JP2011013811A - Device, system and method for word level decision, control program, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To objectively perform level decision of a Japanese word with high accuracy.SOLUTION: A word level decision device includes: a character feature acquisition unit 102 for acquiring a character feature for each character included in the word input to local equipment, from a character feature memory 103 for storing the character feature indicating a characteristic of a character in each character; a word feature acquisition unit 104 for acquiring a word feature of the input word from a word feature memory 105 for storing the word feature indicating a characteristic of the word in each word; and a word level value calculation unit 106 for calculating a word level value indicating difficulty of the word in the input word from a character total feature value obtained by totaling the character features of the respective characters included in the input word acquired by the character feature acquisition unit 102, and the word feature acquired by the word feature acquisition means.

Description

  The present invention relates to a language processing technique for evaluating a word level (difficulty level, etc.), and in particular, a word level determination device, a word level determination system, and a word level determination method that automatically determine the difficulty level of a Japanese word. , A control program, and a recording medium.

  Conventionally, techniques for automatically determining the readability and difficulty of sentences have been widely used (see, for example, Patent Documents 1 and 2 and Non-Patent Documents 1 to 3). Such technology is regarded as important in language education, for example. As an example, it is possible to determine the readability of a sentence by the above technique and appropriately select a teaching material including a vocabulary according to a learner's learning stage.

  Here, it is important to learn kanji and master an enormous amount of words in order to learn Japanese using kanji and spelling sentences with a wide variety of words. However, there are great difficulties for Japanese learners to acquire these skills.

  Therefore, in order to realize more effective learning support, the selection of teaching materials uses not only the readability based on the superficial features of the text but also the semantic content of the word and the word. It is important to perform level determination in units of words in consideration of usage characteristics (appearance frequency, familiarity, etc.). Patent Documents 1 and 2 describe that the difficulty level of a sentence is determined by referring to a predetermined level for each word, but that the level determination of a word itself is automatically performed. Therefore, a dictionary to which a level is assigned for each word must be held in advance.

  On the other hand, Non-Patent Document 4 discloses a technique for automatically determining a level (degree of difficulty) in units of words. Non-Patent Document 4 proposes a detailed word level determination system that takes into account the meaning and usage characteristics of words.

  Specifically, the above system uses three types of information such as the frequency of words appearing in newspaper articles, the word familiarity indicating the degree of familiarity with words, and the question range for each class in the Japanese Language Proficiency Test. Judging the level.

JP 2006-244252 A (published September 14, 2006) JP 2007-121484 A (published May 17, 2007)

Yuika Tateishi, two others, “Evaluation formula for readability of Japanese text”, document processing and human interface 18-4, May 9, 1998 Hideko Shibasaki and one other, “Basic research for building Japanese readability by applying a corpus of Japanese language textbooks”, IEICE Technical Report, p. 19-24, October 2007 Sato Satoshi, “Measure the difficulty of Japanese text”, Monthly Language, Vol. 37, no. 8, P.I. 54-57, August 1, 2008 Yoshiko Kawamura and 1 other, “Development of word familiarity checker for determining difficulty of sentences”, Journal of Japanese Language Education Methodology, Vol. 15, No. 2, P.I. 24-25, September 20, 2008 Ono Rikio, “Multimedia Computing”, Corona Inc., October 24, 2008, p. 182-185

  However, the conventional system causes the following problems.

  Specifically, only the features obtained in units of words, such as appearance frequency, word familiarity, and question range, are considered, and objective features (character complexity, etc.) for each character constituting the word are considered. Not considered.

  The word familiarity is an index obtained from the subjective judgment of the subject, such as “sense of feeling” and “familiarity”, through a survey. Although this is one of the important indicators for determining the difficulty level, it may not be suitable as an indicator for maintaining unbiasedness and objectivity depending on the scope and timing of the investigation. In addition, the range of questions is limited, and the range of words covered cannot be flexibly dealt with vocabulary that changes while being used every day.

  For the above reasons, the conventional system has a problem that the level of words cannot be determined objectively and accurately.

  As described above, learning an enormous amount of kanji and words is a difficult task for the learner. In order to support such a learner, it is more important to objectively and accurately determine the level of the word in consideration of not only the word unit but also the character difficulty level in character units. It becomes a problem.

  Note that the above-mentioned problem does not occur only in the field of Japanese language education / learning, but occurs in situations where the level of Japanese words is determined in all fields that use Japanese.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to provide a word level determination device, a word level determination system, and a word level determination that objectively and accurately determine a word level in a Japanese sentence. To realize a method, a control program, and a recording medium.

  In order to solve the above-described problem, the word level determination device of the present invention provides a character feature amount indicating a character feature for each character, and a character feature amount storage unit for each character included in a word input to the device. In addition, a word feature amount acquisition unit that acquires a character feature amount, and a word feature amount acquisition unit that stores a word feature amount indicating a word feature for each word, a word feature that acquires a word feature amount of the input word A character total feature amount obtained by combining the character feature amount acquired by the character feature amount acquisition unit for each character included in the input word, and the word feature amount acquisition unit It is characterized by comprising a word level value calculating means for calculating a word level value indicating the difficulty of the word in the inputted word from the acquired word feature amount.

  According to the above configuration, when a word is input to the word level determination device, the character feature amount acquisition unit and the word feature amount acquisition unit respectively acquire the character feature amount and the word feature amount. The character feature amount acquisition unit acquires a character feature amount from the character feature amount storage unit for each character included in the input word. The word feature amount acquisition unit acquires the word feature amount stored in association with the input word from the word feature amount storage unit.

  Next, the word level value calculating means, from the character total feature amount obtained by combining the character feature amount acquired for each character of the input word, and the word feature amount acquired for each word, The word level value of the input word is calculated. The word level value is a value indicating the difficulty of the word.

  The character feature amount obtained from the character feature is an objective and fair measure that is not influenced by the subjective measure of the person, such as the word sense and familiarity. According to the above configuration, an objective word level value can be calculated by considering a character feature amount in addition to a word feature amount, so that an objective and accurate word level determination can be realized. It becomes possible.

  The word level determination device further includes a word extraction unit that extracts a word from a Japanese sentence input to the device, and the character feature amount acquisition unit acquires a character feature amount for each character included in the word. Is performed for each word extracted by the word extraction unit, and the word feature amount acquisition unit performs a process of acquiring a word feature amount of the word for each word extracted by the word extraction unit. The value calculation means includes, for each word extracted by the word extraction means, the character total feature quantity obtained by integrating the character feature quantity, and the word feature quantity acquired for each extracted word. From the above, it is preferable to calculate the word level value for each extracted word.

  According to the above configuration, when a sentence is input to the word level determination device, the word extracting unit extracts a word from the input Japanese sentence. The word extraction means extracts one or more words included in the text data of the Japanese sentence. Next, the character feature amount acquisition unit and the word feature amount acquisition unit respectively acquire the character feature amount and the word feature amount. Here, the character feature amount acquisition unit performs the process of acquiring the character feature amount for each character included in the word for each word extracted by the word extraction unit. Therefore, the character total feature amount is obtained for each extracted word. Further, the word feature amount acquisition unit performs a process of acquiring a word feature amount of the word for each word extracted by the word extraction unit.

  Then, the word level value calculating means calculates a word level value for each extracted word from the character total feature amount and the word feature amount obtained for each extracted word.

  Thereby, objective and accurate word level determination can be realized for each of a plurality of words included in the sentence.

  If it becomes possible to collectively determine the word level for each word in the sentence, it will be possible to support Japanese teachers in the field of Japanese language education. In the past, teachers have rewritten and used materials that serve as Japanese teaching materials (for example, newspaper articles) according to the learner's skills. However, this word replacement and sentence proofing are laborious tasks for teachers. This is because, after appropriately determining the level of the word in the material, it is necessary to search for an appropriate (more simple or difficult) word according to the skill of the learner.

  According to the present invention, a teacher can obtain an objective word level value simply by inputting a sentence to the word level determination device of the present invention. The degree can be easily and objectively determined. This makes it easy for the teacher to determine whether or not the learning material corresponds to the learner's level and to select the learning material according to the level.

  The word level determination apparatus further includes a correspondence information storage unit that stores correspondence information in which the word level value is associated with a word difficulty level that is a character string representing the difficulty of the word to be presented to the user; It is preferable that a word difficulty level acquisition unit that acquires a word difficulty level associated with the word level value calculated by the word level value calculation unit from the correspondence information storage unit.

  According to the above configuration, the word difficulty level acquisition unit refers to the correspondence information stored in the correspondence information storage unit and acquires the word difficulty level corresponding to the word level value calculated by the word level value calculation unit. The correspondence information has a data structure in which the word level value is associated with a word difficulty level, which is a character string expressed to present the difficulty of the word to the user. Therefore, the word difficulty level acquisition means can obtain the word difficulty level to be presented to the user from the word level value by referring to the correspondence information.

  As a result, the word difficulty level representing the difficulty of the word as a character string is expressed in a manner that is intuitively understandable to the user, compared to the word level value indicating the difficulty, and the degree of difficulty of the word is indicated to the user. Can be presented. The user can obtain a determination result in which the objectively calculated word level value is easily expressed by a character string such as a rank, a grade, and a word, for example, so that convenience for the user is further improved.

  Furthermore, the correspondence information is stored in the correspondence information storage unit, so that the user can easily create and edit the correspondence between the word level value and the word difficulty character string and the contents of the character string in the correspondence information. It becomes possible.

  For example, when a teacher who provides Japanese language learning materials to a learner is a user, the teacher can grasp the difficulty level of words for learners in Japanese language education from experience in the education field. By creating and storing correspondence information reflecting the opinions of such teachers, it is possible to use correspondence information that is the best standard for Japanese language education when determining the difficulty of words in Japanese teaching materials. And a more appropriate difficulty level determination can be performed.

  The character feature amount storage unit may store the number of character strokes as a character feature amount, and the character feature amount acquisition unit may acquire the number of character strokes from the character feature amount storage unit.

  As a result, the word level value can be calculated by using an objective and fair scale that is not affected by the human subjective scale, ie, the number of character strokes, as the character feature amount. Can be performed with high accuracy.

  The word feature storage unit stores a word IDF (Inverse Document Frequency) as a word feature, and the word feature acquisition unit acquires the IDF of the word from the word feature storage unit. May be.

  IDF (Inverse Document Frequency) is a measure representing the importance of words in the entire document set. Similar to the appearance frequency of words appearing in a document, in IDF, when a word appears in the entire document set, the appearance of the word is given a greater weight with respect to a word that appears frequently in a specific document. give.

  By using such an IDF as a word feature amount, since the appearance frequency is high, the word is a specific document for a word that tends to be judged as an easy word because it is widely used in general. If it appears frequently only in a specific article (such as a specific article in a newspaper), it can be determined that it is a more difficult word as a special word.

  Therefore, by considering the bias of appearance, it is possible to perform the level determination of the word with higher accuracy than in the case of determining based on the simple appearance frequency.

  Further, the wider the document set for obtaining the IDF, the more words can be covered from many documents, and the bias can be correctly determined. Can be requested. Therefore, the level determination of the word can be performed with higher accuracy than the determination using the question range in which the word is limited.

  In particular, if a document set serving as a corpus for obtaining an IDF is widely and regularly obtained from recent newspaper articles or documents published on the Internet, based on the periodically updated IDF, The word level value of the word can be calculated.

  Since recent newspaper articles and documents on the Internet are frequently updated, it is considered that the documents reflect the current Japanese usage characteristics faithfully. By obtaining the IDF from a document set including many such documents, the IDF reflects the current Japanese usage characteristics. Therefore, by using the IDF, it becomes possible to perform word level determination in response to daily vocabulary changes, trend changes, and the like.

  Furthermore, by adopting more documents on the Internet as a corpus, IDF can be used not only from hard sentences created by specific authors such as newspaper articles but also from everyday sentences written by people of various positions. Can be obtained. This makes it possible to perform fair word level determination without bias.

  The character feature amount storage unit stores the number of character strokes as a character feature amount, and the character feature amount acquisition unit calculates the number of character strokes acquired from the character feature amount storage unit for each word. The total number of words is calculated by combining the words, and the word feature storage unit stores the IDF of the word as the word feature, and the word feature acquisition unit receives the word feature from the word feature storage unit. An IDF (Inverse Document Frequency) of a word is acquired, and the word level value calculating means performs at least one of an addition process, a multiplication process, and a weighted average process on the word total stroke count and the IDF, Preferably, the word level value is calculated.

  According to the above configuration, addition, multiplication, weighted average, etc. are added to the total number of strokes (character total feature amount) for each word and IDF (word feature amount) for each word obtained from the number of strokes of objective characters. It is possible to calculate the word level value by integrating at least one of the above arithmetic processes.

  Therefore, objective and accurate word level determination can be easily realized.

  The correspondence information storage unit further stores the correspondence information for each learning stage of the learner, and the word difficulty level acquisition unit learns based on the learner information about the learner input to the device. In the correspondence information corresponding to the designated learning stage, the word difficulty level associated with the calculated word level value is preferably acquired from the correspondence information storage unit.

  According to the above configuration, the correspondence information storage unit includes a plurality of pieces of correspondence information indicating a correspondence relationship between the word level value and the word difficulty level instead of one. Each correspondence information is provided for each learning stage, and shows various correspondences depending on the learning stage.

  Specifically, for learners who are not learning at the same word level value (learning level is low), there is correspondence information that associates a high difficulty level, and for learners with a high learning level There is also correspondence information that associates a low difficulty level. This is an attempt to determine the difficulty level in consideration of learners at various stages of learning when converting word level values, which are objective numerical values, into subjective difficulty levels for learners. is there.

  The word difficulty level acquisition means, when performing the word level determination for the input sentence or word, if the learner's learning stage is specified based on the input learner information, the plurality of types of correspondence information described above The correspondence information corresponding to the designated learning stage is referred to, and the character string of the word difficulty level corresponding to the word level value is acquired in the correspondence information. The learner information is information related to the learner serving as a criterion for determining the difficulty level, and is any information that enables the learning stage to be specified. As an example, the learner information may be information indicating the learning stage itself, or identification information for identifying the learner, and the learning stage may be specified from the identification information.

  Thereby, based on the said word level value, a word difficulty level can be determined according to a learner's skill. Therefore, user convenience is further improved.

  In order to solve the above problems, a word level determination system of the present invention includes the above word level determination device and a terminal device that communicates with the word level determination device, and the terminal device is determined by a user of the terminal device. The word level determination device requests the word level value of the input word, and the word level determination device sends the word level value calculated for the word received from the terminal device to the terminal device as a response to the request. It is characterized by transmitting.

  According to the above configuration, the above-described word level determination device of the present invention can be applied not only to a stand-alone information processing device but also to a communication network system in which a plurality of devices are connected to be able to communicate with each other. The word level determination system constructed in this way also falls within the scope of the present invention.

  In order to solve the above problems, the word level determination method of the present invention is included in a word input to the word level determination device from a character feature amount storage unit that stores a character feature amount indicating a character feature for each character. For each character, the character feature amount acquisition step for acquiring the character feature amount and the word feature amount storage unit for storing the word feature amount indicating the word feature for each word acquire the word feature amount of the input word. A word feature amount acquisition step, a character total feature amount obtained by combining the character feature amounts acquired in the character feature amount acquisition step of each character included in the input word, and the word feature amount A word level value calculating step of calculating a word level value indicating the difficulty of the word in the input word from the word feature amount acquired in the acquiring step, To have.

  According to the above method, an objective character feature amount can be adopted together with a word feature amount, and a word level value can be calculated. Therefore, an objective and accurate word level determination method can be realized.

  The word level determination device and the word level determination system may be realized by a computer. In this case, the word level determination device or the word level determination system is operated by causing the computer to operate as each means. The control program for the word level determination apparatus or the word level determination system realized by a computer and a computer-readable recording medium on which the control program is recorded also fall within the scope of the present invention.

  In order to solve the above-described problem, the word level determination device of the present invention provides a character feature amount indicating a character feature for each character, and a character feature amount storage unit for each character included in a word input to the device. In addition, a word feature amount acquisition unit that acquires a character feature amount, and a word feature amount acquisition unit that stores a word feature amount indicating a word feature for each word, a word feature that acquires a word feature amount of the input word A character total feature amount obtained by combining the character feature amount acquired by the character feature amount acquisition unit for each character included in the input word, and the word feature amount acquisition unit It is characterized by comprising a word level value calculating means for calculating a word level value indicating the difficulty of the word in the inputted word from the acquired word feature amount.

  In order to solve the above problems, a word level determination system of the present invention includes the above word level determination device and a terminal device that communicates with the word level determination device, and the terminal device is determined by a user of the terminal device. The word level determination device requests the word level value of the input word, and the word level determination device sends the word level value calculated for the word received from the terminal device to the terminal device as a response to the request. It is characterized by transmitting.

  In order to solve the above problems, the word level determination method of the present invention is included in a word input to the word level determination device from a character feature amount storage unit that stores a character feature amount indicating a character feature for each character. For each character, the character feature amount acquisition step for acquiring the character feature amount and the word feature amount storage unit for storing the word feature amount indicating the word feature for each word acquire the word feature amount of the input word. A word feature amount acquisition step, a character total feature amount obtained by combining the character feature amounts acquired in the character feature amount acquisition step of each character included in the input word, and the word feature amount A word level value calculating step of calculating a word level value indicating the difficulty of the word in the input word from the word feature amount acquired in the acquiring step, To have.

  Thereby, an objective character feature amount can be adopted together with a word feature amount, and a word level value can be calculated, so that an objective word level determination can be performed with high accuracy.

It is a figure which shows typically the word level determination system 100 of this invention. It is a block diagram which shows the principal part structure of the word level determination apparatus (personal computer 1) in embodiment of this invention. (A)-(c) is a figure which shows the specific example of the table memorize | stored in the stroke number database 23. FIG. It is a figure which shows the specific example of the table memorize | stored in the IDF database. 6 is a diagram illustrating a specific example of a table stored in a storage unit 11 as a correspondence information storage unit 108. FIG. It is a flowchart which shows the flow of the word level determination process of the word level determination apparatus in this embodiment. It is a figure which shows the other specific example of the table recorded on the memory | storage part 11 as the corresponding | compatible information storage part. It is a figure which shows the specific example of the learner information table memorize | stored in the learner information storage part. It is a flowchart which shows the flow of the word difficulty level determination process of the word level determination apparatus in this embodiment. It is a figure which shows an example of the GUI screen for inputting a text. It is a figure which shows an example of the GUI screen for showing a word level determination result to a user.

  An embodiment of the present invention will be described below with reference to the drawings.

[Outline of word level judgment system]
First, with reference to FIG. 1, the word level determination system which implement | achieves the word level determination method of this invention is demonstrated. The word level determination system of the present invention may be realized by a stand-alone information processing device, or by a network system constructed by a plurality of information processing devices and storage devices connected to each other via a communication network. It may be realized.

  FIG. 1 is a diagram schematically showing a word level determination system 100 of the present invention. As shown in FIG. 1, the word level determination system 100 includes a word extraction unit 101, a character feature amount acquisition unit 102, a word feature amount acquisition unit 104, and a word level value calculation unit (word level value calculation means) as functional blocks. ) 106, and more preferably a word difficulty determination unit (word difficulty acquisition means) 107. The word level determination system 100 further includes at least a character feature amount storage unit 103 and a word feature amount storage unit 105 that each functional block of the system reads / writes, and more preferably, includes a correspondence information storage unit 108. Contains.

  The word extraction unit 101 extracts words from a sentence (text data) D1 input to the word level determination system 100 by a user according to Japanese grammar. The word extraction unit 101 can appropriately employ a conventionally known technique such as morphological analysis software, for example. Information on the word (morpheme) extracted by the word extraction unit 101 is supplied as word information D2 to the following feature amount acquisition units.

  The character feature amount acquisition unit 102 acquires, from the character feature amount storage unit 103, a character unit feature amount (character feature amount) representing the character feature of each character constituting the word extracted by the word extraction unit 101. Is. In one embodiment of the present invention, the character feature amount acquisition unit 102 calculates the character total feature amount D3 by combining the character feature amounts acquired for each character into one for each word. There is no particular limitation on how to combine the character feature values of each character for each word, but an example of a method in which the character feature value acquisition unit 102 calculates the character total feature value D3 for each word is described in the first embodiment. Will be described in detail. In addition, this invention is not limited to the said structure. The character feature amount acquisition unit 102 may supply the character feature amount of each character of the input word to the word level value calculation unit 106.

  The character feature amount includes the number of character strokes, the character type, the character complexity, and the like, and the character feature amount acquisition unit 102 can extract them as a character feature amount alone or in combination. As a method for extracting the character complexity, it is conceivable to digitize the character shape and obtain the character complexity numerically. For example, for one character of a certain font, the complexity of the character is objectively digitized from the ratio of the character pixel to the background pixel, the variation of the character pixel, and the like. More specifically, in the word “spoken language”, the character “mouth” has a simple shape in which the pixels of the character are arranged to form four lines, and the character “word” has more characters. Pixels are sparsely arranged, and a complicated shape in which a number of short lines are combined in a complex manner. Therefore, the character feature amount acquisition unit 102 assigns a larger value indicating the character complexity to the character “word” based on the number of pixels of the character and its variation. Thereby, it can be determined that the character “word” is more complicated than the character “mouth”.

  The character feature amount storage unit 103 stores character features in advance for each character. The character feature amount acquisition unit 102 stores one or more features stored for each character from the character feature amount storage unit 103. By reading, the character feature amount may be obtained for every character constituting the word.

  The word feature amount acquisition unit 104 acquires, from the word feature amount storage unit 105, a word unit feature amount (word feature amount) that represents the feature of the word for the word extracted by the word extraction unit 101. As the word feature amount, usage characteristics such as appearance frequency (number of words appearing in the document), familiarity (vocabulary friendliness), test range, etc. may be extracted, but more preferably IDF (Inverse A measure representing the importance of a word such as (Document Frequency) is acquired as a word feature amount. The word feature amount acquisition unit 104 obtains a word feature amount by combining these feature amounts alone or in combination. Details of the IDF will be described later. The word feature value D4 acquired by the word feature value acquisition unit 104 is supplied to the word level value calculation unit 106.

  The word feature quantity storage unit 105 stores word features in advance for each word. The word feature quantity acquisition unit 104 stores one or more features stored for each word from the word feature quantity storage unit 105. Read and acquire the word feature amount of each word in the sentence.

  The word level value calculation unit 106 outputs the character total feature amount D3 (or the character feature amount of each character of the word) output from the character feature amount acquisition unit 102 and the word feature amount D4 extracted by the word feature amount acquisition unit 104. And a word level value D5 that is a scale for determining the difficulty of the word is calculated.

  The word level value calculation unit 106 makes each feature quantity so that the numerical value of each feature quantity obtained from each feature quantity acquisition unit becomes a numerical value indicating the degree of difficulty of the word, that is, the word level value, by the calculation. Execute the process of integrating. The calculation method employed for the integration by the word level value calculation unit 106 is not particularly limited, and an optimum one is used as appropriate depending on the characteristics of the obtained feature amount. For example, the word level value calculation unit 106 performs arithmetic processing (for example, addition processing, multiplication processing, average value calculation processing, and / or weighted average calculation processing) on the obtained character feature amount and / or word feature amount. Etc.) to calculate the word level value D5. In the present embodiment, the word level value calculation unit 106 performs arithmetic processing so that the more difficult the word is, the higher the word level value is, but the present invention is not limited to this.

  The word level value D5 calculated by the word level value calculation unit 106 may be output as it is and presented to the user.

  The word difficulty level determination unit 107 refers to the correspondence information storage unit 108 and determines the word difficulty level based on the word level value D5 calculated by the word level value calculation unit 106. The word difficulty level is a character string obtained by replacing the word level value D5 that expresses the difficulty of the word with a numerical value into an expression that is easy to understand for the user by dividing the difficulty of the word into several stages. For example, word level values are roughly classified (for example, about 5 categories), and the level of ease and difficulty of words is expressed in stages (for example, in a 5-level display) in an intuitively understandable expression for the user. To do. Specifically, the correspondence information storage unit 108 has five levels as word difficulty levels, that is, level A (simple), level B (slightly simple), level C (normal), level D (slightly difficult), Level E (difficult) is determined in advance, and each word difficulty level is associated with the numerical value of the word level value and stored as correspondence information.

  The word difficulty level determination unit 107 calculates the word level value calculated by the word level value calculation unit 106 based on the correspondence information indicating the correspondence relationship between the word level value and the word difficulty level stored in the correspondence information storage unit 108. The word difficulty level D6 can be specified from D5, and the word difficulty level can be determined. The word difficulty level D6 determined by the word difficulty level determination unit 107 is output for presentation to the user as a character string. As a result, the word level determination result can be presented to the user in a form that is easier for the user to understand.

  Note that the word difficulty level D6 is different from a quantitative value such as the word level value D5, and uses an intuitive expression. Therefore, depending on the learner's Japanese skills, the word difficulty level D6 is not necessarily an expression suitable for all learners. It may not be possible. Therefore, in order to eliminate such inconvenience, the word difficulty level determination unit 107 may adjust the correspondence between the word level value D5 and the word difficulty level D6 according to the Japanese language skill of the Japanese learner. Good. This can be realized by the word difficulty level determination unit 107 acquiring learner information (hereinafter referred to as learner information D7). Details of the method for determining the word difficulty level D6 will be described later.

  Furthermore, the word level determination system 100 determines the difficulty level of the entire sentence based on the word level values or word difficulty levels of all (or a part of) words included in the input sentence D1. A determination unit (not shown) may be provided. The sentence difficulty determined by the sentence difficulty determination unit may be presented to the user in a form that is easy for the user to understand.

  According to the above configuration, the word level (difficulty level) can be determined more objectively and accurately by integrating the character total feature amount D3 (or each character feature amount) and the word feature amount D4. It becomes. Hereinafter, an embodiment of the present invention will be described in more detail using specific examples.

Embodiment 1
Next, an embodiment in which the word level determination system 100 of the present invention shown in FIG. 1 is realized by a stand-alone word level determination device will be described. In this embodiment, the word level determination apparatus of the present invention is applied to a personal computer (hereinafter referred to as a personal computer) as an example. The word level determination device of the present invention is not limited to a personal computer, and can be applied to any information processing device capable of inputting and outputting Japanese (for example, an electronic dictionary, a digital TV, a home (portable) ) Game console, mobile phone, PDA (Personal Digital Assistant), etc.).

[Configuration of word level judgment device (PC)]
FIG. 2 is a block diagram showing a main configuration of the personal computer (word level determination device) 1 according to the embodiment of the present invention. As shown in FIG. 2, the personal computer ((word level determination system / word level determination device) 1 includes a control unit 10, a storage unit 11, an input unit 12, a display unit 13, and a communication unit 14.

  The input unit 12 is an input device for inputting the text D1 to the personal computer 1 in a state that the personal computer 1 can read, that is, text data. In this embodiment, the input part 12 shall be comprised by the keyboard, the touchpad, the mouse | mouth etc. as an example. Alternatively, when the personal computer 1 has a voice recognition function for obtaining text data from voice data or a character recognition function for obtaining text data from image data, a microphone, a scanner, or the like is applied as the input unit 12. It is possible. The input unit 12 can also be used to input the learner information D7 into the personal computer 1.

  Note that the text data may be acquired from an e-mail, a website, or the like via the communication unit 14.

  The display unit 13 is a display device for displaying determination results such as the word level value D5, the word difficulty level D6, and the text difficulty level, which are output from the functional block of the control unit 10. Further, when the tool for inputting the text D1 is realized by a graphical user interface (GUI), the display unit 13 displays the GUI screen. The display unit 13 is realized by, for example, an LCD (liquid crystal display).

  The communication unit 14 transmits and receives data to and from other devices via a wide area communication network such as the Internet or a local area network (LAN) or a local area communication network such as a wireless LAN. In the present embodiment, the communication unit 14 downloads various websites and files published on the Internet, receives electronic newspaper articles, etc., and becomes a document serving as a reference corpus in the personal computer 1. Get a set. The document set D8 acquired by the communication unit 14 in this manner is used to derive the characteristics of Japanese words (IDF in this embodiment).

  The storage unit 11 stores a control program executed by the control unit 10, an OS program, and various data read when the control unit 10 executes various functions of the personal computer 1. In this embodiment, an application program for determining a word level is stored, and information used in the application (at least the stroke number database 23, the IDF database 35, and the correspondence information storage unit 108, preferably further learner information storage). Part 38) is stored. The data structure of these various information will be described in detail later.

  The control unit 10 performs overall control of each unit included in the personal computer 1. The control unit 10 includes a morphological analysis unit (word extraction unit) 21, a stroke number determination unit (character feature amount acquisition unit) 22, an IDF acquisition unit (word feature amount acquisition unit) 24a, and a word level value calculation unit 106 as functional blocks. And a word difficulty level determination unit 107. Preferably, an IDF calculation unit (IDF calculation means) 24b may be provided. Furthermore, you may provide the sentence difficulty level determination part not shown. In addition, each said part shown as a functional block is a RAM (random access memory) etc. in the said application program memorize | stored in memory | storage devices (memory | storage part 11), such as ROM (read only memory) CPU (central processing unit). This is realized by reading and executing, thereby providing the word level determination device of the present invention.

  In the personal computer 1 in the present embodiment, the word extraction unit 101 is realized as the morphological analysis unit 21. The morpheme analysis unit 21 performs morpheme analysis on the input sentence D1 and divides the sentence into words. The morpheme analysis unit 21 supplies the obtained word (morpheme) to the stroke number determination unit 22 as word information D2 together with information on the part of speech.

  In the present embodiment, the character feature amount acquisition unit 102 is realized as the stroke number determination unit 22, and the character feature amount storage unit 103 is realized as the stroke number database 23. The stroke number determination unit 22 acquires the number of strokes of each character constituting the word for each word analyzed by the morphological analysis unit 21. In the present embodiment, these are further integrated to determine the total number of strokes of the word (hereinafter referred to as the total number of words) as the character total feature amount D3. The stroke number determination unit 22 is not limited to simply adding the stroke number of each character when summing the number of strokes for each character as the total number of strokes (character total feature amount) for one word, and refers to the stroke number database 23. It is more preferable to obtain the total number of strokes after weighting characters according to the character type, the part of speech of the word, and the like.

  3A to 3C are diagrams showing specific examples of tables stored in the stroke number database 23. FIG. Specifically, FIG. 3A is an example of a stroke number database, and an actual stroke number is stored in association with each character. FIG. 3B is an example of a rule table for assigning weights to the actual number of strokes for each character type, and rules for assigning weights to the actual number of strokes for each character type are associated with each other. It is remembered. FIG. 3C is an example of a rule table for assigning a weight to the actual number of strokes for each part of speech of a word including the character, and rules for how to weight each part of speech. Are stored in association with each other.

  For example, it is assumed that the stroke number determination unit 22 acquires the word “love” and the word “ga (particle)”. The stroke number determination unit 22 determines that the total number of strokes of the word = 13 × 2 × 2 = 52 for the word “love” with reference to the tables of FIGS. The stroke number determination unit 22 selects the stroke number “13” from the characters “Ai” in the table of FIG. 3A and the character type “Kanji” in the table of FIG. By acquiring “× 2” from the part-of-speech “noun” in the table of FIG. On the other hand, the word “ga (particle)” is determined as the total number of strokes = 1 × 0 = 0 even if the actual number of strokes is five. The stroke number determination unit 22 obtains “constant 1” from the character type “Hiragana” in the table of FIG. 3B and “× 0” from the part of speech “particle” in the table of FIG. By doing so, the above-mentioned total number of strokes of the word can be obtained.

  Major independent words such as nouns, verbs, and adjectives have a greater influence on the difficulty of sentences than adjuncts such as particles, and are considered more suitable for judging the difficulty of teaching materials. Therefore, by ignoring the word level value of the attached word or calculating extremely low as described above, it is possible to efficiently realize the determination of the word level (sentence difficulty level) with higher accuracy.

  Moreover, as shown in the table of FIG. 3B, hiragana and katakana in a word may be counted as one stroke regardless of the number of strokes. This is based on the idea that it is the most difficult for Japanese learners to learn kanji, and that hiragana and katakana words are less difficult. Therefore, the above configuration is particularly effective in a scene where the difficulty level of a word in a sentence is measured as a teaching material for Japanese language education because an effective numerical value can be obtained for the total number of word strokes and the word level value can be accurately calculated.

  In the present embodiment, the word feature quantity acquisition unit 104 is realized as the IDF acquisition unit 24a. The word feature quantity acquisition unit 104 may further include an IDF calculation unit 24b. Further, the word feature amount storage unit 105 is realized as the IDF database 35.

  Here, IDF (Inverse Document Frequency) is a global word weight in the entire document set, and is assigned in units of words. IDF is a measure for seeing how many documents a word in the entire document set appears in. In the document set, a greater weight is given to words that appear biased in a small number of documents. Therefore, the IDF (weight) of words that appear uniformly in many different documents is reduced.

  In general, a weight is a value that gives to each word a value that contributes to a feature such as the content and meaning of the document. If a word is characteristic for the document, the weight is assigned. Is a large value. Such word weights (importance) are usually used in fields such as automatic summary generation and collection of search keywords.

  Here, the personal computer 1 of the present invention calculates the word level value using the IDF. More specifically, based on the idea that a word with a smaller IDF (a word that appears more widely in any document) is a less difficult word, the IDF is used to calculate a word level value.

  Details of the IDF calculation method are as described in, for example, Non-Patent Document 5 and the like, and thus the description of the IDF calculation method is omitted.

  The IDF acquisition unit 24 a acquires the IDF from the IDF database 35 for each word analyzed by the morpheme analysis unit 21.

  FIG. 4 is a diagram illustrating a specific example of a table stored in the IDF database 35. As shown in FIG. 4, the IDF database 35 stores the IDF of the word in association with each word. If necessary, the part-of-speech information of the word may be associated and stored. In this case, the IDF acquisition unit 24a can weight the IDF according to the part of speech. For example, the table shown in FIG. 3C can also be used for IDF weighting. The IDF acquisition unit 24a can acquire the IDF for each word with reference to both the table shown in FIG. 4 and the table shown in FIG.

  The IDF acquisition unit 24a refers to the IDF database 35 and acquires the IDF as the word feature amount D4 for each word analyzed by the morpheme analysis unit 21.

  The IDF calculation unit 24b calculates an IDF for each word from the document set D8. The IDF calculation unit 24b stores the calculated IDF in the IDF database 35 in association with the word. In the present embodiment, in particular, the IDF calculation unit 24b regularly acquires the latest set of documents as a corpus, such as a newspaper article database, a textbook for Japanese language education, a sentence uploaded on the Internet, and the like. IDF is calculated. The IDF calculation unit 24b preferably calculates (updates) the IDF periodically. Therefore, it is possible to always obtain the latest IDF. IDF is always integrated and used as an index to automatically calculate high-accuracy word level values that can always reflect buzzwords, traditional terms, Japanese language education, or current affairs that change over time. can do. In addition, in this embodiment, documents close to spoken language such as Internet blogs are also used as a corpus, so that words that are important in terms of learning, such as the scope of test questions, and words that are important for daily use in daily life. Can be eliminated. In this way, the IDF uses Japanese documents that are distributed evenly to Japanese people, such as documents on the Internet and recent newspaper articles, in the corpus, so it has excellent word coverage and fairness. Therefore, the use of IDF for the calculation of the word level value is particularly effective.

  The word level value calculation unit 106 calculates the word level value D5 as described above. In the present embodiment, the word level value D5 is obtained by integrating the total number of words as the character total feature amount D3 determined by the stroke number determination unit 22 and the IDF as the word feature amount D4. For example, the word level value calculation unit 106 may simply obtain the word level value by adding or multiplying the total number of strokes and the IDF, but is not limited thereto. For example, each of the total number of strokes and the IDF may be weighted and added / multiplied, or a weighted average may be obtained from the total number of strokes and the IDF.

  The word difficulty level determination unit 107 determines the word difficulty level D6 as described above. In the present embodiment, the word difficulty level is determined with reference to correspondence information stored in the correspondence information storage unit 108 shown in FIG.

  FIG. 5 is a diagram illustrating a specific example of correspondence information stored in the correspondence information storage unit 108. In the example shown in FIG. 5, the correspondence information includes word level values that can be calculated by the word level value calculation unit 106 divided into six sections (levels A to E and others). A data structure is stored in which character strings representing the difficulty level of words are stored in association with respective values (five levels A to E).

  For example, when the word level value calculation unit 106 calculates the word level value for a certain word as “229”, the word difficulty level determination unit 107 determines the difficulty level of the word based on the correspondence information shown in FIG. Level C: Normal ”is determined.

  Note that the character string “level A” to “level E” or the character strings “simple” to “difficult” may be presented to the user as the word difficulty level D6, or both May be presented. The word difficulty level D6 is displayed on the display unit 13.

  The personal computer 1 further includes a sentence difficulty determination unit (not shown). The sentence difficulty determination unit determines the difficulty of the entire input sentence from the ratio of the number of words of each level value (difficulty) included in the sentence or the document. The degree may be calculated.

[Word level judgment processing flow]
FIG. 6 is a flowchart showing the flow of word level determination processing of the personal computer 1 in this embodiment.

  As shown in FIG. 6, a sentence D <b> 1 that is text data is input to the control unit 10 of the personal computer 1 through the input unit 12. When a sentence is input (YES in S101), the morphological analyzer 21 performs a morphological analysis of the input sentence and divides the sentence into words (S102). The morphological analysis unit 21 outputs the word information D2 of the divided words to the stroke number determination unit 22 and the IDF acquisition unit 24a.

  The stroke number determination unit 22 and the IDF acquisition unit 24a output a character total feature amount D3 (word total stroke number) and a word feature amount D4 (IDF) for each of the output words. More specifically, when the stroke number determination unit 22 acquires the word output from the morpheme analysis unit 21 (S103), the stroke number determination unit 22 refers to the stroke number database 23 (FIG. 3A) and configures each acquired word. The number of strokes is acquired for the character (S104). Then, referring to the stroke number database 23 (FIGS. 3B and 3C), the acquired stroke number is weighted according to the character type and the part of speech of the acquired word, and the total stroke number of the word is calculated. (S105). The stroke number determination unit 22 outputs the calculated total number of strokes of the word to the word level value calculation unit 106. On the other hand, when the IDF acquisition unit 24a acquires the word output from the morphological analysis unit 21 (S106), the IDF acquisition unit 24a refers to the IDF database 35 (FIG. 4) and acquires the IDF associated with the acquired word (S107). . The IDF acquisition unit 24a outputs the acquired IDF to the word level value calculation unit 106 as the word feature amount D4.

  Subsequently, the word level value calculation unit 106 acquires the total number of words from the stroke number determination unit 22 and the IDF from the IDF acquisition unit 24a (S108). Then, the word level value calculation unit 106 integrates the total number of word strokes and the IDF in the word level value calculation process such as (weighting) addition, multiplication, (weighted) average as described above, and calculates the word level value. (S109). The word level value calculation unit 106 outputs the calculated word level value D5 to the word difficulty level determination unit 107. Here, the word level value calculation unit 106 may output the word level value D5 to the display unit 13.

  Next, the word difficulty level determination unit 107 refers to the correspondence information storage unit 108 illustrated in FIG. 5, and levels (A to E) and / or difficulty levels corresponding to the word level values calculated by the word level value calculation unit 106. Is determined (S110). As described above, when the difficulty level of one word is determined, a series of processes from S103 to S110 is performed until the above-described word level value / difficulty level is obtained for all the words in the sentence input in S101. Is repeated (NO in S111).

  Finally, when the control unit 10 obtains the word level value / difficulty level for all the words in the sentence (YES in S111), the control unit 10 displays the word level value / difficulty level on the display unit 13 (S112).

  In the example illustrated in FIG. 6, the series of processes of S103 to S105 and the series of processes of S106 to S107 are repeatedly executed every time the word level value / difficulty level is calculated for one word. The order of is not intended to limit the present invention. For example, for each word divided by the morphological analysis unit 21, the above-described series of processing is repeated, and the total number of strokes (IDF) is summed up as a word level value. A processing order of outputting to the calculation unit 106 may be used. The series of processes may be executed in parallel or sequentially in series (in this case, the processing order may be changed).

  According to the above configuration and method, objective and accurate word level determination can be realized by integrating the character-by-word feature amount and the word-unit feature amount. Specifically, the number of strokes for each character included in the word (and the total number of strokes of all the characters in the word) and the IDF can be integrated to calculate a word level value. It is possible to achieve objective and accurate word level determination for determining the difficulty level of Japanese teaching materials).

  Furthermore, according to the above-described configuration and method, the word level is determined by taking in characteristics (number of strokes of characters, complexity of characters) obtained in units of characters, so that the meaning content, word feeling, familiarity of words An objective, fair and accurate word level determination method is realized without being influenced by subjective measures such as.

  In addition, according to the above configuration and method, since IDF is adopted as a feature obtained in units of words, the word level can be determined widely and fairly for every word. In addition, the IDF is always calculated using the latest document such as a recent newspaper article or a digital document on the Internet as a corpus. Therefore, the word level can be accurately determined for every word in response to daily vocabulary changes.

  The present invention makes it possible to support teachers (users) who provide Japanese language teaching materials to learners in, for example, Japanese language education (particularly Japanese language education for foreigners). In the past, teachers have rewritten and used materials that serve as Japanese teaching materials (for example, newspaper articles) according to the learner's skills. However, this word replacement and sentence proofing are laborious tasks for teachers. This is because, after appropriately determining the level of the word in the material, it is necessary to search for an appropriate (more simple or difficult) word according to the skill of the learner.

  According to the present invention, the teacher can obtain an objective word level value (and corresponding word difficulty level) simply by inputting a sentence to the personal computer 1, and the word level value is used as a scale. It becomes possible to easily and objectively determine the difficulty level of every word. Also, in the Japanese language education for foreigners who have difficulty in learning kanji, the word level value can be determined with emphasis on the number of strokes of kanji, so it is particularly effective.

  The present invention is not limited to Japanese language education, but can be used to keep the word level used in a sentence constant.

<< Embodiment 2 >>
The present invention is an invention in which an objective word level value is obtained and word level determination is performed objectively. Further, the objective determination result is determined according to the ability and characteristics of the learner (user), It is converted into a subjective difficulty level, which is presented to each learner, and the learner can be supported. In addition, it provides education support for teachers (users) who provide learning materials to the learners.

  Below, the case where the function of the personal computer 1 shown in FIG. 2 is implement | achieved with the information processing system which consists of a some apparatus is demonstrated. That is, the case where the word level determination system 100 of the present invention is applied to a client-server type network system will be described as an example.

[Configuration of each device of the word level judgment system]
In this embodiment, the word level determination system 100 includes a terminal device used by a learner, a word level determination device that executes word level determination of a sentence in response to a request from the terminal device, and returns the result to the terminal device. It is the composition which includes. Both devices have a communication function and can be connected to each other via a communication network such as the Internet. In addition, about each member in this embodiment, the same code | symbol is attached | subjected to the member which has the same function as the member shown in the personal computer 1 of FIG. 2, and those are shown substantially the same. Therefore, description of these members will not be repeated except for different portions.

  First, the configuration of the terminal device on the client side will be described. The terminal device includes an input unit 12 and a display unit 13 illustrated in FIG. 2 in addition to a communication unit, a storage unit, and a control unit.

  In the present embodiment, the display unit 13 displays the GUI screen of the word level determination application acquired by the communication unit from the word level determination device via the Internet. Specific examples of the GUI screen displayed by the display unit 13 are shown in FIGS. The GUI screen displayed on the terminal device is not limited to the above specific example.

  FIG. 10 is a diagram illustrating an example of a GUI screen for inputting text. In the present embodiment, the learner can input a sentence whose level is to be determined in the sentence input area 50 by operating the input unit 12. When the input of the text is completed, the learner continues to operate the input unit 12 to input the learner ID assigned to himself / herself into the ID input area 51 and clicks the determination button 52. At this time, the communication unit of the terminal device transmits a request including a sentence (sentence D1) received on the GUI screen and a learner ID as learner information D7 to the word level determination apparatus. The learner ID is for the word level determination device to uniquely identify a client (terminal device / learner) that receives a service provided by the word level determination device. As the learner ID, any information may be used as long as the above purpose can be achieved. The word level determination device may be uniquely assigned in advance to each terminal device.

  FIG. 11 is a diagram illustrating an example of a GUI screen for presenting the word level determination result to the user. The determination result generated by the word level determination device in response to the request is received via the communication unit of the terminal device and displayed on the display unit 13. The determination result is generated as display data for displaying the word level value D5, the word difficulty level D6, and / or the text difficulty level on the display unit 13 of the terminal device. The word level determination device generates a determination result in a format such as HTML (HyperText Markup Language) or XML (eXtensible Markup Language), and supplies the determination result to the terminal device.

  In the example shown in FIG. 11, only the difficulty level determination result obtained by converting the word level value D5 into the word difficulty level D6 is displayed. For example, the determination result GUI screen includes an original text display area 60 that displays an input sentence by color (typeface) according to the difficulty level of the word, a word list 61 that displays a list of words for each difficulty level, and a difficulty level A count table 62 that displays the number of word appearances is included. By displaying these, the difficulty determination result for each word can be presented to the learner from various angles. Further, a text difficulty level display area 63 for displaying the text difficulty level may be provided to present the difficulty level of the entire text to the learner.

  In this embodiment, the GUI screen of this determination result is generated for each learner (ID: A004). Therefore, a message display area 64 for displaying a message addressed to the learner may be provided. In addition, when the terminal device can use the sentence proofreading function for changing the difficulty level without changing the meaning of the sentence, the sentence proofreading button 65 may be provided. For example, when a “convert to easier text” button is clicked, the original input text and a text simplification request are transmitted to the execution unit of the above function.

  According to the above configuration, the learner can use the word level determination application of the word level determination device using the terminal device, and obtain the determination result of the word level value and the word difficulty of the word of the sentence input by the learner. Can be obtained.

  Next, the configuration of the server-side word level determination device for supplying the determination result for each learner will be described. The word level determination apparatus includes each unit of the control unit 10 illustrated in FIG. 2, a communication unit 14, and a storage unit 11. The various data stored in the storage unit 11 may be stored in a storage device built in the word level determination device, or may be stored in an external storage device that can be connected via the communication unit 14. It may be stored.

  In the present embodiment, unlike the personal computer 1, the communication unit 14 has a function of receiving the request transmitted by the terminal device in addition to the function of acquiring the document set D <b> 8. That is, the sentence D <b> 1 and the learner information D <b> 7 included in the request are input to the control unit 10 via the communication unit 14. The input sentence D1 is supplied to the morpheme analysis unit 21, and the learner information D7 is supplied to the word difficulty level determination unit 107.

  The input sentence is subjected to the same processing as described above by each unit of the control unit 10, and the word level value calculation unit 106 calculates a word level value D5 for each word of the sentence (see FIG. 6 S101 to S109). The word level value D5 is used by the word difficulty level determination unit 107 to determine the word difficulty level D6.

  As described above, the word difficulty level D6 differs from a quantitative numerical value such as the word level value D5, and may not necessarily be an expression suitable for all learners. In order to eliminate this inconvenience, in the present embodiment, the word difficulty level determination unit 107 refers to the learner information storage unit 38 and the correspondence information storage unit 108 and based on the supplied learner information D7, The Japanese skill can be specified, and the correspondence between the word level value D5 and the word difficulty level D6 can be changed according to the learner.

  FIG. 7 is a diagram illustrating another specific example of correspondence information stored in the correspondence information storage unit 108. In the example shown in FIG. 7, unlike the example shown in FIG. 5, a plurality of pieces of correspondence information R70 to R75 are stored for each learner skill (learning stage). In each piece of correspondence information shown in FIG. 7, the correspondence between the word level value and the character string (levels A to E) of the word difficulty varies depending on the learning stage. Alternatively, the correspondence information table as shown in FIG. 5 may have a data structure in which a plurality of correspondence information tables are stored in advance for each learning stage. In this case, since the correspondence information storage unit 108 has six learning stages in the example of FIG. 7, it has a structure including six correspondence information tables as shown in FIG. In the example shown in FIG. 7, six pieces of correspondence information are indicated as R70 to R75 (one line corresponds to one table in FIG. 5).

  For example, assuming that the word level value of a certain word is calculated as “250”, the word difficulty level determination unit 107 determines that the word difficulty level is “ordinary” for a learner having a skill of “about junior high school students”. It can be determined as “slightly simple” for a learner having a skill of “about a high school student”.

  The level of skill of the learner who is the sender of the request can be identified by referring to the learner information storage unit 38 by the word difficulty level determination unit 107.

  FIG. 8 is a diagram illustrating a specific example of a learner information table stored in the learner information storage unit 38. As shown in FIG. 8, learner skills are stored in association with learner IDs uniquely assigned to learners.

  The word difficulty level determination unit 107 identifies the skill of the learner who is the sender of the request based on the learner ID as the learner information D7 included in the request by referring to the learner information table. It becomes possible to do. Then, by referring to the word level value-difficulty correspondence table corresponding to the identified skill, it is possible to determine the word difficulty with an expression that is optimal for the sender (learner) of the request.

[Word difficulty determination processing flow]
FIG. 9 is a flowchart showing the flow of word difficulty level determination processing of the word level determination apparatus according to this embodiment.

  Similar to S109 of FIG. 6, when the word level value calculation unit 106 calculates a word level value of a certain word, the word difficulty level determination unit 107 acquires the word level value from the word level value calculation unit 106 (S201). . And if there is the learner information D7 regarding the input learner other than the sentence D1 containing the said word, it will be acquired. For example, a learner ID is acquired as learner information D7 (S202). Based on the learner ID obtained in S202, learner skills (learning stage) are obtained (S203). For example, as shown in FIG. 10, when “learner ID: A004” is input together with the request, the learner skill (learning stage) of the sender of the request is specified as “about the middle school grade of elementary school”. The learner information D7 may be information indicating the learning stage itself.

  When the learner skill is designated based on the learner information in S203, the word difficulty level determination unit 107 stores the correspondence information of the designated “about the elementary school middle grade” stored in the correspondence information storage unit 108 ( In the example of FIG. 7, the word difficulty level is determined by referring to the R71 line) and identifying the classification (character string) corresponding to the word level value acquired in S201 in the correspondence information (S204). ).

  According to the above-described configuration and method, the objective determination result obtained by the word level value calculation unit 106 is converted into a subjective difficulty level according to the ability and characteristics of the learner, and is converted to each learner. It is possible to support the learner. When the present invention is applied in a Japanese language education field, it is possible to provide an appropriate level of text as a teaching material in a timely manner in consideration of the learner's skill and learning stage. When the learner information D7 is not input or the learning stage cannot be specified in S203 from the input learner information D7, default correspondence information (for example, default correspondence information is shown in FIG. 7). The word difficulty level may be determined on the basis of R73).

[Modification]
In each of the embodiments described above, the word level value calculation unit 106 determines the word level value D5 by integrating the total number of words as the character total feature amount D3 and the IDF as the word feature amount D4 as described above. . At this time, the word level value calculation unit 106 may perform addition / multiplication by weighting each value, or may obtain a weighted average of both values. Here, which value is weighted and how it is weighted may be made variable according to the characteristics of the learner.

  For example, in a scene where the level of a sentence / word is judged for a learner from a kanji-speaking country that uses kanji as a character, such as China, Taiwan, Korea, etc., the complexity and difficulty of kanji is difficult. However, it is not necessarily directly connected. Therefore, in the above scene, it is conceivable to calculate the word level value after weighting the IDF rather than the total number of words. On the other hand, if a learner is from a non-kanji-speaking country such as Europe or America, the complexity of kanji is likely to be a factor that directly increases the difficulty of writing. Thus, in such a scene, it is conceivable to weight the total number of strokes rather than the IDF. In this way, the word level value calculation unit 106 can change whether to calculate the word level value with emphasis on the total number of strokes of the word or the IDF according to the characteristics of the learner.

  In this modification, the learner information storage unit 38 (FIG. 2) is configured as shown in FIG. In other words, country-of-origin information indicating whether the learner's home country is within the kanji range (non-kanji range) is stored in association with the learner ID.

  The word level value calculation unit 106 refers to the learner information table shown in FIG. 8, so that the learner who is the sentence input person who has requested the determination based on the learner ID input to the own device It is possible to determine whether or not you are from home. If the text input person is from a kanji-speaking area, the word level value is calculated by using a calculation method that emphasizes the IDF, and if the text input person is from a non-kanji-speaking area, the calculation method is weighted by the total number of strokes of the word. Execute.

  According to the above configuration and method, the word level value of each word can be calculated more accurately and more appropriately according to the ability and characteristics of the learner.

[Modification 2]
In each of the above-described embodiments, the word level determination system 100 (the personal computer 1, the server-side word level determination device) may be configured to determine the difficulty level of the entire sentence based on the word difficulty level of each word. Good. Here, the word level determination system 100 may further include a sentence proofreading unit (not shown) that changes the difficulty level without changing the meaning of the sentence.

  For example, the sentence proofreading unit triggers that the sentence proofreading button 65 shown in FIG. 11 is clicked, and makes all (or a part of) words in the sentence before proofreading difficult for the word difficulty level by one level (or The new sentence is generated by replacing it with another word having a different difficulty level (and having the same meaning).

  As a result, the teacher can convert learning materials (newspaper articles, novels, etc.) according to the level of the learner into teaching materials. In addition, it is possible to easily prepare the same teaching material with different difficulty levels for learners of various learning skills. This function is also effective for learners. For example, if it is difficult to read due to the use of special expressions in newspaper articles, novels, public bulletins of government offices (providing information for foreigners in the event of a disaster, etc.), convert those sentences into plain sentences Can be converted.

[Modification 3]
It is also possible to implement the word level determination system 100 of the present invention in conjunction with a dictionary service. Specifically, the word level determination system 100 refers to the dictionary service usage history stored for each learner, identifies the learner's learning stage, and an appropriate word difficulty level according to the identified learning stage. This result is presented to the learner. The usage history is, for example, a log that records which words the learner is searching for using the service.

  The word level determination system 100 collects information on what kind of word the learner is examining from the use history. The word whose meaning has been examined is regarded as an unknown word to the learner, and the word level determination system 100 determines the learner's learning depending on what kind of word the learner does not know. The stage can be specified.

  In each of the above-described embodiments, a word is input to the word level determination system 100 (PC 1) of the present invention as a trigger, and the word level determination system 100 (PC 1) Although described as performing word level determination, the word level determination system and the word level determination apparatus of the present invention are not limited to this. For example, the word level determination system 100 (the personal computer 1) is configured to receive one Japanese word as an input instead of the sentence D1 and output a word level value D5 (or word difficulty D6) for the word. May be. In this case, even if the word extraction unit 101 is not provided, the input word is supplied to the character feature amount acquisition unit 102 and the word feature amount acquisition unit 104.

  The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  Finally, each block of the word level determination system 100 (personal computer 1, terminal device, word level determination device), in particular, a word extraction unit 101 (morpheme analysis unit 21), a character feature amount acquisition unit 102 (stroke number determination unit 22), The word feature amount acquisition unit 104 (IDF acquisition unit 24a / IDF calculation unit 24b), the word level value calculation unit 106, and the word difficulty level determination unit 107 may be configured by hardware logic, or the CPU may be configured as follows. And may be realized by software.

  That is, the word level determination system 100 includes a central processing unit (CPU) that executes instructions of a control program that realizes each function, a read only memory (ROM) that stores the program, and a random access memory (RAM) that expands the program. ), A storage device (recording medium) such as a memory for storing the program and various data. An object of the present invention is a recording medium in which a program code (execution format program, intermediate code program, source program) of a control program of the word level determination system 100, which is software that realizes the functions described above, is recorded so as to be readable by a computer This can also be achieved by supplying the above to the word level determination system 100 and reading and executing the program code recorded on the recording medium by the computer (or CPU or MPU).

  Examples of the recording medium include a tape system such as a magnetic tape and a cassette tape, a magnetic disk such as a floppy (registered trademark) disk / hard disk, and an optical disk such as a CD-ROM / MO / MD / DVD / CD-R. Card system such as IC card, IC card (including memory card) / optical card, or semiconductor memory system such as mask ROM / EPROM / EEPROM / flash ROM.

  Further, the word level determination system 100 may be configured to be connectable to a communication network, and the program code may be supplied via the communication network. The communication network is not particularly limited. For example, the Internet, intranet, extranet, LAN, ISDN, VAN, CATV communication network, virtual private network, telephone line network, mobile communication network, satellite communication. A net or the like is available. Also, the transmission medium constituting the communication network is not particularly limited. For example, even in the case of wired such as IEEE 1394, USB, power line carrier, cable TV line, telephone line, ADSL line, etc., infrared rays such as IrDA and remote control, Bluetooth ( (Registered trademark), 802.11 wireless, HDR, mobile phone network, satellite line, terrestrial digital network, and the like can also be used. The present invention can also be realized in the form of a computer data signal embedded in a carrier wave in which the program code is embodied by electronic transmission.

  According to the word level determination system of the present invention, the level of a word included in a Japanese sentence can be easily and objectively determined with higher accuracy. Therefore, the present invention can be applied in various fields as long as the Japanese language is used in consideration of the difficulty level. For example, in the field of Japanese language education, particularly in the field of Japanese language education for foreigners, it is necessary to provide appropriate teaching materials to learners of various learning stages and learning skills, so that the present invention is particularly preferably used. be able to.

1 PC (word level determination system / word level determination device / information processing device)
DESCRIPTION OF SYMBOLS 10 Control part 11 Memory | storage part 12 Input part 13 Display part 14 Communication part 21 Morphological analysis part (word extraction means)
22 Stroke number determination unit (character feature amount acquisition means)
23 Stroke database 24a IDF acquisition unit (word feature acquisition means)
24b IDF calculation part (IDF calculation means)
35 IDF database 38 Learner information storage unit 50 Text input area 51 ID input area 52 Determination button 60 Original text display area 61 Word list 62 Count table 63 Text difficulty display area 64 Message display area 65 Sentence correction button 100 Word level determination system ( Word level determination device / information processing device)
101 word extraction unit (word extraction means)
102 Character feature acquisition unit (character feature acquisition means)
103 character feature amount storage unit 104 word feature amount acquisition unit (word feature amount acquisition means)
105 word feature amount storage unit 106 word level value calculation unit (word level value calculation means)
107 Word difficulty determination unit (word difficulty acquisition means)
108 Correspondence information storage unit D1 Sentence D2 Word information (word)
D3 Character total feature (total number of strokes)
D4 Word feature amount D5 Word level value D6 Word difficulty D7 Learner information D8 Document set

Claims (11)

Character feature amount acquisition means for acquiring a character feature amount for each character included in a word input to the device from a character feature amount storage unit that stores a character feature amount indicating a character feature for each character;
Word feature amount acquisition means for acquiring a word feature amount of the input word from a word feature amount storage unit that stores a word feature amount indicating a word feature for each word;
A character total feature amount obtained by combining the character feature amounts acquired by the character feature amount acquisition unit of each character included in the input word, and a word feature acquired by the word feature amount acquisition unit And a word level value calculating means for calculating a word level value indicating the difficulty of the word in the inputted word from the quantity. It further comprises word extraction means for extracting words from Japanese sentences input to its own device,
The character feature amount acquisition unit performs a process of acquiring a character feature amount for each character included in the word for each word extracted by the word extraction unit,
The word feature amount acquisition unit performs a process of acquiring a word feature amount of a word for each word extracted by the word extraction unit,
The word level value calculating means includes, for each word extracted by the word extracting means, the character comprehensive feature value obtained by combining character feature values, and the word acquired for each extracted word. The word level determination apparatus according to claim 1, wherein the word level value is calculated for each extracted word from a feature amount. A correspondence information storage unit that stores correspondence information in which the word level value is associated with a word difficulty level that is a character string representing the difficulty of a word to be presented to the user;
The word difficulty level acquisition means for acquiring the word difficulty level associated with the word level value calculated by the word level value calculation means from the correspondence information storage unit. 2. The word level determination device according to 2. The character feature amount storage unit stores the number of strokes of characters as the character feature amount,
The word level determination device according to any one of claims 1 to 3, wherein the character feature amount acquisition unit acquires the number of strokes of characters from the character feature amount storage unit. The word feature storage unit stores a word IDF (Inverse Document Frequency) as a word feature,
5. The word level determination apparatus according to claim 1, wherein the word feature amount acquisition unit acquires an IDF of a word from the word feature amount storage unit. The character feature amount storage unit stores the number of strokes of characters as the character feature amount,
The character feature amount acquisition unit calculates the total number of strokes of words by totaling the number of strokes of each character acquired from the character feature amount storage unit for each word,
The word feature storage unit stores a word IDF as a word feature,
The word feature amount acquisition unit acquires an IDF (Inverse Document Frequency) of a word from the word feature amount storage unit,
The word level value calculation means calculates a word level value by performing at least one of an addition process, a multiplication process, and a weighted average process on the total number of words and the IDF. Item 6. The word level determination device according to any one of Items 1 to 5. The correspondence information storage unit further stores the correspondence information for each learning stage of the learner,
The word difficulty level acquisition means is:
When the learning stage is specified by the learner information related to the learner input to the own device, the word difficulty level associated with the calculated word level value in the correspondence information corresponding to the specified learning stage is set. The word level determination device according to claim 3, wherein the word level determination device is acquired from the correspondence information storage unit. The word level determination device according to any one of claims 1 to 7,
A terminal device that communicates with the word level determination device,
The terminal device is
Requesting the word level determination device for a word level value of a word input by a user of the terminal device;
The word level determination device
A word level determination system, wherein a word level value calculated for a word received from the terminal device is transmitted to the terminal device as a response to the request. A character feature amount acquisition step for acquiring a character feature amount for each character included in a word input to the word level determination device from a character feature amount storage unit that stores a character feature amount indicating a character feature for each character;
A word feature amount acquisition step of acquiring a word feature amount of the input word from a word feature amount storage unit storing a word feature amount indicating a word feature for each word;
The character total feature amount obtained by combining the character feature amount acquired in the character feature amount acquisition step of each character included in the input word, and acquired in the word feature amount acquisition step A word level determination method comprising: a word level value calculating step for calculating a word level value indicating difficulty of a word in the input word from the word feature amount.   The control program for functioning a computer as each means of the word level determination apparatus of any one of Claim 1-7.   The computer-readable recording medium which recorded the control program of Claim 10.

Images