A method for checking plagiarism of an English short article

Abstract

The present application provides an English short article plagiarism checking method, which is an inspection model composed of sequentially connected English short article copy-paste plagiarism checking module, English short article rewriting plagiarism checking module and English short article plagiarism checking comprehensive evaluation module. After an English short article is processed by the inspection model, the plagiarism checking result and comments of the English short article can be finally obtained. The checking method solves the problem of automatic checking of English short article plagiarism.

Description

Technical Field

[0001] This invention relates to natural language processing technology, specifically a method for automatically checking for plagiarism in different English short passages using a computer system. Background Technology

[0002] The final similarity between English passages determines whether plagiarism exists. Existing plagiarism detection techniques mainly include string comparison methods, word frequency statistics methods, digital fingerprinting methods, and machine learning methods. String comparison and word frequency statistics methods require a full scan of both passages, resulting in high time complexity. Digital fingerprinting maps strings to represent numbers, and direct comparison of these numbers determines string similarity, offering lower time complexity than string comparison and word frequency statistics methods. However, this method may suffer from hash collisions during fingerprint calculation, affecting the accuracy of the results. Machine learning methods compare semantic similarity between English passages to determine plagiarism, but this method has high time complexity and low accuracy. Summary of the Invention

[0003] To address the aforementioned problems, this invention provides a method for detecting plagiarism in English short passages.

[0004] The present invention provides a method for checking plagiarism in English short passages, comprising: a sequentially connected English short passage copy-paste plagiarism checking module, an English short passage rewriting plagiarism checking module, and an English short passage plagiarism checking comprehensive evaluation module. The overall processing flow is as follows: Figure 1 As shown.

[0005] The processing flow of the English short passage copy-paste plagiarism detection module of this invention is as follows: First, input the full text of two English short passages and process them into sentences respectively; Second, use an N-gram model and select a five-gram window as needed, i.e., a five-gram model, and cut the sentences in the two English short passages to be checked into windows according to the direction of sentence extension; Third, perform numerical mapping on the content of the cut windows and map out the corresponding values ​​respectively; Fourth, match the window values ​​corresponding to the sentences in the two English short passages. If the values ​​match the same, the content of the cut window is salted and recombined, and numerical mapping is performed again. If the numerical matching of the numerical mapping is different, the cut window is slid in the direction of sentence extension. If the numerical matching of the numerical mapping is the same again, it jumps to the next cut window; Fifth, iteratively match the sentences in the two English short passages until both English short passages are matched.

[0006] The processing flow of the English short passage rewriting plagiarism detection module of the present invention is as follows: First, iterate through the non-copy-paste plagiarized English sentences in the two English short passages respectively, and perform word segmentation, spelling correction, removal of special symbols, and deletion of stop words on the traversed English sentences, and output the results after the operation is completed; Second, iterate through the English sentences in the two English short passages processed in the first step respectively, and concatenate the two English sentences; Third, map the concatenated English sentence into a distributed space to generate a vector representation of the English sentence; Fourth, extract the semantic features from the vector representation of the English sentence, and analyze whether the English sentence contains semantic plagiarism based on the semantic feature information; Fifth, iterate through the non-copy-paste plagiarized sentences in the two English short passages respectively, and analyze whether each pair of English sentences contains semantic plagiarism in turn.

[0007] The processing flow of the English short passage plagiarism detection and comprehensive evaluation module of the present invention is as follows: First, calculate the proportion of copied and pasted sentences and rewritten and plagiarized sentences in the English short passage to obtain the comprehensive plagiarism rate of the English short passage; Second, generate the coordinates of English words with copied and pasted content, the coordinates of English sentences with rewritten content, and the comprehensive plagiarism rate of the English short passage into a plagiarism detection result record table; Third, generate the plagiarism detection result and comments of the English short passage based on the coordinates of copied and pasted English words, the coordinates of rewritten English sentences, and the comprehensive plagiarism rate of the English short passage in the plagiarism detection result record table.

[0008] The specific processing flow of the English short passage copy-paste plagiarism detection module, the English short passage rewriting plagiarism detection module, and the English short passage plagiarism detection comprehensive evaluation module of the present invention is as follows.

[0009] like Figure 2 As shown, the specific processing flow of the English short passage copy-paste plagiarism detection module is as follows:

[0010] Starting from page 201;

[0011] Page 202 reads the full text of two short English passages respectively;

[0012] Page 203 uses regular expressions to segment the entire English passage into sentences and outputs the segmented results.

[0013] P204 reads the sentence segmentation results of the two English short articles after processing in step P203;

[0014] Page 205 shows the sentence segmentation results of the first English passage;

[0015] P206 performs a five-grammar model window segmentation of the first English passage sentence traversed;

[0016] P207 performs digital mapping on the segmented windows according to the calculation formula (1) and outputs the value of the first English short article window segmentation;

[0017] Page 208 shows the sentence segmentation results of the second English passage.

[0018] P209 performs a five-grammar model window segmentation of the sentences in the second English passage that is traversed;

[0019] P210 performs digital mapping on the segmented windows according to the calculation formula (1) and outputs the value of the segmentation of the second English short article window;

[0020] P211 matches the values ​​mapped in step P210 with the values ​​mapped in step P207;

[0021] P212 determines whether the numerical matching results in step P211 are the same. If they are the same, proceed to step P213; otherwise, proceed to step P216.

[0022] P213 salts and recombines the contents of two windows with the same numerical matching, and performs numerical mapping again according to the calculation formula (1), and matches the mapped values ​​again.

[0023] P214 checks whether the numerical matching results in step P213 are the same. If they are the same, proceed to step P215; otherwise, proceed to step P216.

[0024] P215 saves the clipping window where numerical values ​​match in the traversed sentences;

[0025] On page 216, determine whether the traversal of the clipping window in the second English passage has ended. If the traversal has ended, proceed to step P217; otherwise, proceed to step P210.

[0026] P217 determines whether the copy-paste degree value of the sentences traversed in the second English short passage exceeds 0.5 according to calculation formula (2) and calculation formula (3). If it exceeds the threshold, the position information of the window is saved to the copy-paste position information file and P218 is executed. Otherwise, P218 is executed directly.

[0027] On page 218, determine whether the traversal of sentences in the second English passage has ended. If the traversal has ended, proceed to step P219; otherwise, proceed to step P208.

[0028] On page 219, determine whether the traversal of the clipping window in the first English passage has ended. If the traversal has ended, proceed to step P220; otherwise, proceed to step P207.

[0029] P220 determines whether the copy-paste degree value of the sentences traversed in the first English short article exceeds 0.5 according to calculation formula (2) and calculation formula (3). If it exceeds the threshold, the position information of the window is saved to the copy-paste position information file and P221 is executed. Otherwise, P221 is executed directly.

[0030] On page 221, determine whether the traversal of the first English passage has ended. If the traversal has ended, proceed to step P222; otherwise, proceed to step P205.

[0031] End of page 222.

[0032] like Figure 3 As shown, the specific processing flow of the English short passage rewriting plagiarism detection module is as follows:

[0033] Starting from P301;

[0034] On page 302, read out the non-copy-paste plagiarized sentences from the two short English passages respectively;

[0035] Page 303 shows the results of iterating through the sentences in the first English passage that are not copied and pasted.

[0036] Page 304 shows the results of iterating through the sentences in the second English passage that are not copied and pasted.

[0037] P305 performs word segmentation on the English sentences in the two English short articles iterates through, and outputs the word segmentation results of the English sentences respectively;

[0038] P306 uses regular expressions to match the stop word list and performs stop word removal on two English sentences respectively;

[0039] P307 performs spelling correction on the remaining words in two English sentences after removing stop words and outputs the two final processed English sentences.

[0040] P308 will concatenate the two processed English sentences and output the concatenation result;

[0041] P309 represents the concatenation results output in step P308 as sentence vectors;

[0042] P310 uses a bidirectional gated loop encoder to encode the output of step P309;

[0043] P311 decodes based on the output of step P310 and the similarity category;

[0044] P312 classifies sentences based on semantic similarity between them;

[0045] P313 Determine whether the classification results are semantically similar. If they are similar, proceed to step P314; otherwise, proceed to step P315.

[0046] P314 records the location information of two semantically similar sentences;

[0047] On page 315, determine whether the sentence results of the second English passage have been traversed. If the traversal is complete, proceed to step P316; otherwise, proceed to step P304.

[0048] On page 316, determine whether the sentence results of the first English passage have been traversed. If the traversal is complete, proceed to step P317; otherwise, proceed to step P303.

[0049] P317 outputs all sentences in the final two English passages that contain text rewriting and plagiarism, and saves the location information of the rewritten and plagiarized sentences to the text rewriting and plagiarism location information file.

[0050] End of P318.

[0051] like Figure 4 As shown, the specific processing flow of the comprehensive evaluation module for checking plagiarism in English short passages is as follows:

[0052] Starting from P401;

[0053] P402 reads the location information of copied and pasted sentences in two short English passages;

[0054] P403 reads the location information of the plagiarized sentences in two short English articles;

[0055] P404 calculates and outputs the combined plagiarism rate of the two English short articles according to the calculation formula (4);

[0056] P405 determines whether the output result of step P404 is plagiarized based on the calculation formula (5). If plagiarism is found, proceed to step P406; otherwise, proceed to step P407.

[0057] P406 generates a comprehensive evaluation of plagiarism detection for English short passages and saves it to the English short passage plagiarism evaluation result file;

[0058] End of P407.

[0059] The calculation formulas involved in the English short passage plagiarism detection method of this invention are defined as follows:

[0060] 1. Formula for calculating the mapping of strings to numerical values

[0061]

[0062] In the calculation formula (1), n ​​represents the character length in the English short passage clipping window, and i represents the character position in the clipping window.

[0063] 2. Formula for calculating the degree of copy-pasting in English short passages

[0064]

[0065] In the calculation formula (2), n represents the total number of English words in the English short passage, and i represents the i-th English word in the sentence that is copied and pasted.

[0066] 3. Formula for calculating the plagiarism tendency of copying and pasting sentences in English short passages

[0067]

[0068] In the calculation formula (3), the degree of copying and pasting of English short sentences is calculated by the calculation formula (2).

[0069] 4. Formula for calculating the overall plagiarism rate of English short passages

[0070]

[0071] In the calculation formula (4), i represents the i-th sentence in the English passage that is copied and pasted, n represents the total number of sentences in the English passage that are copied and pasted, j represents the j-th sentence in the English passage that is rewritten and copied, m represents the total number of sentences in the English passage that are rewritten and copied, and 0.7 is the weight parameter for rewriting and copying detection.

[0072] 5. Formula for calculating plagiarism tendency in English short passages

[0073]

[0074] In formula (5), the overall plagiarism rate of English short passages is calculated by formula (4).

[0075] After an English short passage is processed by the plagiarism detection model of this invention, the final result and comments on the plagiarism detection of the English short passage can be obtained. The detection method of this invention solves the problem of automatic detection of plagiarism in English short passages with high accuracy. Attached Figure Description

[0076] Figure 1 This is a flowchart illustrating the overall processing flow of the method of the present invention;

[0077] Figure 2 This is a flowchart of the English short text copy-paste plagiarism detection module processing method of the present invention;

[0078] Figure 3This is a flowchart of the English short passage rewriting plagiarism detection module processing method of the present invention;

[0079] Figure 4 This is a flowchart of the English short passage plagiarism detection and comprehensive evaluation module of the method of the present invention. Detailed Implementation

[0080] The specific implementation method of the English short passage plagiarism detection method of the present invention consists of the following five steps.

[0081] First step: Execute the "English Short Passage Copy and Paste Plagiarism Check Module"

[0082] In this embodiment of the invention, the English short passage copy-paste plagiarism detection module first segments the entire text of the two English short passages to be checked into sentences, then segments the sentences into windows, and performs numerical mapping on the segmented windows. The module matches the window mapping values ​​of the corresponding sentences in the two English short passages. If the values ​​match, the content of the window segment is salted and recombined, and numerical mapping is performed again. If the numerical matching is the same again, the module jumps to the next segmented window. When the degree of copy-paste in a sentence exceeds 0.5, the position of the corresponding word in the English short passage is saved, and the sentence in that English short passage is recorded as having copy-paste plagiarism. Finally, the information on copy-paste plagiarism in the two English short passages is saved to a copy-paste plagiarism location information file. In this embodiment of the invention, the two input English short passages are taken from the English short passage titled "My View on NEET," and the contents of the two English short passages are as follows:

[0083] The plagiarized English short essay

[0084] In recent years, under the pretext of employment difficulty, some students return home to live with their parents after graduation. While their parents are busy at work, they idle away their time day after day, least concerned with finding a job to have their own source of income. Worse still, they exhaust every penny of their parents and even grandparents to buy an apartment for their own enjoyment. In doing so, they even plunge their parents into serious debts.

[0085] This group of people has been named the "ken tribe". This phenomenon arises from long-term over-spoiling by parents and the inertia of the students themselves. Students should immediately stop the "ken" behavior and become independent and self-reliant.

[0086] Plagiarized English essay

[0087] 近年来,部分学生毕业后以就业难为借口回到老家和父母一起生活。在父母忙于工作时,他们日复一日虚度光阴,完全不主动找工作获取自己的收入。更糟糕的是,他们花光父母甚至祖父母的钱买房来享乐,甚至还让父母背负巨额债务。

[0088] 这类学生被赋予了一个称呼——“啃老族”。这是父母溺爱和学生自身懒惰共同造成的结果。因此,学生应当立即行动,停止啃老行为,成为独立自主、自力更生的人。

[0089] (1) Carry out clause segmentation processing on the full text of the two English short passages, and the results are shown below:

[0090] Plagiarized English short passage

[0091] [[In recent years, under the pretext of difficulty in finding employment, some students return home to live with their parents after graduation.]-1, [When their parents are busy with work, they idle away day after day, and do not care about finding a job to obtain their own source of income at all.]-2, [Even worse, they use up all the savings of their parents and even grandparents to buy a house for their own pleasure.]-3, [In this regard, they even push their parents into heavy debt.]-4, [This group of students has been given a name - "the Elderly-Devouring Community".]-5, [The emergence of this group is the result of long-term excessive doting from parents and the laziness of the students themselves.]-6, [Students should take immediate actions to stop this behavior of eating the elderly, and become independent and self-reliant.]-7]

[0092] Plagiarized English short passage

[0093] [[In recent years, some students return to their hometown to live with their parents after graduation on the pretext of difficult employment.]-1,[While their parents are busy at work, they idle away their time day after day, least concerned with finding a job to have their own source of income.]-2,[What's worse, they spend all their parents' and even grandparents' money to buy a house in order to enjoy themselves.]-3,[Thus, they even plunge their parents into serious debts.]-4,[This group of students have come to acquire a name-"the Elderly-Devouring Community".]-5,[This is the result of parents' coddling and students' laziness.]-6,[So, Students should take immediate actions to terminate their elderly-devouring behavior and become independent and self-reliant.]-7]

[0094] (2) Perform window clipping processing on the sentences in the full text of the two English short passages, and the results are shown as follows:

[0095] Plagiarized English short passage

[0096] Sentence 1[[In recent years,under],[recent years,under the],[years,underthe pretext],[,under the pretext of],[under the pretext of employment],[thepretext of employment difficulty],[pretext of employment difficulty,],[ofemployment difficulty,some],[employment difficulty,some students],[difficulty,some students return],[,some students return home],[some studentsreturn home to],[students return home to live],[return home to live with],[home to live with their],[to live with their parents],[live with theirparents after],[with their parents after graduation],[their parents aftergraduation.]]

[0097] Sentence 2[[While their parents are busy],[their parents are busy at],[parents are busy at work],[are busy at work,],[busy at work,they],[at work,they idle],[work,they idle away],[,they idle away their],[they idle away their time],[idle away their time day],[away their time day after],[their time day after day],[time day after day,],[day after day,least],[after day,least concerned],[day,least concerned with],[,least concerned with finding],[least concerned with finding a],[concerned with finding a job],[with finding a job to],[finding a job to have],[a job to have their],[job to have their own],[to have their own source],[have their own source of],[their own source of income],[own source of income.]]

[0098] ……

[0099] Sentence 7[[Students should take immediate actions],[should take immediateactions to],[take immediate actions to terminate],[immediate actions toterminate their],[actions to terminate their elderly],[to terminate theirelderly-],[terminate their elderly-devouring],[their elderly-devouringbehavior],[elderly-devouring behavior and],[-devouring behavior and become],[devouring behavior and become independent],[behavior and become independentand],[and become independent and self],[become independent and self-],[independent and self-reliant],[and self-reliant.]]

[0100] Plagiarized English short passage

[0101] Sentence 1[[In recent years,some],[recent years,some students],[years,somestudents return],[,some students return to],[some students return to their],[students return to their hometown],[return to their hometown to],[to theirhometown to live],[their hometown to live with],[hometown to live withtheir],[to live with their parents],[live with their parents after],[withtheir parents after graduation],[their parents after graduation on],[parentsafter graduation on the],[after graduation on the pretext],[graduation on thepretext of],[on the pretext of difficult],[the pretext of difficultemployment],[pretext of difficult employment.]]

[0102] Sentence 2[[While their parents are busy],[their parents are busy at],[parents are busy at work],[are busy at work,],[busy at work,they],[at work,they idle],[work,they idle away],[,they idle away their],[they idle away their time],[idle away their time day],[away their time day after],[their time day after day],[time day after day,],[day after day,least],[after day,least concerned],[day,least concerned with],[,least concerned with finding],[least concerned with finding a],[concerned with finding a job],[with finding a job to],[finding a job to have],[a job to have their],[job to have their own],[to have their own source],[have their own source of],[their own source of income],[own source of income.]]

[0103] ……

[0104] Sentence 7 [[So, Students should take],[Students should take immediate],[Students should take immediate actions],[should take immediate actions to],[take immediate actions to terminate],[immediate actions to terminate their],[actions to terminate their elderly],[to terminate their elderly-],[terminatetheir elderly-devouring],[their elderly-devouring behavior],[elderly-devouring behavior and],[-devouring behavior and become],[devouring behaviorand become independent],[behavior and become independent and],[and becomeindependent and self],[become independent and self-],[independent and self-reliant],[and self-reliant.]]

[0105] (3) The clipping windows of the two English short articles were digitally mapped, and the results are shown below:

[0106] The plagiarized English short essay

[0107] Sentence 1 [1851440784,-2059393946,-1091057125,-1442870670,-1604089266,-1984342907,856699202,1270645474,230275231,1143971589,1119327231,-1450840376,271515544,-598029626,2029264496,1739868024,-691032551,157920521,1521160893]

[0108] Sentence 2 [-687514255,1979304337,1625963578,635602255,-1010508251,1406138774,-1302102171,-2058999056,141502705,-367834891,-1835445187,-1957578681,504024973,1 13423657,-839947380,635061270,-713755525,1185268616,1224769742,-1706589906,-229887756,-978916795,2054952362,295904334,738071780,1964291757,-1448512011]

[0109] ...

[0110] Sentence 7 [-1514319003,1950383792,-205350089,1752465499,-2127324872,355182162,980058623,330307307,743844611,-218432723,-1460195360,966448242,1647615704,-1752371907,-1631747833,-143813368,529359664,-811960582,-676698390,220036265]

[0111] Plagiarized English short essay

[0112] Sentence 1 [1583686204,-1824006655,-856414070,-1411123813,-480985207,-1724307756,689250767,-1057045619,-1890170466,1632202270,1739868024,-691032551,157920521,-88650448,798881863,-1234262386,485852389,-202131275,-1083483656,952489973]

[0113] Sentence 2 [-687514255,1979304337,1625963578,635602255,-1010508251,1406138774,-1302102171,-2058999056,141502705,-367834891,-1835445187,-1957578681,504024973,1 13423657,-839947380,635061270,-713755525,1185268616,1224769742,-1706589906,-229887756,-978916795,2054952362,295904334,738071780,1964291757,-1448512011]

[0114] ...

[0115] Sentence 7 [1265111888,355595571,1924389815,481490241,-301370738,1961425042,1148146888,-1752371907,-1631747833,-143813368,529359664,-811960582,-676698390,220036265,-20224038,834388950,620935779,597782223,557266165,1241242444]

[0116] (4) For the window content with the same numerical matching in the full text of the two English short articles, salting and re-mapping were performed, and the results are shown below:

[0117] The plagiarized English short essay

[0118] Same fragments of sentence 1 [[to live with their parents], [live with their parentsafter], [with their parents after graduation]]

[0119] The numerical mapping results for the same segment are [1739868024, -691032551, 157920521].

[0120] Salt-assisted reorganization of identical segments[[rth rnsttvaeiwi ep othiel],[t e patrietrrefthiiwavelh sn],[itoret epdagasaaunttihrrt irhfe nw]]

[0121] Result of the second digital mapping[1509780928,-1301579601,-338578603]

[0122] Identical segments in sentence 2[[While their parents are busy],[their parents are busyat],[parents are busy at work],[are busy at work,],[busy at work,they],[atwork,they idle],[work,they idle away],[,they idle away their],[they idle awaytheir time],[idle away their time day],[away their time day after],[theirtime day after day],[time day after day,],[day after day,least],[after day,least concerned],[day,least concerned with],[,least concerned with finding],[least concerned with finding a],[concerned with finding a job],[with findinga job to],[finding a job to have],[a job to have their],[job to have theirown],[to have their own source],[have their own source of],[their own sourceof income],[own source of income.]]

[0123] The numerical mapping results for the same segment are: [-687514255,1979304337,1625963578,635602255,-1010508251,1406138774,-1302102171,-2058999056,141502705,-367834891,-1835445187,-1957578681,50402497] 3,113423657,-839947380,635061270,-713755525,1185268616,1224769742,-1706589906,-229887756,-978916795,2054952362,295904334,738071780,1964291757,-1448512011]

[0124] Salt-induced recombination of identical fragments[[ethinsWe rhiaersbtaeu y lrp],[their parents arebusy at],[parents are busy at work],[are busy at work,],[busy at work,they],[at work,they idle],[work,they idle away],[,they idle away their],[they idleaway their time],[idle away their time day],[away their time day after],[their time day after day],[time day after day,],[day after day,least],[afterday,least concerned],[day,least concerned with],[,least concerned withfinding],[least concerned with finding a],[concerned with finding a job],[with finding a job to],[finding a job to have],[a job to have their],[job tohave their own],[to have their own source],[have their own source of],[theirown source of income],[own source of income.]]

[0125] The results of the digital mapping again are: [-716232339,-1323029135,-420225958,-617693841,-1207981563,1208665462,-2095393979,-881241328,1134136529,1880942869,-842811363,291199079,-1322680403,-6 79868151,-401582228,1341938742,1921907867,-474035288,1929187566,1795085582,-1023179564,-1176390107,1261660554,-1750285202,-1308117756,-1623798067,-270754283]

[0126] ...

[0127] Plagiarized English short essay

[0128] Same fragments of sentence 1 [[to live with their parents], [live with their parentsafter], [with their parents after graduation]]

[0129] The hash mapping result for the same fragment is [1739868024, -691032551, 157920521].

[0130] Recombining identical fragments with salt: [[rth rnsttvaeiwi ep othiel],[te patrietrrefthiiwavelh sn],[itoret epdagasaaunttihrrt irhfe nw]]

[0131] The result of the digital mapping is [1509780928, -1301579601, -338578603].

[0132] Sentence 2 identical fragments[[While their parents are busy],[their parents are busyat],[parents are busy at work],[are busy at work,],[busy at work,they],[atwork,they idle],[work,they idle away],[,they idle away their],[they idle awaytheir time],[idle away their time day],[away their time day after],[theirtime day after day],[time day after day,],[day after day,least],[after day,least concerned],[day,least concerned with],[,least concerned with finding],[least concerned with finding a],[concerned with finding a job],[with findinga job to],[finding a job to have],[a job to have their],[job to have theirown],[to have their own source],[have their own source of],[their own sourceof income],[own source of income.]]

[0133] The numerical mapping results for the same segment are: [-687514255,1979304337,1625963578,635602255,-1010508251,1406138774,-1302102171,-2058999056,141502705,-367834891,-1835445187,-1957578681,50402497] 3,113423657,-839947380,635061270,-713755525,1185268616,1224769742,-1706589906,-229887756,-978916795,2054952362,295904334,738071780,1964291757,-1448512011]

[0134] Salting recombination of identical fragments[[ethinsWe rhiaersbtaeu y lrp],[their parents arebusy at],[parents are busy at work],[are busy at work,],[busy at work,they],[at work,they idle],[work,they idle away],[,they idle away their],[they idleaway their time],[idle away their time day],[away their time day after],[their time day after day],[time day after day,],[day after day,least],[afterday,least concerned],[day,least concerned with],[,least concerned withfinding],[least concerned with finding a],[concerned with finding a job],[with finding a job to],[finding a job to have],[a job to have their],[job tohave their own],[to have their own source],[have their own source of],[theirown source of income],[own source of income.]]

[0135] The results of the digital mapping again are: [-716232339,-1323029135,-420225958,-617693841,-1207981563,1208665462,-2095393979,-881241328,1134136529,1880942869,-842811363,291199079,-1322680403,-6 79868151,-401582228,1341938742,1921907867,-474035288,1929187566,1795085582,-1023179564,-1176390107,1261660554,-1750285202,-1308117756,-1623798067,-270754283]

[0136] ...

[0137] (5) Save the position information of sentences and words in the two English short passages whose copy-paste score exceeds 0.5 to a copy-paste position information file. The result is shown below:

[0138] The plagiarized English short essay

[0139] 1 / [15,16,17,18,19,20,21]^2 / [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]^4 / [3,4,5,6,7,8,9,1 0,11,12]^5 / [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]^6 / [0,8,9,23,28]^7 / [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19]^

[0140] Plagiarized English short essay

[0141] 1 / [10,11,12,13,14,15,16]^2 / [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]^4 / [1,2,3,4,5,6,7,8,9 ,10]^5 / [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]^6 / [0,3,4,5,6,7,8,9]^8 / [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21]^.

[0142] Step Two: Execute the "English Passage Rewriting Plagiarism Check Module"

[0143] The English short passage rewriting plagiarism detection module checks for non-copy-paste plagiarized sentences in English short passages. This step first preprocesses the two English sentences to be matched, then maps the preprocessed sentences into a distributed space to generate English sentence vector representations. Next, encoding and decoding operations are performed on these sentence vector representations, and the concatenated sentences are classified based on semantic similarity according to the decoding results. Finally, the positional information of sentence pairs with semantic similarity is saved to a text rewriting plagiarism positional information file. The analysis results of this implementation are as follows:

[0144] (1) Read sentences from the English passage that are not copied and pasted. The results are shown below:

[0145] The plagiarized English short essay

[0146] [[Worse still, they exhaust every penny of their parents and evengrandparents to buy an apartment for their own enjoyment.]-3, [This community has come into existence as a result of long-time over-indulgence on the partof parents and the indolence on the part of students themselves.]-6]

[0147] Plagiarized English short essay

[0148] [[What's worse, they spend all their parents'and even grandparents'money to buy a house in order to enjoy themselves.]-3,[This is the result of parents'coddling and students'laziness.]-6]

[0149] (2) The English short passage sentences from step (1) are segmented and stop words are removed. The results are shown below:

[0150] The plagiarized English short essay

[0151] Sentence 3 [[Worse],[still],[exhaust],[penny],[parents],[grandparents],[buy],[apartment],[enjoyment]]

[0152] Sentence 6[[community],[existence],[result],[time],[indulgence],[parents],[indolence],[students]]

[0153] Plagiarized English short essay

[0154] Sentence 3 [[worse],[spend],[parents'],[grandparents'],[money],[buy],[house],[enjoy]]

[0155] Sentence 6[[result],[parents'],[coddling],[students'],[laziness]]

[0156] (3) The word segmentation results from step (2) are processed for word spelling correction, and the results are shown below:

[0157] The plagiarized English short essay

[0158] Sentence 3 [[Worse],[still],[exhaust],[penny],[parents],[grandparents],[buy],[apartment],[enjoyment]]

[0159] Sentence 6[[community],[existence],[result],[time],[indulgence],[parents],[indolence],[students]]

[0160] Plagiarized English short essay

[0161] Sentence 3 [[worse],[spend],[parents'],[grandparents'],[money],[buy],[house],[enjoy]]

[0162] Sentence 6[[result],[parents'],[coddling],[students'],[laziness]]

[0163] (4) The correction results from step (3) are processed by sentence splicing, and the results are shown below:

[0164] Check sentence pair 1

[0165] [[Worse],[still],[exhaust],[penny],[parents],[grandparents],[buy],[apartment],[enjoyment],[worse],[spend],[parents'],[grandparents'],[money],[buy],[house],[enjoy]]

[0166] Check sentence pair 2

[0167] [[Worse],[still],[exhaust],[penny],[parents],[grandparents],[buy],[apartment],[enjoyment]

[0168] Check sentence pair 3

[0169] [[community],[existence],[result],[time],[indulgence],[parents],[indolence],[students],[worse],[spend],[parents'],[grandparents'],[money],[buy],[house],[enjoy]]

[0170] Check sentence pair 4

[0171] [[community],[existence],[result],[time],[indulgence],[parents],[indolence],[students],[result],[parents'],[coddling],[students'],[laziness]]

[0172] (5) The output words in step (4) are processed into 300-dimensional vector representations, and the results are shown below: (The number of columns in the vector representation of each sentence pair is determined by the number of words in the sentence pair, and the word dimension is 300-dimensional)

[0173] Check sentence pair 1 [...]

[0174] Check sentence pair 2 [...]

[0175] Check sentence pair 3 [...]

[0176] Check the sentence pair 4[[-0.26584,-0.38287,0.13745,-0.34137,0.49913,-0.39343,-0.36459,0.042284,0.12432,-1.3025,0.079668,0.016803,0.18624,0.34 975, 0.4584, 0.11253, -0.17025, 0.01747, 0.14395, -0.18153, 0.18186, 0.09162, 0.27202, 0.33658, -0.50694, 0.082877, -0.14796, 0.41669, -0.4 6604, 0.4833, 0.35791, 0.12277, -0.41215, 0.56583, -0.20488, 0.28528, -0.01814, 0.52943, -0.067425, -0.43782, -0.20926, -0.18644, 0.31697, 0.18412, 0.71971, 0.3551, 0.19574, 0.20682, 0.21684, 0.26609, -0.071323, 0.075019, 0.24228, -0.10614, -0.21407, 0.022161, -0.10246, 0.40095 ,0.33074,-0.28223,0.32636,-0.24872,0.42714,0.10538,0.18978,-0.21392,0.077358,0.25456,-0.035788,-0.24849,-0.31267,-0.74889,-0 .031123,0.28418,0.055868,-0.27982,0.12521,0.23108,-0.20629,0.088269,-0.31153,-0.40711,0.063917,-0.42026,0.18822,-0.42677,-0. 051717, -0.052392, -0.1341, -0.18726, -0.56798, -0.31525, 0.063111, 0.29355, 0.33446, -0.22276, 0.55856, 0.25568, -0.27641, 0.13162, 0.211 09, -0.19135, -0.46096, -0.48333, 0.50242, 0.28792, 0.69155, -0.40548, 0.20882, -0.24729, -0.4067, -0.92679, 0.14147, -0.26459, -0.1219, -0.18663,0.37792,-0.27525,-0.15002,0.41089,0.086282,0.36707,0.39134,0.24567,-0.14968,-0.0232,0.40821,-0.027494,0.26942,0.1215,0.11944,-0.40082,0.044711,-0.45641,0.081264,-0.28871,-0.3351,-0.46428,-0.2659,-0.25424,0.27057,0.2254,0.47827,0.0089617,-0.12849,-0.030031,0.10856,-0.52442,-0.22589,-0.013136,0.36234,-0.20808,-0.072111,0.30285,0.26291,0.16197,0.23639,-0.0071087,-0.14576,-0.17559,0.26325,0.34905,0.10604,-0.31117,-0.54556,-0.89422,0.017616,-0.11261,-0.30832,0.061763,0.38819,0.058281,-0.54015,0.22642,0.47922,0.16835,-0.28717,-0.20311,0.42449,0.51578,-0.34063,-0.1145,0.35968,0.028322,-0.039375,0.11155,-0.72148,-0.070369,-0.18859,0.15754,-0.045399,-0.18787,0.016147,0.0093634,-0.28285,0.12067,0.35622,-0.24238,0.082042,-0.29993,0.54735,-0.091844,0.12827,0.24579,0.23052,0.20662,0.0022189,0.10875,-0.052753,0.73033,-0.31618,0.18756,-0.35548,-0.12272,0.20605,0.49178,-0.20901,0.075118,-0.16374,0.2307,0.043809,-0.4622,-0.4966,-0.11847,0.65149,-0.21991,-0.29969,0.039375,0.039327,0.59476,0.0059422,0.1241,-0.37176,-0.075128,0.061737,0.037677,0.18445,-0.10028,-0.23085,-0.13843,0.033648,0.18056,0.15014,-0.02941,-0.80613,0.28564,-0.32619,0.18798,-0.077773,-0.11521,0.13683,0.57268,0.25394,-0.048348,0.36895,-0.14142,0.17504,0.11282,-0.099034,0.01617,-0.28301,0.47614,0.47416,0.074527,-0.5138,-0.60652,0.013668,0.38312,0.098792,0.26468,-0.22388,0.44729,0.026868,-0.13904,0.20608,-0.16199,-2.3809,1.0467,1.1725,0.49076,-0.24992,-0.31623,-0.50482,0.31675,-0.31728,0.26221,0.13543,0.17883,0.42534,0.027469,-0.20934,-0.96892,-0.4379,0.03178,-0.11146,0.35721,0.16911,0.20465,-0.40038,-0.19199]，.

[0177] ……，

[0178] [0.45507,0.42837,0.073782,0.15588,0.3627,0.70686,0.10398,-0.22026,0.22404,-0.014677,0.15565,-0.23118,-0.64793,-0.45301,-0.19011,-0.018863,0.47468,-0.5492,0.31743,-0.17016,0.40326,0.40293,-0.53091,-0.44148,-0.080265,0.050296,0.10055,0.49174,0.16233,-0.29158,0.056588,0.13533,-0.44015,-0.30531,0.25688,0.18758,-0.39885,0.13044,-0.38209,-0.16148,0.083757,-0.029892,-0.087639,-0.52905,0.064181,-0.25512,0.34413,0.39288,-0.058174,0.098793,0.49826,0.051931,-0.09178,-0.47997,-0.60441,0.22357,0.10482,0.36093,0.28643,0.21742,0.82428,0.35106,-0.092973,-0.31426,-0.33214,0.11562,-0.24508,0.076538,0.34152,0.28615,-0.27172,-0.40387,0.43126,-0.35038,0.32242,0.33454,-0.31363,0.40072,-0.2257,0.12225,-0.23982,-0.11948,-0.055398,0.33762,0.052755,-0.24335,0.050043,-0.097635,0.27947,0.27523,-0.29075,-0.1217,0.2989,-0.18077,-0.25628,-0.24463,-0.025946,0.5039,-0.44225,0.11159,0.66951,-0.27193,-0.17196,-0.2134,0.10764,0.68568,0.23575,0.34649,-0.034088,0.094132,0.19916,-0.20284,-0.15727,0.038761,0.24125,0.5603,0.36364,-0.10046,-0.50317,-0.35339,0.12964,0.17248,-0.16493,0.35361,0.11483,-0.26352,-0.13249,-0.4742,0.34514,0.53094,0.37833,0.010242,0.3396,-0.30235,-0.26962,0.32345,0.11133,-0.082411,0.10582,-0.34896,-0.36941,0.36404,0.038351,0.47927,0.38482,-0.19814,-0.10485,0.21717,-0.48906,0.33122,-0.97608,0.29086,-0.039408,-0.030497,-0.12542,0.6837,0.13649,0.6373,0.40654,0.20084,0.031585,0.018098,0.22733,0.18354,0.13438,0.20076,-0.22296,-0.07787,0.019405,-0.84281,-0.15405,0.15307,-0.31749,-0.285,-0.43171,-0.17062,-0.064241,0.11738,-0.056049,-0.29681,0.54809,0.13307,-0.21344,-0.29585,-0.37227,-0.16964,0.034767,0.12121,0.22892,0.11892,-0.40955,-0.93508,0.18685,0.1313,0.3254,0.30791,-0.79626,-0.42484,0.33481,-0.21558,-0.13629,-0.35317,-0.15976,0.33566,0.36364,0.18006,0.13422,0.056216,0.030953,0.053569,-0.34772,0.19759,-0.1258,-0.26858,-0.31279,0.47831,0.39307,-0.08399,0.36008,0.48483,-1.1631,0.16649,-0.43008,-0.14762,-0.54182,-0.061401,-0.0067102,0.2304,0.21312,-0.047432,0.14503,-0.35008,0.088607,-0.19538, -0.94988, -0.39186, 0.2809, -0.021577, 0.020267, 0.31474, -0.14015, -0.090809, -0.31406, 0.055445, -0.37695, -0.50607, 0.21944, 0.11507, 0.40068, 0.10032, -0.1238, -0.084609, 0.19852, -0.038276, -0.12997, 0.017946, -0.45262, -0.32439, -0.14462, -0.73513, 0.31342, -0.3244, 0.11155, -0.098187, 0.10951, 0.343, -0.24029 ,0.054174,-0.12866,-0.19828,0.16169,-0.37709,-0.070296,-0.27317,0.076799,0.056824,0.81607,0.029843,0.31444,-0.49576,0.37763,-0.44894,-0.36281,0.83 307, -0.095801, 0.13511, 0.096772, -0.31248, 0.93476, -0.18458, -0.49332, -0.37064, -0.06909, 0.082717, 0.42247, -0.46211, 0.12589, 0.10234, 0.011343, -0.21237]].

[0179] (6) Perform semantic encoding on the vector representation output in step (5), and the result is shown below:

[0180] Check sentence pair 1 [-0.000096, 0.000220]

[0181] Check sentence pair 2 [-0.001008, 0.000407]

[0182] Check sentence pair 3 [-0.000099, 0.000217]

[0183] The sentence pair 4 was checked and returned [-0.001017, 0.000397].

[0184] (7) Decode the encoding result of step (6), and the result is shown below:

[0185] Check sentence pair 1 [0.499921, 0.500079]

[0186] Check sentence pair 2 [0.500354, 0.499646]

[0187] The sentence pair 3 was checked [0.500079, 0.499921].

[0188] The sentence pair 4 was checked. [0.499647, 0.500353]

[0189] (8) Based on the decoding results, the concatenated sentences are classified according to semantic similarity. The category labels are as follows:

[0190] Check sentence pair 1[1]

[0191] Check sentence pair 2[0]

[0192] Check sentence pair 3[0]

[0193] Check sentence pair 4[1]

[0194] (9) Record the location information of text rewriting in the English short passage sentences into the text rewriting plagiarism location information file. The result is shown below:

[0195] The plagiarized English short essay

[0196] 3^7^

[0197] Plagiarized English short essay

[0198] 3^7^.

[0199] The third step: Execute the "Comprehensive Evaluation Module for English Short Passage Plagiarism Detection"

[0200] The comprehensive evaluation module for English short passage plagiarism checks uses the results from the first step (copy and paste plagiarism check module) and the second step (rewrite plagiarism check module) to conduct a comprehensive plagiarism evaluation and generate the plagiarism evaluation result for the English short passage.

[0201] The format of the comprehensive evaluation result for plagiarism detection of English short passages in this implementation method is as follows:

[0202] Copying and pasting plagiarized sentences:

[0203] The first, second, fourth, fifth, sixth, and eighth sentences of the copied English passage are copied and pasted from the first, second, fourth, fifth, sixth, and eighth sentences of the original English passage, respectively.

[0204] Text rewriting plagiarized sentences:

[0205] The third and seventh sentences of the plagiarized English passage are suspected of being textual paraphrases and plagiarism, respectively.

[0206] The overall plagiarism rate for the English short passages was 83.117%.

Claims

1. A method for detecting plagiarism in English short passages, characterized by: The detection model includes a sequentially connected English short passage copy-paste plagiarism detection module, an English short passage rewriting plagiarism detection module, and an English short passage plagiarism detection comprehensive evaluation module. Its detection method includes the following steps: (1) The processing flow of the English short passage copy-paste plagiarism detection module is as follows: First, input the full text of two short English passages and process them into sentences; Second, using the N-gram model, select the five-gram window as needed, which is the five-gram model, and perform window segmentation on the sentences in the two English short passages to be examined according to the direction of sentence extension; Third, perform digital mapping on the cut window content to map out the corresponding values; Fourth, match the window values ​​corresponding to the sentences in the two English short articles. If the values ​​match the same, salt the contents of the clipping window and reassemble them, and then perform a number mapping again. If the values ​​match differently in the second number mapping, slide the clipping window in the direction of sentence extension. If the values ​​match the same again, jump to the next clipping window. Fifth, iteratively match sentences in the two English passages until both passages are completely matched; (2) The processing flow of the English short passage rewriting plagiarism detection module is as follows: First, iterate through the non-copy-pasted English sentences in the two English passages, and perform word segmentation, spelling correction, removal of special symbols, and deletion of stop words on the iterated English sentences. Output the results after the operation is completed. Second, iterate through the English sentences of the two short English articles after the first step and then concatenate the two English sentences. Third, the assembled English sentence is mapped into a distributed space to generate a vector representation of the English sentence. Fourth, extract semantic features from the vector representation of English compound sentences, and analyze whether the English compound sentences contain semantic plagiarism based on the semantic feature information; Fifth, after iteratively matching the non-copy-paste plagiarized sentences in the two English short articles, analyze whether there is semantic plagiarism in each pair of English sentences. (3) The processing flow of the comprehensive evaluation module for checking plagiarism in English short passages is as follows: First, the proportion of copied and pasted sentences and paraphrased sentences in the English passage is calculated to obtain the overall plagiarism rate of the English passage. Second, the coordinates of English words found to be copied and pasted, the coordinates of English sentences found to be rewritten and copied, and the overall plagiarism rate of the English passage are generated into the plagiarism check result record table. Third, based on the coordinates of the copied and pasted English words, the coordinates of the copied English sentences, and the overall plagiarism rate of the English passage in the plagiarism check result record table, generate the plagiarism check results and comments for the English passage.

2. The method for detecting plagiarism in English short passages according to claim 1, characterized in that: The specific processing flow of the English short passage copy-paste plagiarism detection module is as follows: Start at page 201; Read the full text of the two short English passages on page 202. P203 uses regular expressions to segment the entire English passage into sentences and outputs the segmented results. P204 Read the sentence segmentation results of the two English short articles after processing in step P203; Page 205 shows the sentence segmentation results of the first English passage; P206 The sentences of the first English passage traversed are split into five-grammar model windows; P207 The segmented windows are mapped digitally according to the calculation formula (1), and the numerical values ​​of the first English short article window segmentation are output; Page 208 shows the sentence segmentation results of the second English passage; P209 The sentences of the second English passage are segmented using the five-grammar model window; P210 The segmented windows are mapped digitally according to the calculation formula (1), and the numerical values ​​of the second English short article window segmentation are output; P211 matches the values ​​mapped in step P210 with the values ​​mapped in step P207; P212 Determine whether the numerical matching results in step P211 are the same. If they are the same, proceed to step P213; otherwise, proceed to step P216. P213 Salt the contents of two windows with the same numerical matching, and perform numerical mapping again according to the calculation formula (1), and match the mapped values ​​again; P214 Determine whether the numerical matching results in step P213 are the same. If they are the same, proceed to step P215; otherwise, proceed to step P216. P215 Save the clipping window with the same numerical match in the traversed sentences; P216 Determine whether the traversal of the clipping window in the second English passage has ended. If the traversal has ended, proceed to step P217; otherwise, proceed to step P210. P217 Based on the calculation formula (2) and calculation formula (3), determine whether the copy-paste degree value of the sentences traversed in the second English short article exceeds 0.

5. If it exceeds the threshold, save the position information of the clipping window to the copy-paste position information file and execute step P218. Otherwise, directly proceed to step P218. P218 Determine whether the traversal of sentences in the second English passage has ended. If the traversal has ended, proceed to step P219; otherwise, proceed to step P208. P219 Determine whether the traversal of the clipping window in the first English passage has ended. If the traversal has ended, proceed to step P220; otherwise, proceed to step P207. P220 Based on the calculation formula (2) and calculation formula (3), determine whether the copy-paste degree value of the sentences traversed in the first English short article exceeds 0.

5. If it exceeds the threshold, save the position information of the clipping window to the copy-paste position information file and execute step P221. Otherwise, directly proceed to step P221. P221 Determine whether the traversal of the first English passage has ended. If the traversal has ended, proceed to step P222; otherwise, proceed to step P205. End of page 222.

3. The method for detecting plagiarism in English short passages according to claim 2, characterized in that: The calculation formula (1) mentioned in the specific processing flow of the English short passage copy-paste plagiarism detection module is the calculation formula for string mapping numerical values: ； In the calculation formula (1), n ​​represents the character length in the English short passage clipping window, and i represents the character position in the clipping window; Formula (2) is the formula for calculating the degree of copying and pasting of English short passage sentences: ； In the calculation formula (2), n represents 'i' indicates that the sentence contains the i-th copied and pasted English word; Formula (3) is the formula for calculating the plagiarism tendency value of copying and pasting sentences in English short passages: ； In the calculation formula (3), the degree of copying and pasting of English short sentences is calculated by the calculation formula (2).

4. The method for detecting plagiarism in English short passages according to claim 1, characterized in that: The specific processing flow of the English short passage rewriting plagiarism detection module is as follows: Start at page 301; P302 Read out the non-copy-paste plagiarized sentences from the two short English passages respectively; P303 Results of iterating through the sentences in the first English passage that are not copied and pasted; P304 Results of iterating through the sentences in the second English passage that are not copied and pasted; P305 performs word segmentation on the English sentences in the two English short articles iterates through, and outputs the word segmentation results of the English sentences respectively; P306 uses regular expressions to match the stop word list and performs stop word removal on two English sentences respectively; P307 For the remaining words in two English sentences after removing stop words, perform spelling correction on each sentence and output the two final English sentences. P308 concatenates the two processed English sentences and outputs the concatenation result; P309 Represents the concatenated results output in step P308 as sentence vectors; P310 uses a bidirectional gated loop encoder to encode the output of step P309; P311 Decode based on the output of step P310 and the similarity category; P312 Classify sentences based on semantic similarity; P313 Determine whether the classification results are semantically similar. If they are similar, proceed to step P314; otherwise, proceed to step P315. P314 records the location information of two semantically similar sentences; P315 Determine whether the sentence results of the second English passage have been traversed. If the traversal is complete, proceed to step P316; otherwise, proceed to step P304. P316 Determine whether the sentence results of the first English passage have been traversed. If the traversal is complete, proceed to step P317; otherwise, proceed to step P303. P317 Output all sentences in the final two English passages that contain text rewriting and plagiarism, and save the location information of the rewritten and plagiarized sentences to the text rewriting and plagiarism location information file; End of page 318.

5. The method for detecting plagiarism in English short passages according to claim 1, characterized in that: The specific processing flow of the English short passage plagiarism detection and comprehensive evaluation module is as follows: Start at page 401; P402 Read the location information of copied and pasted sentences in two short English passages; P403 Read the location information of plagiarized sentences in two short English articles; P404 Calculate and output the combined plagiarism rate of the two English short articles according to the calculation formula (4); P405 Determine whether the output result of step P404 is plagiarized based on the calculation formula (5). If plagiarism is found, proceed to step P406; otherwise, proceed to step P407. P406 generates a comprehensive evaluation of plagiarism detection for English short passages and saves it to the English short passage plagiarism evaluation result file; Page 407 is the end.

6. The method for detecting plagiarism in English short passages according to claim 5, characterized in that: The calculation formula (4) mentioned in the specific processing flow of the English short passage plagiarism detection and comprehensive evaluation module is the calculation formula for the comprehensive plagiarism rate of English short passages: ； In the calculation formula (4), i represents the i-th sentence in the English passage that is copied and pasted, n represents the total number of sentences in the English passage that are copied and pasted, j represents the j-th sentence in the English passage that is rewritten and copied, m represents the total number of sentences in the English passage that are rewritten and copied, and 0.7 is the weight parameter for rewriting and copying detection. Formula (5) is the formula for calculating the plagiarism tendency value of English short passages: ； In the calculation formula (5), the overall plagiarism rate of English short articles is calculated by the calculation formula (4).

Images