is the set of labeled words of the qth reading text, K q is the set of labeled knowledge points for the qth reading text; When the text fits the value X q If the value is greater than a preset text adaptation threshold, the reader is determined to have text adaptation with the qth reading text, and the qth reading text is recorded as the target reading text; Step S304: Acquire and aggregate target reading texts of the reader to obtain the target text data of the reader; Step S400: Obtain the reader's target text data, simplify the target reading text in the target text data in combination with the English mastery data, and push the simplified target reading text to the reader; Wherein, step S400 includes: Step S401: acquiring each target reading text from the target text data, and randomly selecting a number of target reading texts from each target reading text as reading materials for the reader in the current cycle; Step S402: simplifying a plurality of target reading texts according to the English mastery data of the readers, wherein the specific simplification process for the vth target reading text among the plurality of target reading texts is as follows: Read the labeled word set Y´ in the text for the vth target v and the set of labeled knowledge points K´ v To obtain, when the vth target reads several words in a certain paragraph of the text, which are not in the marked word set Y´ v When reading, select synonyms from the word set corresponding to the reader's English level for replacement; For example, in the target reading text, “utilize” can be replaced with “use”; When a certain paragraph in the vth target reading text contains only a few knowledge points, it is not in the marked knowledge point set K´ v In the code, split the text content corresponding to a certain paragraph; When a certain paragraph in the vth target reading text contains several knowledge points that are not in the marked knowledge point set K´ v If a certain paragraph contains more than a certain number of words than a preset threshold, annotate the words in the paragraph that are not in the word set corresponding to the reader's English level, and split the paragraph into several short sentences according to the text content of the paragraph to simplify the text content of the paragraph; Step S403: obtaining a plurality of simplified target reading texts, and pushing them to readers in the current cycle; In order to better implement the above method, an English graded reading text simplification system based on multimodal data fusion is also proposed, which includes an English level assessment module, a reference value assessment module, a text adaptation assessment module, and an intelligent simplification module; English proficiency assessment module, used to assess the reader's English knowledge mastery level in the current cycle and obtain English proficiency data; The reference value evaluation module is used to obtain the historical English mastery data of other readers, and evaluate the reading reference value of other readers to the reader in combination with the reader's English mastery data, and obtain the reference reader; A text adaptation evaluation module is used to obtain historical text reading data of a reference reader, evaluate the text adaptability between the reading text in the historical text reading data and the reader, and obtain target text data; An intelligent simplification module is used to obtain a target reading text in the target text data and perform intelligent simplification on the target reading text; Among them, the English proficiency assessment module includes a unit for marking historical text reading records and an English proficiency assessment unit; A marked historical text reading record unit is used to obtain a marked historical text reading record of a knowledge point involved in a certain question according to the accuracy rate of a certain question in the historical text reading record; English proficiency assessment unit, used to assess the reader's English knowledge level in the current cycle and obtain the reader's English proficiency data; Wherein, the reference value assessment module includes a reading reference value unit and a reference value assessment unit; The reading reference value unit is used to calculate the reading reference value of other readers to readers in each historical period; A reference value evaluation unit, used for evaluating the reading reference value of other readers to the reader according to the reading reference value, and obtaining a reference reader; Among them, the text adaptation evaluation module includes an adaptation degree analysis unit and a text adaptation evaluation unit; The adaptation degree analysis unit is used to analyze the text adaptation degree between each reading text in the historical text reading data and the reader, and calculate the text adaptation value between each reading text and the reader; A text adaptation evaluation unit, used to evaluate the text adaptability between the reading text in the historical text reading data and the reader according to the text adaptation value, and obtain target text data; Wherein, the intelligent simplification module includes an intelligent simplification unit; The intelligent simplification unit is used to intelligently simplify several target reading texts in the target text data according to the readers' English mastery data, and push them to the readers in the current cycle.

[0019] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for simplifying English graded reading text based on multimodal data fusion, characterized in that: The method comprises: Step S100: obtaining a reader's history text reading record, obtaining history question answering data from the history text reading record, evaluating the reader's English knowledge mastery level in the current period, and obtaining English mastery data; Step S200: Acquire the historical English mastery data of other readers, acquire the English mastery data of the reader, evaluate the reading reference value of the other readers to the reader, and obtain the reference reader; Step S300: Acquire the historical text reading data and historical English mastery data of the reference reader, acquire the English mastery data, evaluate the text adaptability between the reading text in the historical text reading data and the reader, and obtain target text data; Step S400: acquiring the target text data of the reader, simplifying the target reading text in the target text data in combination with the English mastery data, and pushing the simplified target reading text to the reader.

2. The English graded reading text simplification method based on multimodal data fusion according to claim 1 is characterized in that: The step S100 includes: Step S101: acquiring each historical text reading record of the reader, and acquiring historical answer data from the historical text reading record, wherein the historical answer data includes the correct rate of each question in the reading text in the historical text reading record, and acquiring a number of knowledge points involved in each question; Step S102: obtaining a knowledge point involved in a certain question among the questions, the correctness of the certain question is greater than a preset correctness threshold, recording the question as a characteristic question, and recording the historical text reading record as a marked historical text reading record of the knowledge point; Step S103: Evaluate the reader's English knowledge level in the current cycle. The specific evaluation process is as follows: Acquire the text contents involved in several characteristic topics in the historical text reading records, collect the words in the text contents, obtain the word set of the reader in the historical text reading records, and take the union of the word sets of the reader in the various historical text reading records to obtain the mastered word set of the reader; Step S104: Obtain the total number of words A in the reading text in the historical text reading record sum , obtain the total reading time t of the reader in the historical text reading record sum , calculate the marked reading efficiency E=A of the historical text reading record sum / t sum , obtain the average value E of the marked reading efficiency of each historical text reading record △ , calculate the characteristic reading efficiency γ=E / E of the historical text reading record △ , and normalize the characteristic reading efficiency; Get the total number C of historical text reading records of the questions containing the knowledge point sum , the characteristic reading efficiency of each marked historical text reading record of the knowledge point is accumulated and divided by the total number C sum , obtain the knowledge mastery value of the knowledge point; When the knowledge mastery value is greater than a preset knowledge mastery threshold, it is determined that the reader has mastered the knowledge point, and the knowledge point is recorded as the reader's mastered knowledge point; Acquire and aggregate various knowledge points mastered by the reader to obtain the mastered knowledge set of the reader; Step S105: Acquire the knowledge set and vocabulary set mastered by the reader in the current cycle, and collect them to obtain the English mastery data of the reader.

3. The English graded reading text simplification method based on multimodal data fusion according to claim 2 is characterized in that: The step S200 includes: Step S201: obtaining the English levels of other readers and the reader in the reading platform, wherein the English levels of other readers in the current period are higher than those of the reader, obtaining historical English mastery data of other readers in the reading platform in various historical periods, and obtaining mastery knowledge sets and mastery vocabulary sets of other readers from the historical English mastery data; Step S202: Evaluate the reading reference value of the other readers to the reader. The specific evaluation process is as follows: Calculate the vocabulary similarity value between the other readers and the reader in each historical period, where the vocabulary similarity value between the other readers and the reader in the αth historical period is r α =(G´ α ∩G) / G´ α ∪G,G´ α is the mastered word set of the other readers in the αth historical period, and G is the mastered word set of the reader; Step S203: setting a value β, where β≠0, obtaining various knowledge points in the English text reading on the reading platform, and obtaining the knowledge set mastered by the reader; When a certain knowledge point among the various knowledge points is a mastered knowledge point in the mastered knowledge set of the reader, the characteristic value of the certain knowledge point is assigned to 0, otherwise, the characteristic value of the certain knowledge point is assigned to β, and based on the mastered knowledge set of the reader, the characteristic values ​​of the various knowledge points are obtained and aggregated to obtain the characteristic knowledge vector H of the reader; Step S204: Obtain the characteristic knowledge vector H' of the other readers in the αth historical period α , calculate the knowledge similarity value W between the other readers and the reader in the αth historical period α : ， Calculate the reading reference value P of the other readers to the reader in the αth historical period α =λ1×W α +λ2×r α , where λ1 and λ2 are respectively the preset first similarity coefficient and the second similarity coefficient, where λ1>0, λ2>0; Step S205: When the reading reference value P α If the reading reference value is greater than a preset reading reference threshold, it is determined that the other readers have reading reference value to the reader, the other readers are recorded as reference readers of the reader, and the α+1th historical period is recorded as the reference historical period between the reference reader and the reader, so as to obtain each reference reader of the reader.

4. The English graded reading text simplification method based on multimodal data fusion according to claim 3 is characterized in that: The step S300 includes: Step S301: obtaining a reference reader of the reader, and obtaining historical English mastery data of the reference reader in several historical periods after a reference historical period; The mastery change ratio of the reference reader in the several historical periods is calculated, wherein the mastery change ratio L in the δth historical period among the several historical periods is δ =[(G △ （sum,δ） -G △ （sum,δ-1） ) / G △ （sum,δ-1） ] / 2×[(S △ （sum,δ） -S △ （sum,δ-1） ) / S △ （sum,δ-1） ] / 2, where G △ （sum,δ-1） , G △ （sum,δ） are the total number of words in the mastered word set of the reference reader in the δ-1th historical period and the δth historical period, respectively, △ （sum,δ-1） , S △ （sum,δ） are the total number of mastered knowledge points in the mastered knowledge set of the reference reader in the δ-1th historical period and the δth historical period respectively; Obtaining an average value of the mastery change ratio of the reference reader in the several historical periods, and retaining the reference reader when the average value is greater than a preset average threshold, otherwise, eliminating the reference reader; Step S302: obtaining the historical text reading data of several reference readers in the reference historical period, wherein the historical text reading data includes the reading text of the reference reader in the reference historical period, wherein the content of the reading text is the text content that has been artificially simplified; Step S303: evaluating the text compatibility between the reading texts in the historical text reading data of the reference readers and the readers, and the specific evaluation process is as follows: Respectively obtaining the reading texts of the reference readers within the reference history period, sorting and collecting them in the order of reading time, and obtaining a reference reading text set; Acquire several knowledge points and several words contained in each reading text in the reference reading text set, and collect them to obtain a marked knowledge point set and a marked word set of each reading text; Calculate the text adaptation value between each reading text and the reader, wherein the text adaptation value X between the qth reading text in each reading text and the reader is q =1 / 2×[(S∩Y q ) / Y q +(G∩K q ) / K q ], where S is the knowledge set mastered by the reader, G is the vocabulary set mastered by the reader, and Y q is the set of labeled words of the qth reading text, K q is a set of marked knowledge points of the qth reading text; When the text fits the value X q If the value is greater than a preset text adaptation threshold, it is determined that the reader and the qth reading text have text adaptation, and the qth reading text is recorded as a target reading text; Step S304: Acquire and aggregate the target reading texts of the reader to obtain the target text data of the reader.

5. The English graded reading text simplification method based on multimodal data fusion according to claim 4 is characterized in that: The step S400 includes: Step S401: acquiring each target reading text from the target text data, and randomly selecting a number of target reading texts from the target reading texts as reading materials for the reader in the current cycle; Step S402: simplifying the target reading texts according to the English mastery data of the reader, wherein the specific simplification process of the vth target reading text among the target reading texts is as follows: The set of labeled words Y´ in the vth target reading text v and the set of labeled knowledge points K´ v To obtain, when several words in a certain paragraph of the v-th target reading text are not in the marked word set Y´ v When the reader's English level is reached, a synonym is selected from the word set corresponding to the reader's English level for replacement; When the certain paragraph in the vth target reading text contains only a few knowledge points, which are not in the marked knowledge point set K´ v In the example, the text content corresponding to the certain paragraph is split; When several knowledge points contained in the certain paragraph of the v-th target reading text are not in the marked knowledge point set K´ v and the number of words in the certain paragraph exceeds a preset number threshold, annotating the words in the certain paragraph that are not in the word set corresponding to the reader's English level, and splitting the certain paragraph into a number of short sentences according to the text content of the certain paragraph to simplify the text content of the certain paragraph; Step S403: obtaining a plurality of simplified target reading texts, and pushing them to the readers in the current cycle.

6. An English graded reading text simplification system based on multimodal data fusion, used to execute the English graded reading text simplification method based on multimodal data fusion according to any one of claims 1 to 5, characterized in that: The system includes an English proficiency assessment module, a reference value assessment module, a text adaptation assessment module, and an intelligent simplification module; The English proficiency assessment module is used to assess the reader's English knowledge mastery level in the current cycle to obtain English mastery data; The reference value evaluation module is used to obtain the historical English mastery data of other readers, and evaluate the reading reference value of the other readers to the reader in combination with the English mastery data of the reader, so as to obtain the reference reader; The text adaptation evaluation module is used to obtain the historical text reading data of the reference reader, evaluate the text adaptability between the reading text in the historical text reading data and the reader, and obtain the target text data; The intelligent simplification module is used to obtain the target reading text in the target text data and perform text intelligent simplification on the target reading text.

7. The English graded reading text simplification system based on multimodal data fusion according to claim 6 is characterized in that: The English proficiency assessment module includes a marked historical text reading record unit and an English proficiency assessment unit; The marked historical text reading record unit is used to obtain the marked historical text reading record of the knowledge point involved in a certain question according to the accuracy rate of the certain question in the historical text reading record; The English proficiency assessment unit is used to assess the reader's English knowledge mastery level in the current cycle to obtain the reader's English proficiency data.

8. The English graded reading text simplification system based on multimodal data fusion according to claim 6 is characterized in that: The reference value assessment module includes a reference value reading unit and a reference value assessment unit; The reading reference value unit is used to calculate the reading reference values ​​of other readers to the reader in each historical period; The reference value evaluation unit is used to evaluate the reading reference value of the other readers to the reader according to the reading reference value, and obtain a reference reader.

9. The English graded reading text simplification system based on multimodal data fusion according to claim 6, characterized in that: The text adaptation evaluation module includes an adaptation degree analysis unit and a text adaptation evaluation unit; The adaptation degree analysis unit is used to analyze the text adaptation degree between each reading text in the historical text reading data and the reader, and calculate the text adaptation value between each reading text and the reader; The text adaptation evaluation unit is used to evaluate the text adaptability between the reading text in the historical text reading data and the reader according to the text adaptation value, so as to obtain target text data.

10. The English graded reading text simplification system based on multimodal data fusion according to claim 6, characterized in that: The intelligent simplification module includes an intelligent simplification unit; The intelligent simplification unit is used to intelligently simplify several target reading texts in the target text data according to the English mastery data of the reader, and push them to the reader in the current cycle.

Images