TW202324185A - Item generating method - Google Patents

Item generating method Download PDF

Info

Publication number
TW202324185A
TW202324185A TW110145307A TW110145307A TW202324185A TW 202324185 A TW202324185 A TW 202324185A TW 110145307 A TW110145307 A TW 110145307A TW 110145307 A TW110145307 A TW 110145307A TW 202324185 A TW202324185 A TW 202324185A
Authority
TW
Taiwan
Prior art keywords
word
replacement
phrase
vocabulary
target
Prior art date
Application number
TW110145307A
Other languages
Chinese (zh)
Other versions
TWI773604B (en
Inventor
陳柏熹
謝嘉恩
Original Assignee
國立臺灣師範大學
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 國立臺灣師範大學 filed Critical 國立臺灣師範大學
Priority to TW110145307A priority Critical patent/TWI773604B/en
Application granted granted Critical
Publication of TWI773604B publication Critical patent/TWI773604B/en
Publication of TW202324185A publication Critical patent/TW202324185A/en

Links

Images

Landscapes

  • Making Paper Articles (AREA)
  • Machine Translation (AREA)

Abstract

An item generating method is implemented by a computing device. The computing device stores a plurality of relative words used to express the relationship between sentences. Each of the relative words corresponds to a grammatical rule for extracting a phrase. The item generating method includes: (A) dividing a text file related to a teaching material into paragraphs to obtain a descriptive sentence; (B) Pre-processing the descriptive sentence to obtain a plurality of words and part of speech of each word; (C) locating a target relational word from the descriptive sentence, and the target relational word is one of the relational words; (D) extracting an answer phrase from the descriptive sentence based on the grammatical rule of the target relative word; and (E) generating an item based on the descriptive sentence that excludes the answer phrase and the target relative word, wherein the answer phrase is used as the answer of the item.

Description

試題產生方法test question generation method

本發明是有關於一種試題產生方法，特別是指一種根據一相關於一教材的文字檔產生一試題的試題產生方法。The present invention relates to a method for generating test questions, in particular to a method for generating a test question based on a text file related to a teaching material.

考試最大的目的是讓學生更清楚地瞭解自己目前的學習情況，反思自己學習中的錯誤。一份好的試題，可以協助老師了解學生學習的程度，以及困難所在，有助於老師調適教學的步調，並作為補救教學的參考依據。The biggest purpose of the exam is to let students have a clearer understanding of their current learning situation and reflect on their mistakes in learning. A good test question can help teachers understand the level of students' learning and where the difficulties lie, help teachers adjust the pace of teaching, and serve as a reference for remedial teaching.

目前試卷的命題多仰賴資深且有經驗的教師，依標準分領域來進行命題，然而命題的過程除了需花費大量人力外，也相當曠日廢時。因此，現今的出版社也看到教師們的困境，紛紛提供了大量的題庫供考試組卷之用，但每當課綱修改或試教材更動時，又會使得題目的來源還是得回歸人力創作，故實有必要提出一解決方案。At present, most of the propositions of the test papers rely on senior and experienced teachers to make propositions according to the standards and fields. However, the process of propositioning not only requires a lot of manpower, but also takes a long time. Therefore, today's publishing houses also see the plight of teachers, and have provided a large number of question banks for the use of test papers. However, whenever the syllabus is revised or the test materials are changed, the source of the questions will still have to return to human creation. , it is necessary to propose a solution.

因此，本發明的目的，即在提供一種自動命題以節省人力與時間成本的試題產生方法。Therefore, the object of the present invention is to provide a method for generating test questions that automatically sets questions to save manpower and time costs.

於是，本發明試題產生方法，適用於根據一相關於一教材的文字檔產生一試題，且藉由一運算裝置來實施，該運算裝置儲存有多個用於表達文句之關係的關係詞，每一關係詞對應一用於擷取一詞組的語法規則，該試題產生方法包含以下步驟：Therefore, the test question generation method of the present invention is suitable for generating a test question based on a text file related to a teaching material, and is implemented by a computing device. The computing device stores a plurality of relational words used to express the relationship between sentences, each A relative word corresponds to a grammatical rule for extracting a phrase, and the method for generating test questions includes the following steps:

(A) 將該文字檔進行段落切分，以獲得一描述句；(A) Segment the text file into paragraphs to obtain a descriptive sentence;

(B)將該描述句進行文本前處理，以獲得多個斷詞及其對應之詞性；(B) performing pre-text processing on the descriptive sentence to obtain multiple hyphens and their corresponding parts of speech;

(C)自該描述句定位出一目標關係詞，該目標關係詞為該等關係詞之其中一者；(C) Locating a target relative word from the descriptive sentence, the target relative word being one of the relative words;

(D)根據該目標關係詞之語法規則，自該描述句擷取一答案詞組；及(D) extracting an answer phrase from the descriptive sentence according to the grammatical rules of the target relative word; and

(E)根據排除該答案詞組的該描述句及該目標關係詞，產生該試題，並將該答案詞組作為該試題之試題答案。(E) Generate the test question according to the descriptive sentence and the target relative word excluding the answer phrase, and use the answer phrase as the answer of the test question.

本發明的功效在於：藉由該運算裝置將該文字檔進行段落切分以獲得該描述句，並自該描述句定位出該目標關係詞，且根據該目標關係詞之語法規則，自該描述句擷取該答案詞組，並根據排除該答案詞組的該描述句及該目標關係詞，產生該試題，並將該答案詞組作為該試題之試題答案，藉此以自動根據該文字檔產生該試題，以達成自動命題以節省人力與時間成本之目的。The effect of the present invention is: segmenting the text file into paragraphs by the computing device to obtain the descriptive sentence, and locating the target relative word from the descriptive sentence, and according to the grammatical rules of the target relative word, from the description sentence to extract the answer phrase, and generate the test question according to the descriptive sentence and the target relative word excluding the answer phrase, and use the answer phrase as the answer of the test question, thereby automatically generating the test question according to the text file , in order to achieve the purpose of automatic proposition to save manpower and time costs.

參閱圖1，本發明試題產生方法之實施例，適用於根據一相關於一教材的文字檔產生一試題，並藉由一運算裝置1來實施。該運算裝置1包含一儲存模組11及一電連接該儲存模組11的處理模組12。該運算裝置1之實施態樣例如為一伺服器、一個人電腦、一筆記型電腦、一平板電腦或一智慧型手機等。Referring to FIG. 1 , the embodiment of the test question generation method of the present invention is suitable for generating a test question based on a text file related to a teaching material, and it is implemented by a computing device 1 . The computing device 1 includes a storage module 11 and a processing module 12 electrically connected to the storage module 11 . The implementation of the computing device 1 is, for example, a server, a personal computer, a notebook computer, a tablet computer, or a smart phone.

該儲存模組11儲存有多個用於表達文句之關係的關係詞、一用於擷取一詞組的語法規則、一詞向量轉換模型，及多個詞彙。每一關係詞對應一問題構句，表1示例了每一關係詞所對應的問題構句。在本實施方式中，該等關係詞例如包含具有、稱為、因為、導致、屬於、包含、引起等等，然不以此為限，此外，還可利用該詞向量轉換模型找出相似度高的其他相似詞以擴充該等關係詞。該語法規則可視需求選擇一基本語法規則或一進階語法規則，該基本語法規則用於抓取位於該目標關係詞後且位於除了頓號之標點符號前的斷句中所有連續出現的基本特定詞性之斷詞或頓號，亦即，抓取斷句中的連接於該目標關係詞後且未被其他非基本特定詞性之字詞隔開的所有連續的基本特定詞性之斷詞或頓號，該基本特定詞性可為名詞、形容詞、動詞及連接詞之任一者；該進階語法規則用於抓取位於該目標關係詞後且位於除了頓號之標點符號前的斷句中所有連續出現的進階特定詞性之斷詞或頓號，亦即，抓取斷句中的連接於該目標關係詞後且未被其他非進階特定詞性之字詞隔開的所有連續的進階特定詞性之斷詞或頓號，該進階特定詞性可為名詞、形容詞、動詞、連接詞、副詞及介詞之任一者。該詞向量轉換模型可利用如，gensim或word2vec等套件而訓練出。 關係詞 問題構句 具有 什麼特徵 稱為 下列何者 屬於 下列何種 因為 什麼原因 包含 什麼內容 導致 什麼結果 引起 什麼現象 表1 The storage module 11 stores a plurality of relative words used to express the relationship between sentences, a grammatical rule used to extract a word group, a word vector conversion model, and a plurality of vocabulary. Each relational word corresponds to a question sentence construction, and Table 1 illustrates the question construction sentence corresponding to each relational word. In this embodiment, the relative words include, for example, have, be called, because of, cause, belong to, include, cause, etc., but are not limited thereto. In addition, the word vector conversion model can also be used to find similarity High other similar words to expand these relative words. The grammatical rule can be selected as a basic grammatical rule or an advanced grammatical rule as required, and the basic grammatical rule is used to capture all consecutive basic specific parts of speech that appear after the target relative word and before the punctuation marks except commas. Hyphens or commas, that is, capture all consecutive basic part-of-speech hyphens or commas that are connected to the target relative word in the sentence and are not separated by other non-basic specific part-of-speech words. The basic specific part-of-speech It can be any one of noun, adjective, verb, and conjunction; the advanced grammar rule is used to capture all consecutive occurrences of the advanced specific part of speech in the sentence after the target relative word and before the punctuation mark except the comma Hyphens or commas, that is, all consecutive advanced part-of-speech hyphens or commas that are connected to the target relative word and not separated by other non-advanced specific part-of-speech words in the sentence, the advanced The class-specific part of speech may be any of noun, adjective, verb, conjunction, adverb, and preposition. The word vector conversion model can be trained using packages such as gensim or word2vec. Relationship Words question construction have what feature known as Which of the following belong Which of the following because what reason Include What content lead to what result cause what phenomenon Table 1

參閱圖1與圖2，以下將藉由本發明試題產生方法的實施例來說明該運算裝置1的運作細節。Referring to FIG. 1 and FIG. 2 , the operation details of the computing device 1 will be described below through an embodiment of the method for generating test questions of the present invention.

在步驟21中，該處理模組12將該文字檔進行文字清理及字形轉換，以獲得轉換後的該文字檔。其中，文字清理係過濾該文字檔中之亂碼、網頁標籤等等標記，字形轉換係將該文字檔中之將數字轉為半形、符號轉為全形。In step 21, the processing module 12 performs character cleaning and font conversion on the word file to obtain the converted word file. Among them, the text cleaning is to filter the garbled codes in the text file, webpage labels and other marks, and the font conversion is to convert the numbers in the text file to half-shaped, and the symbols to full-shaped.

在步驟22中，該處理模組12將轉換後的該文字檔進行段落切分，以獲得一描述句。舉例來說，「蕨類植物具有根、莖和葉，是最早演化出維管束的植物」為一示例之描述句，然不以此為限。In step 22, the processing module 12 divides the converted text file into paragraphs to obtain a descriptive sentence. For example, "ferns have roots, stems and leaves, and are the first plants to evolve vascular bundles" is an exemplary descriptive sentence, but it is not limited thereto.

在步驟23中，該處理模組12將該描述句進行文本前處理，以獲得多個斷詞及其對應之詞性。其中，該文本前處理可採用如CKIP tagger或Jeiba等中文分詞技術。以「蕨類植物具有根、莖和葉，是最早演化出維管束的植物」之描述句為例，其經文本前處理後可得到「(蕨類,Na)(植物,Na)(具有,VJ)(根,Na)(、,PAUSE)(莖,Na)(和,Caa)(葉,Na)(，,COMMA)(是,SHI)(最早,D)(演化出,VC)(維管束,Na)(的,DE)(植物,Na)」之結果。In step 23, the processing module 12 performs text pre-processing on the descriptive sentence to obtain a plurality of segmented words and their corresponding parts of speech. Among them, the text pre-processing can adopt Chinese word segmentation technology such as CKIP tagger or Jeiba. Taking the descriptive sentence “Pteridophytes have roots, stems and leaves, and are the first plants to evolve vascular bundles” as an example, after text preprocessing, we can get “(ferns,Na)(plants,Na)(with, VJ) (root, Na) (,, PAUSE) (stem, Na) (and, Caa) (leaf, Na) (,, COMMA) (yes, SHI) (earliest, D) (evolved, VC) (dimensional Tube bundle, Na) (of, DE) (plant, Na)" results.

在步驟24中，該處理模組12判定該描述句是否包含一目標關係詞，該目標關係詞為該等關係詞之其中一者。當該處理模組12判定出該描述句包含該目標關係詞時，流程進行步驟25；當該處理模組12判定出該描述句不包含該目標關係詞時，流程進行步驟29。In step 24, the processing module 12 determines whether the descriptive sentence contains a target relative word, and the target relative word is one of the relative words. When the processing module 12 determines that the descriptive sentence contains the target relative word, the process proceeds to step 25 ; when the processing module 12 determines that the descriptive sentence does not contain the target relative word, the process proceeds to step 29 .

在步驟25中，該處理模組12自該描述句定位出一目標關係詞。以「蕨類植物具有根、莖和葉，是最早演化出維管束的植物」之描述句為例，可定位出「具有」此一目標關係詞。In step 25, the processing module 12 locates a target relative word from the descriptive sentence. Taking the descriptive sentence "the fern has roots, stems and leaves, and is the first plant to evolve vascular bundles" as an example, the target relative word "has" can be located.

在步驟26中，該處理模組12根據該目標關係詞及該語法規則，自該描述句擷取一答案詞組。以「蕨類植物具有根、莖和葉，是最早演化出維管束的植物」之描述句為例，並以該基本語法規則來擷取連接於該目標關係詞後且位於除了逗號前的斷句「根、莖和葉」中所有連續出現的基本特定詞性(亦即，名詞、形容詞、動詞及連接詞之任一者)之斷詞或頓號，即可擷取出「根、莖和葉」此一答案詞組。In step 26, the processing module 12 extracts an answer phrase from the description sentence according to the target relative word and the grammatical rule. Take the descriptive sentence "Ferns have roots, stems and leaves, and are the first plants to evolve vascular bundles" as an example, and use the basic grammatical rules to extract the sentence sentences that are connected after the target relative word and before commas All consecutive occurrences of hyphens or commas of the basic specific part of speech (that is, any of noun, adjective, verb, and conjunction) in "roots, stems, and leaves" can extract the words "roots, stems, and leaves" An answer phrase.

在步驟27中，該處理模組12根據該答案詞組及該等詞彙，產生多個與該答案詞組相似的誘答詞組。In step 27, the processing module 12 generates a plurality of decoy phrases similar to the answer phrase according to the answer phrase and the vocabulary.

值得一提的是，步驟27包含以下子步驟(見圖3及圖4)。It is worth mentioning that step 27 includes the following sub-steps (see FIG. 3 and FIG. 4 ).

在子步驟271中，該處理模組12自該答案詞組選擇一目標詞。In sub-step 271, the processing module 12 selects a target word from the answer phrase.

在子步驟272中，該處理模組12根據該目標詞及該答案詞組中相鄰該目標詞的相鄰詞，獲得多個目標詞組合。在本實施例中，該處理模組12係將位於該目標詞前的相鄰詞與該目標詞組成該等目標詞組合之其中一者，並將位於該目標詞後的相鄰詞與該目標詞組成該等目標詞組合之其中另一者。以「根、莖和葉」之答案詞組為例，若目標詞為「根」，則由於「根」前無相鄰詞，故以＜根＞作為該等目標詞組合之其中一者，而以＜根、莖＞作為該等目標詞組合之其中另一者。In sub-step 272, the processing module 12 obtains a plurality of target word combinations according to the target word and the adjacent words adjacent to the target word in the answer phrase. In this embodiment, the processing module 12 forms one of the target word combinations with the adjacent words before the target word and the target word, and combines the adjacent words after the target word with the target word The target word constitutes the other of the target word combinations. Taking the answer phrase of "root, stem and leaf" as an example, if the target word is "root", since there is no adjacent word before "root", <root> is used as one of the target word combinations, and Take <root, stem> as the other of these target word combinations.

在子步驟273中，對於每一目標詞組合，該處理模組12計算出該目標詞組合之一待配對詞向量。In sub-step 273, for each target word combination, the processing module 12 calculates a word vector to be paired for the target word combination.

值得一提的是，步驟273包含以下子步驟(見圖5)。It is worth mentioning that step 273 includes the following sub-steps (see FIG. 5 ).

在子步驟273a中，該處理模組12係根據該目標詞組合中之目標詞利用該詞向量轉換模型轉換出該目標詞的目標詞向量。In sub-step 273a, the processing module 12 uses the word vector conversion model to convert the target word vector of the target word according to the target word in the target word combination.

在子步驟273b中，該處理模組12係根據該目標詞組合中之相鄰詞利用該詞向量轉換模型轉換出該相鄰詞的相鄰詞向量。In sub-step 273b, the processing module 12 uses the word vector conversion model to convert the adjacent word vectors of the adjacent words according to the adjacent words in the target word combination.

在子步驟273c中，該處理模組12計算該目標詞向量與該相鄰詞向量之中心，以獲得該目標詞組合之一待配對詞向量。In sub-step 273c, the processing module 12 calculates the center of the target word vector and the adjacent word vector to obtain a paired word vector of the target word combination.

在子步驟274中，對於每一詞彙，該處理模組12根據該詞彙利用該詞向量轉換模型轉換出該詞彙的詞彙向量。In sub-step 274, for each vocabulary, the processing module 12 converts the vocabulary vector of the vocabulary using the word vector conversion model according to the vocabulary.

在子步驟275中，對於每一目標詞組合，該處理模組12根據該目標詞組合之待配對詞向量及該等詞彙之詞彙向量，自該等詞彙選取出至少一候選詞彙，其中，該至少一候選詞彙之詞彙向量與該目標詞組合之待配對詞向量的相似度為排序最高或前幾高，當該至少一候選詞彙之數目為一個時，即選擇對應有相似度最高的詞彙作為該候選詞彙，當該至少一候選詞彙之數目為N個時，即選擇對應有相似度前N高的詞彙作為該等候選詞彙。In sub-step 275, for each target word combination, the processing module 12 selects at least one candidate word from the words according to the word vector to be paired of the target word combination and the word vector of the words, wherein, the The similarity between the vocabulary vector of at least one candidate vocabulary and the word vector to be paired of the target word combination is the highest or the first few high. When the number of the at least one candidate vocabulary is one, the corresponding vocabulary with the highest similarity is selected as For the candidate vocabulary, when the number of the at least one candidate vocabulary is N, select the vocabulary corresponding to the top N highest similarity as the candidate vocabulary.

在子步驟276中，該處理模組12根據該目標詞之目標詞向量及每一候選詞彙的詞彙向量，自所有候選詞彙選取出一替換詞彙。其中，該替換詞彙之詞彙向量與該目標詞之目標詞向量的相似度大於一門檻值。In sub-step 276, the processing module 12 selects a replacement word from all candidate words according to the target word vector of the target word and the vocabulary vector of each candidate word. Wherein, the similarity between the vocabulary vector of the replacement vocabulary and the target word vector of the target word is greater than a threshold.

在子步驟277中，該處理模組12將該答案詞組中的目標詞替換為該替換詞彙，以獲得一替換詞組。假設選擇出之替換詞彙為「芽」，則該替換詞組即為「芽、莖和葉」。In sub-step 277, the processing module 12 replaces the target word in the answer phrase with the replacement vocabulary to obtain a replacement phrase. Assuming that the selected replacement word is "bud", the replacement phrase is "bud, stem and leaf".

在子步驟278中，該處理模組12根據該答案詞組及該替換詞組，判定是否將該替換詞組作為該等誘答詞組之其中一者。當該處理模組12判定出不將該替換詞組作為該等誘答詞組之其中一者時，流程進行子步驟279；當該處理模組12判定出將該替換詞組作為該等誘答詞組之其中一者時，流程進行子步驟280。In sub-step 278, the processing module 12 determines whether the replacement phrase is one of the decoy phrases according to the answer phrase and the replacement phrase. When the processing module 12 determines that the replacement phrase is not used as one of the decoy phrases, the process proceeds to sub-step 279; when the processing module 12 determines that the replacement phrase is used as one of the decoy phrases If either, the process proceeds to sub-step 280 .

值得一提的是，子步驟278包含以下子步驟(見圖6)。It is worth mentioning that sub-step 278 includes the following sub-steps (see FIG. 6 ).

在子步驟278a中，該處理模組12計算該替換詞組之一替換詞向量。其中，該處理模組12係將該替換詞組中之每一詞彙利用該詞向量轉換模型轉換出該替換詞組中之每一詞彙的詞彙向量，且計算該替換詞組中之所有詞彙之詞彙向量的中心，以獲得該替換詞組之替換詞向量。In sub-step 278a, the processing module 12 calculates a replacement word vector of the replacement phrase. Wherein, the processing module 12 uses the word vector transformation model to convert each vocabulary in the replacement phrase to the vocabulary vector of each vocabulary in the replacement phrase, and calculates the lexical vector of all vocabulary in the replacement phrase Center to get the replacement word vector for the replacement phrase.

在子步驟278b中，該處理模組12計算該答案詞組之一答案詞向量。其中，該處理模組12係將該答案詞組中之每一詞彙利用該詞向量轉換模型轉換出該答案詞組中之每一詞彙的詞彙向量，且計算該答案詞組中之所有詞彙之詞彙向量的中心，以獲得該答案詞組之答案詞向量。In sub-step 278b, the processing module 12 calculates an answer word vector of the answer phrase. Wherein, the processing module 12 utilizes the word vector transformation model to convert each vocabulary in the answer phrase to the vocabulary vector of each vocabulary in the answer phrase, and calculates the lexical vector of all vocabulary in the answer phrase center to obtain the answer word vector of the answer phrase.

在子步驟278c中，該處理模組12判定該替換詞向量與該答案詞向量之相似度是否大於一基準值，以判定是否將該替換詞組作為該等誘答詞組之其中一者。當該處理模組12判定出該替換詞向量與該答案詞向量之相似度不大於該基準值時，即判定不將該替換詞組作為該等誘答詞組之其中一者，流程進行子步驟279；當該處理模組12判定出該替換詞向量與該答案詞向量之相似度大於該基準值時，即判定將該替換詞組作為該等誘答詞組之其中一者，流程進行子步驟280。In sub-step 278c, the processing module 12 determines whether the similarity between the replacement word vector and the answer word vector is greater than a reference value, so as to determine whether the replacement phrase is one of the decoy phrases. When the processing module 12 determines that the similarity between the replacement word vector and the answer word vector is not greater than the reference value, it is determined not to use the replacement phrase as one of the decoy phrases, and the flow proceeds to substep 279 ; when the processing module 12 determines that the similarity between the replacement word vector and the answer word vector is greater than the reference value, it determines that the replacement phrase is one of the decoy phrases, and the flow proceeds to substep 280.

在子步驟279中，該處理模組12自該答案詞組選擇另一目標詞，並回到步驟272。In sub-step 279 , the processing module 12 selects another target word from the answer phrase, and returns to step 272 .

在子步驟280中，該處理模組12將該替換詞組作為該等誘答詞組之其中一者。In sub-step 280, the processing module 12 takes the replacement phrase as one of the decoy phrases.

在子步驟281中，該處理模組12根據該替換詞彙及該替換詞組中相鄰該替換詞彙的相鄰詞，獲得多個替換詞組合。在本實施例中，該處理模組12係將位於該替換詞彙前的相鄰詞與該替換詞彙組成該等替換詞組合之其中一者，並將位於該替換詞彙後的相鄰詞與該替換詞彙組成該等替換詞組合之其中另一者。In sub-step 281 , the processing module 12 obtains a plurality of replacement word combinations according to the replacement vocabulary and the adjacent words adjacent to the replacement vocabulary in the replacement word group. In this embodiment, the processing module 12 forms one of the replacement word combinations with the adjacent words before the replacement vocabulary and the replacement vocabulary, and combines the adjacent words after the replacement vocabulary with the replacement vocabulary. A substitute word constitutes the other of those combinations of substitute words.

在子步驟282中，對於每一替換詞組合，該處理模組12計算出該替換詞組合之另一待配對詞向量。類似的，該處理模組12亦是根據該替換詞組合中之替換詞彙利用該詞向量轉換模型轉換出該替換詞彙的替換詞彙向量，並根據該替換詞組合中之相鄰詞利用該詞向量轉換模型轉換出該相鄰詞的相鄰詞向量，且計算該替換詞彙向量與該相鄰詞向量之中心，以獲得該替換詞組合之另一待配對詞向量。In sub-step 282, for each replacement word combination, the processing module 12 calculates another word vector to be paired for the replacement word combination. Similarly, the processing module 12 also uses the word vector transformation model to convert the replacement vocabulary vector of the replacement vocabulary according to the replacement vocabulary in the replacement word combination, and uses the word vector according to the adjacent words in the replacement word combination The conversion model converts the adjacent word vector of the adjacent word, and calculates the center of the replacement word vector and the adjacent word vector to obtain another paired word vector of the replacement word combination.

在子步驟283中，對於每一替換詞組合，該處理模組12根據該替換詞組合之待配對詞向量及該等詞彙之詞彙向量，自該等詞彙選取出至少另一候選詞彙，其中，該至少另一候選詞彙之詞彙向量與該替換詞組合之待配對詞向量的相似度為排序最高或前幾高，當該至少另一候選詞彙之數目為一個時，即選擇對應有相似度最高的詞彙作為該另一候選詞彙，當該至少另一候選詞彙之數目為N個時，即選擇對應有相似度前N高的詞彙作為該等另一候選詞彙。In sub-step 283, for each replacement word combination, the processing module 12 selects at least another candidate word from the vocabulary according to the word vector to be paired of the replacement word combination and the vocabulary vector of the vocabulary, wherein, The similarity between the vocabulary vector of the at least another candidate vocabulary and the word vector to be paired of the replacement word combination is the highest or the first few high. When the number of the at least another candidate vocabulary is one, the corresponding selection has the highest similarity. The vocabulary of the other candidate vocabulary is used as the other candidate vocabulary, and when the number of the at least another candidate vocabulary is N, the vocabulary corresponding to the top N highest similarities is selected as the other candidate vocabulary.

在子步驟284中，該處理模組12根據該目標詞之目標詞向量及每一另一候選詞彙的詞彙向量，自所有另一候選詞彙選取出另一替換詞彙。其中，該另一替換詞彙之詞彙向量與該目標詞之目標詞向量的相似度大於該門檻值。In sub-step 284, the processing module 12 selects another replacement word from all other candidate words according to the target word vector of the target word and the vocabulary vector of each other candidate word. Wherein, the similarity between the vocabulary vector of the another replacement vocabulary and the target word vector of the target word is greater than the threshold value.

在子步驟285中，該處理模組12將該替換詞組中的替換詞彙替換為另一替換詞彙，以獲得另一替換詞組。In sub-step 285, the processing module 12 replaces the replacement vocabulary in the replacement phrase with another replacement vocabulary to obtain another replacement phrase.

在子步驟286中，該處理模組12根據該替換詞組及該另一替換詞組，判定是否將該另一替換詞組作為該等誘答詞組之其中一者。當該處理模組12判定出不將該另一替換詞組作為該等誘答詞組之其中一者時，流程進行子步驟287；當該處理模組12判定出將該另一替換詞組作為該等誘答詞組之其中一者時，流程進行子步驟288。值得一提的是，該處理模組12判定是否將該另一替換詞組作為該等誘答詞組之其中一者的判定流程與子步驟278a~子步驟278c類似，故於此不再重述其細節。In sub-step 286, the processing module 12 determines whether to use another replacement phrase as one of the decoy phrases according to the replacement phrase and the other replacement phrase. When the processing module 12 determines not to use another replacement phrase as one of these decoy phrases, the process proceeds to sub-step 287; when the processing module 12 determines that another replacement phrase is used as the When one of the decoy phrases is selected, the flow proceeds to sub-step 288. It is worth mentioning that the processing module 12 determines whether to use another replacement phrase as one of the decoy phrases. The determination process is similar to sub-step 278a ~ sub-step 278c, so it will not be repeated here. detail.

在子步驟287中，該處理模組12自該答案詞組選擇另一目標詞，並回到步驟272。In sub-step 287 , the processing module 12 selects another target word from the answer phrase, and returns to step 272 .

在子步驟288中，該處理模組12將該另一替換詞組作為該等誘答詞組之其中一者。In sub-step 288, the processing module 12 takes another replacement phrase as one of the decoy phrases.

在子步驟289中，該處理模組12回到步驟281以獲得不同之誘答詞組，其中在下一次執行步驟281時，該另一替換詞彙作為該替換詞彙，該另一替換詞組作為該替換詞組。In sub-step 289, the processing module 12 returns to step 281 to obtain different decoy phrases, wherein when performing step 281 next time, the other replacement vocabulary is used as the replacement vocabulary, and the other replacement phrase is used as the replacement phrase .

在步驟28中，該處理模組12根據排除該答案詞組的該描述句、該目標關係詞及該等誘答詞組，產生該試題，並將該等誘答詞組作為該試題之誘答選項，且將該答案詞組作為該試題之試題答案。其中，該處理模組12係根據排除該答案詞組的該描述句、該目標關係詞及該目標關係詞所對應之問題構句，產生該試題。藉此，即可產生選擇題型之試題。表2示例出所產生之選擇題型的試題。 蕨類植物具有什麼特徵，是最早演化出維管束的植物 (A)芽、莖和葉 誘答詞組 (B)根、莖和葉 答案詞組 (C)塊根、莖和葉 誘答詞組 (D)根、莖和葉脈 誘答詞組 表2 In step 28, the processing module 12 generates the test question according to the descriptive sentence excluding the answer phrase, the target relative word and the decoy phrases, and uses the decoy phrases as the decoy options of the test question, And use the answer phrase as the answer of the test question. Wherein, the processing module 12 constructs sentences according to the descriptive sentence excluding the answer phrase, the target relative word and the question corresponding to the target relative word to generate the test question. In this way, multiple-choice test questions can be generated. Table 2 shows examples of the multiple-choice test questions that were generated. What are the characteristics of ferns, the first plants to evolve vascular bundles (A) Buds, stems and leaves Decoy phrase (B) Roots, stems and leaves answer phrase (C) roots, stems and leaves Decoy phrase (D) roots, stems and veins Decoy phrase Table 2

值得特別說明的是，經由步驟27之執行，即可產生選擇題型的試題，然而，在其他實施方式中，亦可不執行步驟27而產生簡答題型，此時，在步驟28中，該處理模組12係根據排除該答案詞組的該描述句、該目標關係詞，及該目標關係詞所對應之問題構句，產生該試題，並將該答案詞組作為該試題之試題答案。藉此，即可產生簡答題型之試題。舉例來說，簡答題型之試題可為「蕨類植物具有什麼特徵，是最早演化出維管束的植物」。It is worth noting that, through the execution of step 27, multiple-choice test questions can be generated. However, in other embodiments, step 27 can also be executed without performing short-answer questions. At this time, in step 28, the processing Module 12 generates the test question according to the descriptive sentence excluding the answer phrase, the target relative word, and the question corresponding to the target relation word, and uses the answer phrase as the answer of the test question. In this way, short-answer test questions can be generated. For example, a short-answer test question could be "What are the characteristics of ferns, the first plants to evolve vascular bundles".

在步驟29中，對於該描述句之每一斷詞，該處理模組12計算該斷詞的斷詞權重值。In step 29, for each segmented word of the descriptive sentence, the processing module 12 calculates the segmented word weight value of the segmented word.

值得特別說明的是，步驟29包含以下子步驟(見圖7)。It is worth noting that step 29 includes the following sub-steps (see FIG. 7 ).

在子步驟291中，對於該描述句之每一斷詞，該處理模組12根據該斷詞利用該詞向量轉換模型轉換出該斷詞的斷詞向量。In sub-step 291 , for each segmented word of the descriptive sentence, the processing module 12 converts the segmented word's segmented word by using the word vector conversion model according to the segmented word.

在子步驟292中，對於該描述句之每一斷詞，該處理模組12計算排除該斷詞後之剩餘的斷詞之斷詞向量的中心，以獲得一剩餘斷詞向量。In sub-step 292, for each segmented word of the descriptive sentence, the processing module 12 calculates the centers of the segmented word vectors of the remaining segmented words after excluding the segmented word to obtain a remaining segmented word vector.

在子步驟293中，對於該描述句之每一斷詞，該處理模組12計算該斷詞之斷詞向量與排除該斷詞後之剩餘斷詞向量間之一餘弦相似度，並以所計算出之餘弦相似度作為該斷詞之斷詞權重值。當所計算出之餘弦相似度越大，即代表該斷詞在該描述句的權重越高。In sub-step 293, for each segmented word of the descriptive sentence, the processing module 12 calculates a cosine similarity between the segmented word vector of the segmented word and the remaining segmented word vector after excluding the segmented word, and uses the resulting The calculated cosine similarity is used as the word segmentation weight value of the word segmentation. When the calculated cosine similarity is larger, it means that the weight of the segmented word is higher in the descriptive sentence.

在步驟30中，該處理模組12將權重最高的斷詞作為一答案，並根據排除該答案的該描述句，產生該試題。藉此，即可產生填空題型之試題。In step 30, the processing module 12 takes the sentence with the highest weight as an answer, and generates the test question according to the descriptive sentence excluding the answer. In this way, test questions of the fill-in-the-blank type can be produced.

綜上所述，本發明試題產生方法，藉由該運算裝置1將該文字檔進行段落切分以獲得該描述句，並自該描述句擷取答案，並產生該試題，藉此以自動根據該文字檔產生該試題，並可依不同情境產生如選擇題型、簡答題型或填空題型之試題，藉此達成自動命題以節省人力與時間成本之目的，故確實能達成本發明的目的。To sum up, the test question generation method of the present invention uses the computing device 1 to segment the text file into paragraphs to obtain the descriptive sentence, extracts the answer from the descriptive sentence, and generates the test question, thereby automatically based on The text file generates the test questions, and can generate test questions such as multiple-choice questions, short-answer questions or fill-in-the-blank questions according to different situations, so as to achieve the purpose of automatic proposition to save manpower and time costs, so it can indeed achieve the purpose of the present invention .

惟以上所述者，僅為本發明的實施例而已，當不能以此限定本發明實施的範圍，凡是依本發明申請專利範圍及專利說明書內容所作的簡單的等效變化與修飾，皆仍屬本發明專利涵蓋的範圍內。But the above-mentioned ones are only embodiments of the present invention, and should not limit the scope of the present invention. All simple equivalent changes and modifications made according to the patent scope of the present invention and the content of the patent specification are still within the scope of the present invention. Within the scope covered by the patent of the present invention.

1:運算裝置 11:儲存模組 12:處理模組 21~30:步驟 271~289:子步驟 273a~273c:子步驟 278a~278c:子步驟 291~293:子步驟 1: computing device 11: Storage module 12: Processing module 21~30: Steps 271~289: sub-steps 273a~273c: sub-steps 278a~278c: sub-steps 291~293: sub-steps

本發明的其他的特徵及功效，將於參照圖式的實施方式中清楚地呈現，其中： 圖1是一方塊圖，說明實施本發明試題產生方法之實施例的一運算裝置； 圖2是一流程圖，說明本發明試題產生方法之實施例； 圖3與圖4皆是一流程圖，配合說明一處理模組如何產生多個誘答詞組； 圖5是一流程圖，說明該處理模組如何計算一目標詞組合之一待配對詞向量； 圖6是一流程圖，說明該處理模組如何判定是否將一替換詞組作為該等誘答詞組之其中一者；及 圖7是一流程圖，說明該處理模組如何計算每一斷詞權重值。 Other features and effects of the present invention will be clearly presented in the implementation manner with reference to the drawings, wherein: Fig. 1 is a block diagram illustrating a computing device implementing an embodiment of the method for producing test questions of the present invention; Fig. 2 is a flow chart, illustrates the embodiment of the method for producing test questions of the present invention; Fig. 3 and Fig. 4 are all a flow chart, cooperating to illustrate how a processing module generates a plurality of decoy phrases; Fig. 5 is a flowchart illustrating how the processing module calculates a word vector to be paired for a target word combination; Fig. 6 is a flowchart illustrating how the processing module determines whether to use a replacement phrase as one of the decoy phrases; and FIG. 7 is a flowchart illustrating how the processing module calculates the weight value of each word segmentation.

21~30:步驟 21~30: Steps

Claims (10)

一種試題產生方法，適用於根據一相關於一教材的文字檔產生一試題，且藉由一運算裝置來實施，該運算裝置儲存有多個用於表達文句之關係的關係詞，及一用於擷取一詞組的語法規則，該試題產生方法包含以下步驟： (A) 將該文字檔進行段落切分，以獲得一描述句； (B)將該描述句進行文本前處理，以獲得多個斷詞及其對應之詞性； (C)自該描述句定位出一目標關係詞，該目標關係詞為該等關係詞之其中一者； (D)根據該目標關係詞及該語法規則，自該描述句擷取一答案詞組；及 (E)根據排除該答案詞組的該描述句及該目標關係詞，產生該試題，並將該答案詞組作為該試題之試題答案。 A method for generating test questions, suitable for generating a test question based on a text file related to a teaching material, and implemented by a computing device, the computing device stores a plurality of relational words used to express the relationship between sentences, and a device for The grammatical rules of a phrase are extracted, and the test question generation method includes the following steps: (A) Segment the text file into paragraphs to obtain a descriptive sentence; (B) performing pre-text processing on the descriptive sentence to obtain multiple hyphens and their corresponding parts of speech; (C) Locating a target relative word from the descriptive sentence, the target relative word being one of the relative words; (D) extracting an answer phrase from the descriptive sentence according to the target relative word and the grammatical rule; and (E) Generate the test question according to the descriptive sentence and the target relative word excluding the answer phrase, and use the answer phrase as the answer of the test question. 如請求項1所述的試題產生方法，該運算裝置儲存有多個詞彙，在步驟(E)之前，還包含以下步驟： (F)根據該答案詞組及該等詞彙，產生多個與該答案詞組相似的誘答詞組； 其中，在該步驟(E)中，不僅根據排除該答案詞組的該描述句及該目標關係詞，還根據該等誘答詞組，產生該試題，其中該等誘答詞組作為該試題之誘答選項。 As the test question generation method described in claim item 1, the computing device stores a plurality of vocabulary, and before step (E), it also includes the following steps: (F) According to the answer phrase and the vocabulary, generate multiple inducement phrases similar to the answer phrase; Wherein, in the step (E), the test question is generated not only based on the descriptive sentence and the target relative word excluding the answer phrase, but also based on the induced answer phrases, wherein the induced answer phrases are used as the induced answers of the test question options. 如請求項2所述的試題產生方法，其中，步驟(F)包含以下子步驟： (F-1)自該答案詞組選擇一目標詞； (F-2)根據該目標詞及該答案詞組中相鄰該目標詞的相鄰詞，獲得多個目標詞組合； (F-3)對於每一目標詞組合，計算出該目標詞組合之一待配對詞向量； (F-4)對於每一目標詞組合，根據該目標詞組合之待配對詞向量及該等詞彙之詞彙向量，自該等詞彙選取出至少一候選詞彙； (F-5)根據該目標詞之目標詞向量及每一候選詞彙的詞彙向量，自所有候選詞彙選取出一替換詞彙，其中，該替換詞彙之詞彙向量與該目標詞之目標詞向量的相似度大於一門檻值； (F-6)將該答案詞組中的目標詞替換為該替換詞彙，以獲得一替換詞組 (F-7) 根據該答案詞組及該替換詞組，判定是否將該替換詞組作為該等誘答詞組之其中一者； (F-8)當判定出將該替換詞組作為該等誘答詞組之其中一者時，將該替換詞組作為該等誘答詞組之其中一者； (F-9)根據該替換詞彙及該替換詞組中相鄰該替換詞彙的相鄰詞，獲得多個替換詞組合； (F-10)對於每一替換詞組合，計算出該替換詞組合之另一待配對詞向量； (F-11)對於每一替換詞組合，根據該替換詞組合之待配對詞向量及該等詞彙之詞彙向量，自該等詞彙選取出至少另一候選詞彙； (F-12)根據該目標詞之目標詞向量及每一另一候選詞彙的詞彙向量，自所有另一候選詞彙選取出另一替換詞彙，其中，該另一替換詞彙之詞彙向量與該目標詞之目標詞向量的相似度大於該門檻值； (F-13)將該替換詞組中的替換詞彙替換為另一替換詞彙，以獲得另一替換詞組， (F-14)根據該替換詞組及該另一替換詞組，判定是否將該另一替換詞組作為該等誘答詞組之其中一者； (F-15) 當判定出將該另一替換詞組作為該等誘答詞組之其中一者時，將該另一替換詞組作為該等誘答詞組之其中一者；及 (F-16)回到步驟(F-9)以獲得不同之誘答詞組，其中在下一次執行步驟(F-9)時，該另一替換詞彙作為該替換詞彙，該另一替換詞組作為該替換詞組。 The method for generating test questions as described in claim item 2, wherein step (F) includes the following sub-steps: (F-1) Select a target word from the answer phrase; (F-2) Obtain a plurality of target word combinations according to the adjacent words adjacent to the target word in the target word and the answer phrase; (F-3) For each target word combination, calculate one of the target word combinations to be paired word vectors; (F-4) For each target word combination, select at least one candidate vocabulary from the vocabulary according to the target word combination's word vector to be paired and the vocabulary vector of the vocabulary; (F-5) Select a replacement word from all candidate words according to the target word vector of the target word and the vocabulary vector of each candidate word, wherein the vocabulary vector of the replacement word is similar to the target word vector of the target word degree is greater than a threshold value; (F-6) replace the target word in this answer phrase with this replacement vocabulary, to obtain a replacement phrase (F-7) Based on the answer phrase and the replacement phrase, determine whether the replacement phrase is one of the inducement phrases; (F-8) When it is determined that the replacement phrase is one of the decoy phrases, the replacement phrase is one of the decoy phrases; (F-9) Obtain a plurality of replacement word combinations according to the replacement vocabulary and the adjacent words adjacent to the replacement vocabulary in the replacement phrase; (F-10) For each replacement word combination, calculate another word vector to be paired for the replacement word combination; (F-11) For each replacement word combination, select at least one other candidate word from the vocabulary based on the word vector to be paired for the replacement word combination and the vocabulary vector of the vocabulary; (F-12) According to the target word vector of the target word and the vocabulary vector of each other candidate vocabulary, select another replacement word from all other candidate words, wherein, the vocabulary vector of the another replacement word is the same as the target word The similarity of the target word vector of the word is greater than the threshold; (F-13) replacing a replacement term in the replacement phrase with another replacement term to obtain another replacement phrase, (F-14) Based on the replacement phrase and the other replacement phrase, determine whether to use the other replacement phrase as one of the decoy phrases; (F-15) when it is determined that the alternative phrase is one of the decoy phrases, the other alternative phrase is one of the decoy phrases; and (F-16) Go back to step (F-9) to obtain different decoy phrases, wherein when step (F-9) is performed next time, the other replacement vocabulary is used as the replacement vocabulary, and the other replacement phrase is used as the replacement vocabulary. Replacement phrase. 如請求項3所述的試題產生方法，其中，子步驟(F-7)包含以下子步驟： (F-7-1)計算該替換詞組之一替換詞向量； (F-7-2)計算該答案詞組之一答案詞向量；及 (F-7-3)判定該替換詞向量與該答案詞向量之相似度是否大於一基準值，以判定是否將該替換詞組作為該等誘答詞組之其中一者。 The method for generating test questions as described in claim item 3, wherein the sub-step (F-7) includes the following sub-steps: (F-7-1) calculating one of the replacement word vectors for the replacement phrase; (F-7-2) calculating one of the answer word vectors of the answer phrase; and (F-7-3) Determine whether the similarity between the replacement word vector and the answer word vector is greater than a reference value, so as to determine whether the replacement phrase is one of the decoy phrases. 如請求項3所述的試題產生方法，其中，在子步驟(F-7)後，還包含以下子步驟： (F-17)當判定出不將該替換詞組作為該等誘答詞組之其中一者時，自該答案詞組選擇另一目標詞，並回到步驟(F-2)。 The test question generation method as described in claim item 3, wherein, after the sub-step (F-7), the following sub-steps are also included: (F-17) When it is determined that the replacement phrase is not used as one of the decoy phrases, select another target word from the answer phrase, and return to step (F-2). 如請求項3所述的試題產生方法，其中，在步驟(F-2)中，係將位於該目標詞前的相鄰詞與該目標詞組成該等目標詞組合之其中一者，並將位於該目標詞後的相鄰詞與該目標詞組成該等目標詞組合之其中另一者。The method for generating test questions as described in claim item 3, wherein, in step (F-2), the adjacent words before the target word and the target word are used to form one of the target word combinations, and The adjacent words after the target word and the target word form the other of the target word combinations. 如請求項6所述的試題產生方法，其中，在步驟(F-3)中，對於每一目標詞組合，係根據該目標詞組合中之目標詞的目標詞向量及該目標詞組合中之相鄰詞的相鄰詞向量，計算出該目標詞組合之一待配對詞向量。The test question generation method as described in claim item 6, wherein, in step (F-3), for each target word combination, it is based on the target word vector of the target word in the target word combination and the target word combination Adjacent word vectors of adjacent words, one of the target word combinations to be paired word vectors is calculated. 如請求項6所述的試題產生方法，該運算裝置還儲存有一詞向量轉換模型，其中，在步驟(F-3)中，該目標詞向量係藉由將該目標詞利用該詞向量轉換模型而轉換出，該相鄰詞向量係藉由將該相鄰詞利用該詞向量轉換模型而轉換出，且在步驟(F-4)中，每一詞彙向量係藉由將所對應之詞彙利用該詞向量轉換模型而轉換出。As in the test question generation method described in claim 6, the computing device also stores a word vector conversion model, wherein, in step (F-3), the target word vector is obtained by using the target word using the word vector conversion model And converted, the adjacent word vector is converted by using the adjacent word using the word vector conversion model, and in step (F-4), each vocabulary vector is converted by using the corresponding vocabulary The word vector conversion model is converted out. 如請求項1所述的試題產生方法，在步驟(A)之前，還包含以下步驟： (G)將該文字檔進行文字清理及字形轉換，以獲得轉換後的該文字檔； 其中，在步驟(A)中，係將轉換後的該文字檔進行段落切分。 The method for generating test questions as described in claim item 1, before step (A), also includes the following steps: (G) performing text cleaning and glyph conversion on the text file to obtain the converted text file; Wherein, in the step (A), the converted text file is segmented into paragraphs. 如請求項1所述的試題產生方法，每一關係詞還對應一問題構句，其中，在步驟(E)中，係根據排除該答案詞組的該描述句、該目標關係詞及該目標關係詞所對應之問題構句，產生該試題。As in the test question generation method described in claim item 1, each relational word also corresponds to a question to form a sentence, wherein, in step (E), it is based on the descriptive sentence, the target relational word and the target relationship that exclude the answer phrase The questions corresponding to the words are constructed into sentences to generate the test questions.
TW110145307A 2021-12-03 2021-12-03 Item generating method TWI773604B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW110145307A TWI773604B (en) 2021-12-03 2021-12-03 Item generating method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW110145307A TWI773604B (en) 2021-12-03 2021-12-03 Item generating method

Publications (2)

Publication Number Publication Date
TWI773604B TWI773604B (en) 2022-08-01
TW202324185A true TW202324185A (en) 2023-06-16

Family

ID=83806968

Family Applications (1)

Application Number Title Priority Date Filing Date
TW110145307A TWI773604B (en) 2021-12-03 2021-12-03 Item generating method

Country Status (1)

Country Link
TW (1) TWI773604B (en)

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI396100B (en) * 2007-10-12 2013-05-11 Hon Hai Prec Ind Co Ltd System and method for generating examination paper
CN102789380B (en) * 2011-05-20 2015-07-29 智学馆科技有限公司 The generation method of electronic test paper
TW201301059A (en) * 2011-06-21 2013-01-01 Univ Nan Kai Technology System for choosing examination questions from file and method thereof
TW201349159A (en) * 2012-05-31 2013-12-01 Han Lin Publishing Co Ltd Method for generating learning test questions and system thereof
TWM546586U (en) * 2017-03-16 2017-08-01 Kang Hsuan Educational Publish Corp Test making and scoring system of examination paper
TWI663583B (en) * 2018-08-13 2019-06-21 國立臺灣師範大學 Automatic scoring method and system for divergent thinking test

Also Published As

Publication number Publication date
TWI773604B (en) 2022-08-01

Similar Documents

Publication Publication Date Title
CN111310458B (en) Subjective question automatic scoring method based on multi-feature fusion
US9342499B2 (en) Round-trip translation for automated grammatical error correction
Sukkarieh et al. Automarking: using computational linguistics to score short ‚free− text responses
US9443513B2 (en) System and method for automated detection of plagiarized spoken responses
CN110851599A (en) Automatic scoring method and teaching and assisting system for Chinese composition
Zhang et al. HANSpeller++: A unified framework for Chinese spelling correction
Kwankajornkiet et al. Automatic multiple-choice question generation from Thai text
Chandra et al. A survey on non-english question answering dataset
Lim et al. Neural automated writing evaluation for Korean L2 writing
Hussain et al. Assessment of bangla descriptive answer script digitally
TWI773604B (en) Item generating method
Björklund et al. Computer, enhence: POS-tagging improvements for nonbinary pronoun use in Swedish
CN111814433B (en) Uygur language entity identification method and device and electronic equipment
Zhao et al. A study on the innovative model of foreign language teaching in universities using big data corpus
CN110674871B (en) Translation-oriented automatic scoring method and automatic scoring system
Khandait et al. Automatic question generation through word vector synchronization using lamma
Liu Approach of the corpus research on medical English prepositions
Dinh et al. Transliterating nôm scripts into Vietnamese national scripts using statistical machine translation
Wu Automatic English essay scoring algorithm based on machine learning
Contractor et al. Text retrieval using sms queries: Datasets and overview of fire 2011 track on sms-based faq retrieval
Bear et al. Evaluating unsupervised approaches to morphological segmentation for wolastoqey
Li An automated English translation judging system based on feature extraction algorithm
Irawati Extensive reading on accuracy, acceptability, and readability in translation study
Rytting et al. Error Correction for Arabic Dictionary Lookup.
Agustawan Borrowing Keep and Its Translation Found In The Text of “Puputan Badung”