JP2011018155A - Method, device and program for creating infant vocabulary development database - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technology for constructing a highly reliable infant vocabulary development database based on data collected in a user participation type system.SOLUTION: An input means 11 accepts word information input through a user terminal, and an obtaining age of the month generation means 12 calculates the obtaining age of the month of the accepted word information, and a meaning category application means 13 applies a meaning category to the accepted word information, and an fraud detection means 14 applies the obtaining age of the month or the meaning category to a plurality of preliminarily prepared definition references to exclude any illicit word information, and a mean obtaining age of the month calculation means 15 calculates the mean obtaining age of the month of the word information recognized to be valid without being excluded, and a reliability determination means 16 determines the reliability of the word information from the mean obtaining age of the month, and a generation means 17 generates an infant vocabulary development database based on the word information whose reliability is recognized.

Description

  The present invention relates to a technology for creating an infant vocabulary development database using data posted from the Web, and more particularly to a technology for ensuring the reliability of posted data and realizing the high-quality database.

  Currently, there are various dictionaries on the web, ranging from electronic book-version dictionaries such as English-Japanese dictionaries and national language dictionaries to browse on the website to user-participation dictionaries represented by “wikipedia”. There is a database browsing service.

  The biggest advantage of publishing dictionaries and databases on the web compared to the book version is that many users can easily add and change information, so that new information can be added in a timely manner. is there. In other words, although there are possible disadvantages related to the accuracy and reliability of posted information, it is possible to efficiently and quickly build a kind of dictionary and database that did not exist so far by taking advantage of these web characteristics. It is.

  Under such circumstances, a database on infant vocabulary development called “Children's Dictionary” has been released on the web, and a dictionary unique to the world is being constructed. This is the data recorded by the user on the web diary tool, and the database created by organizing and processing the recorded information on the web. It is something that can be done. According to this browsing service, when a certain word is acquired (for example, the word “mom”: acquisition average 15.4 months old), the meaning of an infant word (infant word “sha” → meaning: “train”), etc. Can be easily searched and viewed, and is a childcare information providing service useful for parents who have children aged 0 to 3 (Non-patent Document 1).

Tetsuo Kobayashi, Masaaki Nagata “Attempts to collect initial vocabulary development data through the web and its applications” The 8th Annual Meeting of the Japanese Society of Baby Sciences Abstracts held April 12-13, 2008 pp. 73 Kim, M.M. McGregor, K .; K. , & Thompson, C.I. K. (2000) "Early lexical development in England- and Korean-speaking children: language-general and language-specific patterns, Journal of Children 2 25 Tamiko Kotsuki, Toru Watanuki "Standard Research on Vocabulary Development in Japanese Children: From the Japanese MacArthur Infant Language Development Questionnaire" Bulletin of Kyoto International Social Welfare Center "Development and Rehabilitation Research" 2008.11 24th pp. 3-42

  The collection of data related to infant vocabulary development is usually a problem because researchers in the fields of developmental psychology and psycholinguistics observe children's utterance behaviors and directly interview their mothers. Almost never.

  However, when an infant vocabulary development database is constructed based on user-submitted data, the reliability of data due to posting of illegal data becomes a problem. In particular, in the case of the “Children's Word Dictionary”, each individual's information about when a certain word was spoken is an important component, so many users can post data with inaccuracy of data, If there are no children, posting false data maliciously may reduce the accuracy of the aggregated data.

  Further, when the database is opened to the public as a browsing service, a responsibility for presenting more accurate information is required. Therefore, it is necessary to avoid these problems with some technique, and it is important to ensure the accuracy and reliability of the database.

  The present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide a technique for constructing a reliable infant vocabulary development database based on data collected by a user participation type method.

  Therefore, the present invention implements a process for detecting and reliability of fraudulent data that makes use of infant vocabulary development characteristics when creating an infant vocabulary development database based on data collected by a user participation type method. Provide quality database creation technology.

  One aspect of the present invention is a method for creating an infant vocabulary development database using word information posted on the web through a user's terminal, wherein the input means inputs the word information to the terminal. An input receiving step for displaying the interface and receiving the word information input by the user through the terminal, and the fraud detection means, a plurality of definition criteria prepared in advance for the acquired age or meaning category of the word information received in the input receiving step And the fraud detection step for detecting fraudulent word information and eliminating the fraud information, and the average acquisition age calculation means determine the average acquisition age of the word information that is recognized as valid without being excluded in the fraud detection step. The average acquired age calculated by the average acquired age calculated in the average acquired age calculated in the average acquired age A reliability determination step for judging the reliability of the word information recognized as valid based on the above, and generation means for generating the infant vocabulary development database based on the word information whose reliability is recognized in the reliability determination step Steps.

  Another aspect of the present invention is an apparatus for creating an infant vocabulary development database using word information posted on the web through a user terminal, and displaying an interface for inputting word information on the terminal Incorrect word information is detected by applying the input means for receiving word information input by the user through the terminal and the acquired age or semantic category of the word information received by the input means to a plurality of definition criteria prepared in advance. Calculated by the fraud detection means for eliminating the fraud information, the average acquisition age calculation means for calculating the average acquisition age of the word information that is recognized as valid without being excluded by the fraud detection means, and the average age calculation means Reliability determination means for judging the reliability of word information recognized as effective based on average acquired age, and reliability is recognized by the reliability determination means. Based on the word information comprises generation means for generating an infant vocabulary development database, a.

  In addition, this invention is good also as an aspect of the program for functioning a computer as each means of the said apparatus. This program may be provided in a form stored in a recording medium.

  According to the present invention, a highly reliable infant vocabulary development database is constructed based on data collected through a user participation method.

The basic composition figure of the infant vocabulary development database creation device concerning the embodiment of the present invention. FIG. The block diagram of the infant vocabulary development database creation apparatus which concerns on an Example. The image figure of the word input example of the same user interface. The figure which shows an example of a synonym category classification table. The figure which shows an example of a synonym category definition dictionary. The chart figure which shows the process of the same unauthorized data detection part. The figure which shows an example of the same crossing 50% arrival age dictionary. The figure which shows an example of the same infant vocabulary development database. The image figure of the search page which the user interface browsing part provides. The image figure of the search result page which the user interface browsing part provides.

≪Basic structure≫
FIG. 1 shows a basic configuration of an apparatus for creating an infant vocabulary development database according to an embodiment of the present invention. This creation device 1 is connected to a user terminal (not shown) via the Internet.

  Here, the creation device 1 performs a processing process of fraudulent data detection and data reliability determination on post data relating to infant vocabulary development from the user terminal, and creates a high-quality infant vocabulary development database. Specifically, the creation device 1 is configured by a computer and includes hardware resources of a normal computer, such as a CPU, a memory (RAM), a hard disk drive device, and a communication device.

  As a result of the cooperation between the hardware resource and the software resource (OS, application), the creation device 1 provides a user interface that can input a word to the user terminal according to the characteristics of the infant vocabulary. Means 11, word acquisition age generation means 12 for calculating the acquisition age of each word data input through the input means 11, infant vocabulary meaning category assignment means 13 for assigning a semantic category to each input word, both means 12. By using the processing result of 12.13, a plurality of definition criteria are applied to the input word group to detect illegal data, and the invalid data detecting means 14 detects the input word from the valid data set that is not excluded by the detecting means 14. Average acquired age generation means 15 for generating an average acquired age, and average acquired age calculated by the generation means 15 A data reliability determination unit 16 for determining the reliability of the data set using the data set, and an entry word item is generated based on the highly reliable data set determined by the determination unit 16 to form an infant vocabulary development database. Database item generation means 17 for performing the search, and user interface browsing means 18 for providing a user interface capable of browsing and searching the database generated by the generation means 17 on the user terminal.

  Here, the generation unit 17 generates the database on the hard disk drive device. The input unit 11 and the browsing unit 18 provide each user interface to the user terminal through the communication device. The input unit 11 provides a user interface for a web diary tool.

  FIG. 2 shows a database creation process of the creation device 1. Here, first, a word remembered by a child is input to the interface of the web diary tool displayed on the user terminal through the input unit 11, and the web diary tool word that the input unit 11 accepts via the Internet. A reception step (S01) is performed.

  Next, for each input word received in S01, a word acquisition age generation step (S02) in which the generation means 12 calculates an acquisition month age is performed, and then the meaning category of each input word is given by the assignment means 13. An infant vocabulary meaning category assignment step (S03) to be provided is performed.

  Subsequently, for the input word group received in S01, there is an illegal data detection step (S04) in which the detection means 14 detects and eliminates illegal data based on a plurality of reference indices devised from scientific characteristics of infant vocabulary. To be implemented. The detection step is performed by using the word acquisition age calculated in S02 and the semantic category assigned in S03.

  Then, for each valid input word that has not been excluded in S04, the generation means 15 performs an average acquired age generation step (S05) in which an average acquired age is generated. A data reliability determination step in which the determination means 16 evaluates / determines the reliability of each input word by referring to a dictionary in which the acquired age for each word is previously tabulated / listed based on the generated average acquired age. S06) is performed.

  A database item generating step (S07) is performed in which the generating means 17 generates a final database item based on each input word having high reliability in the stage of S06. Here, the generated items are converted into a database and used as an infant vocabulary development database, and a user interface that can be browsed and searched is displayed on the user terminal through the browsing means 18.

<Example>
FIG. 3 shows a configuration of an embodiment of the creation device 1, which includes a user interface input unit 31, a word acquisition age generation unit 32, an infant vocabulary meaning category assignment unit 33, an intermediate data holding unit 34, an unauthorized data detection unit 35, an average It has an acquired age generation unit 36, a data reliability determination unit 37, a database item generation unit 38, and a user interface browsing unit 39. Here, the parts 31 to 33.35 to 39 correspond to the means 11 to 18, respectively. Hereinafter, each part 31-39 is demonstrated individually concretely.

(1) User interface input unit 31
The user interface input unit 31 displays the user interface of a personal web diary tool on the browser of the user terminal via the Internet. According to this interface, as shown in FIG. 4, what word (eg, doggie) and what meaning (eg: dog) is spoken for each date and time (when “example: 200 x year x month x day”). Input field for recording R is displayed. The word information input here, that is, input data in the input field Q is referred to as input word, and input data in the input field R is referred to as input meaning.

  When the input unit 31 accepts each input data, post data necessary for database creation is collected. Each data collected here is transferred to each section 32.33. Basic information such as personal identification ID, gender, date of birth, birth order, and residential area shall be entered separately before using the web diary tool. The input basic information is transferred to the generation unit 32 and the data holding unit 34.

(2) Word acquisition age generation unit 32
The word acquisition age generation unit 32 is configured to input the input fields Q. From the difference between the date and time of data recording in R and the date of birth of the user's child input as the basic information, the age at which the input word was acquired, that is, the number of months after birth was determined. For example, in the case of data recording date “2008.10.21” and date of birth “2007.8.5”, the acquired age is calculated as “14.5 months”. The acquired age calculated here is transferred to the intermediate data holding unit 34 as a pair with the input word.

(3) Infant vocabulary meaning category assigning unit 33
The infant vocabulary meaning category assigning unit 33 assigns a meaning category (eg, animal) to the input meaning (eg, dog) with reference to the meaning category classification table of FIG. At this time, an input word (eg, doggie) is collated with the semantic category definition dictionary 310 in FIG. 3 to determine an input semantic category. As shown in FIG. 6, in this definition dictionary 310, semantic categories are defined in advance for words that may be posted as infant vocabulary.

  For example, according to the definition dictionary 310 of FIG. 6, the input word (eg, doggie) corresponds to the category name “animal” of the category ID “25”. As a result, as shown in the category classification table of FIG. 5, the meaning category of “25. Animal” in “2. The semantic category assigned here is transferred to the intermediate data holding unit 34 in pairs with the input meaning. It is assumed that the semantic category classification table of FIG. 5 and the definition dictionary 310 of FIG. 6 are stored in the hard disk drive device, respectively.

(4) Intermediate data holding unit 34
The intermediate data holding unit 34 organizes and holds the transfer data from the units 31 to 33. Here, for input words (example: doggie), input meaning (example: dog), acquired age (example: 14.5 months), semantic category (example: 25. animal), and personal identification ID (example: F09-) 3-456) shall be a data set.

  Specifically, the intermediate data holding unit 34 sequentially holds the data set for each user through the memory (RAM) or the hard disk drive device.

(5) Unauthorized data detection unit 35
The fraudulent data detection unit 35 includes four indexes defined in the program, that is, the meaningful word age verification definition 311 in FIG. On the basis of the above, illegal data is determined and detected from the data set held by the intermediate data holding unit 34. Hereinafter, the processing process of the unauthorized data detection unit 35 will be described with reference to FIG.

  S11: First, the fraudulent data detection unit 35 selects the words from the smallest acquired age to the 50th word among the input words of each user included in the data set. The word selected here is referred to as “early appearance vocabulary 50 words”. For the 50 early appearance vocabulary words, illegal data is detected based on the respective indices in S12 and thereafter.

  S12: Incorrect data detection based on the meaningful word age collation for the 50 words of early appearance vocabulary selected in S11, that is, the meaningful word age collation definition 311 is performed. Here, if the early appearance vocabulary 50 word includes a meaningful word whose acquired age is 8 months or older, this is regarded as illegal data. The input meaning in the data set is used for checking whether this is a meaningful word.

  As described above, the reason why the data is incorrect when the acquired age is 8 months or older is that the infant's 8 months of age or earlier is the brain function that controls the articulatory organ and its control, and the cognitively related brain function of the speech representation and the target object. This is because there is a natural scientific finding that is not mature enough. For example, even if a 3-month-old child utters a meaningful word such as “give me”, “jump”, or “fallen”, it is presumed that it is actually impossible even if it is entered in the input field Q.

  S13. S14: Next, noun category matching is performed on the 50 early appearing vocabulary words selected in S11, that is, illegal data detection based on the noun category matching definition 312 is performed (S13). Here, the presence / absence of a word corresponding to a noun category (21 “food / drink” to 26 “vehicle” in the semantic category classification table of FIG. 5) in 50 early appearance vocabulary words is collated. If there is no word corresponding to the noun category as a result of collation, the data is regarded as invalid data.

  After that, illegal data detection is performed based on daily / greeting category matching for the 50 words that appear early, that is, daily / greeting category matching definition 313 (S14). Here, the presence / absence of a word corresponding to a daily / greeting category (31 “daily routine / greeting” in the semantic category classification table of FIG. 5) in the 50 early appearing vocabulary words is collated. If there is no word corresponding to the daily / greeting category as a result of the collation, it is regarded as illegal data as in S13. S13. The semantic category in the data set is used for the collation in S14.

  As described above, the reason why the data is invalid when there are no words corresponding to both categories is that, according to the survey of Japanese learning children conducted by the inventors, the words in both categories are not limited to the 50 early appearance vocabulary words. If it has been found that there is no word in this category, it is considered that there is a high possibility of illegal data.

  S15: Finally, NV ratio calculation collation between the noun category (N) and the verb category (V) included in the 50 early appearance vocabulary words selected in S11, that is, illegal data detection based on the NV ratio calculation collation definition 314 is performed. To do.

  Specifically, the noun category (21 “food / drink” to 26 “vehicle” in the semantic category classification table of FIG. 5) included in the 50 words of early appearance vocabulary and the verb category (41 “in the semantic category classification table of FIG. 5). “Non-verb ratio” is calculated from the number of words of “operation word”). Here, “NV ratio = noun category number ÷ verb category number” is assumed, and when the calculated NV ratio is outside a specific range, it is regarded as illegal data.

  According to the research conducted by the inventors, the NV ratio of the 50 early vocabulary words of Japanese learning children is an average of “3.32” and a standard deviation of “1.78”, and is based on twice the standard deviation It was found that the data of “95.45%” corresponding to 2SD, which falls under the above, falls within the value of “0 to 6.87”. If this natural scientific phenomenon is used as an index and the NV ratio is outside the range of “0 to 6.87”, it is regarded as illegal data.

  However, it has been found that the NV ratio varies depending on the language to be acquired, and the above range of the NV ratio cannot be applied to languages other than Japanese. According to Non-Patent Document 2, the average NV ratio in English is about 12.0, which is considerably higher than Japanese, and the average in Korean is about 1.8, which is lower than Japanese. The range of the NV ratio may be adjusted.

  As described above, in the processes of S12 to S15, if all the four indicators 311 to 314 are finally satisfied, the respective data sets of the user are validated and used as valid data to be analyzed thereafter. This valid data is held in the intermediate data holding unit 34.

  On the other hand, if any one of the indicators 311 to 314 corresponds to illegal data, the data set of the user is invalidated and excluded from the subsequent analysis targets. At that time, the intermediate data holding unit 34 may be deleted. By applying a criterion based on an index using the characteristics of infant language development, it becomes possible to detect illegal data due to mischief and the like with high accuracy.

(6) Average acquired age generation unit 36
The average acquired age generation unit 36 generates the average acquired age of each input word and the number of posted data for the valid data recognized by the unauthorized data detection unit 35. For example, by selecting all the words registered in the meaning of “mama” and averaging the acquired age of each user of the word, the average acquired age of the word “mama” (eg, 16.7 months) The number of post data input (for example, 123) is generated. The average acquired age and the number of posted data generated here are stored in the memory (RAM).

(7) Data reliability determination unit 37
The data reliability determination unit 37 calculates the average acquired age value (eg, word “Mama”, 16.7 months) of each input word stored in the memory (RAM) and the number of posted data (eg, number value word). The reliability of the effective data held in the intermediate data holding unit 34 is determined using “Mama” (123 cases).

  More specifically, first, the average acquired age value of each word (eg, the word “mama”, 16.7 months old) is collated with the crossing 50% reaching age dictionary 315 in FIG. Determine sex. As shown in FIG. 8, the dictionary 315 assumes that “50% reached age” values of each word are listed in advance and stored in the hard disk drive.

  “50% Reaching Age” stored here is for parents with children aged 10-36 months, answering the checklist (question paper) what words my child currently remembers The data of the answer is tabulated for each age, and the percentage of children in each age group that each word is acquired is calculated (for example, 65% The age at which the acquisition rate of each word reaches 50% is provisionally set as the acquisition age of the word (see Non-Patent Document 3).

  For example, if the acquisition rate is 46% in the 13-month-old group, 51% in the 14-month-old group, and 60% in the 15-month-old group for the word “manma”, the 50-month-old age of “manma” is 14 Set as months of age. This “50% reached age” is completely different from the average acquired age value calculated by the average acquired age generation unit 36 in terms of data collection and calculation methods, but according to the research results of the inventors. , It was found that the value of “50% reaching age” matched at a statistically significant level (the intraclass correlation coefficient was 0.7, which matched significantly at the 5% level).

  Based on such scientific facts, the average acquired age value in the transferred data is collated with the dictionary 315 for each word, and the range of 2 months before and after 50% acquired age (for example, the word “Manma: 14 months old”) (for example, , 12-16 months old), the reliability of the effective data of the average acquired age is affirmed. This reliability affirmative range (two months before and after the average acquired age) is set in the program.

  However, when the number of posted data included in the transfer data is too small, it is difficult to accurately estimate the average acquired age. Therefore, a criterion for validating only the input words having the posted data number of 10 or more is set, and when the posted data number is 9 or less, it is excluded from the valid data as unreliable data. This reference number is also set in the program and can be adjusted as necessary.

(8) Database item generation unit 38
The database item generation unit 38 obtains effective data whose reliability has been affirmed through a series of processing procedures of the units 32 to 37 for each input word (for example, Anpanman) on the average acquired age (for example, 18.3 months). ) ・ Category ID of category (example: 52 characters) ・ Number of posted data (example: 133 cases) ・ Examples of utterances (example: Anpan, Panpan) are summarized and processed as shown in FIG. Create a database. The database generated here is an infant vocabulary development database.

  Therefore, according to the creation apparatus 1, fraudulent data detection (S11 to S15) utilizing infant vocabulary development characteristics of Japanese learning children from web posting data from the user terminal, that is, data collected by a user participation type method. ) And a reliability verification process, a high-precision and high-quality infant vocabulary development database is created.

(9) User interface browsing unit 39
According to the user interface browsing unit 39, a user interface capable of searching and browsing the infant vocabulary development database created by the database item generating unit 38 on the web is provided to the user terminal. That is, as shown in FIG. 10, the browsing unit 39 displays a search page for searching the database on the browser of the user terminal.

  The search result is output in response to the search request input to the search page, and the browser display of the user terminal is switched to the search result page as shown in FIG. Thereby, the user can easily acquire information related to infant vocabulary development through the user terminal. For example, if the pronunciation “Wanwan” is entered in the T column “Search from free keyword” on the search page of FIG. 10 and a search is requested from the pronunciation, the “meaning” (dog) for the voice (wanwan) as shown in FIG. , Animals, NHK characters), “category” (animals), “average acquired age” (18, 1 month old), and acquisition distribution (frequency distribution table) W indicating the percentage (%) spoken It is displayed on the browser of the user terminal.

  Also, if you want to find out what words 16-month-old children tend to memorize on average, you can click on “16 months” on the browser in the item “Check by age” in FIG. Words having an acquired age of “16.0 months to 16.9 months” are displayed as a list on the browser of the user terminal. In addition, by clicking the item “Check from index” and the item “Check from semantic category” in FIG. 10, the corresponding result can be displayed on the browser.

  This provides a “children dictionary” service that allows easy search / browsing of the language development process of infants, and makes it easy to search for information on the growth of children's words that parents want to know. At this time, the infant vocabulary development database is constructed on the basis of reliable post data in which invalid data is rejected through the invalid data detection unit 35 and the data reliability determination unit 37, so that high accuracy regarding infant vocabulary development is obtained.・ It is constructed as a high-quality database.

  Therefore, the growth of the child's words can be estimated and understood more accurately, and evidence-based information disclosure is possible. In addition, if the average acquired age of each word contained in the infant vocabulary development database is used, it can be applied to an early vocabulary development education support system and a spoken dialogue system for infants according to the age. Applicable in various industrial situations.

≪Programs≫
In the present invention, each means 12-18. Some or all of the units 31 to 39 can be configured as a program for causing a computer to function. In this program, S01 to S07. It is possible to cause the computer to execute all or some of steps S11 to S15.

  This program can be provided through a network such as a website or e-mail. The program is recorded on a recording medium such as a CD-ROM, DVD-ROM, CD-R, CD-RW, DVD-R, DVD-RW, MO, HDD, Blu-ray Disk (registered trademark). It is also possible to save and distribute. This recording medium is read using a recording medium driving device, and the program code itself realizes the processing of the above embodiment, so that the recording medium also constitutes the present invention.

1 ... Infant vocabulary development database creation device 11 ... User interface input means (input means)
12 ... Word acquisition age generation means (acquisition age generation means)
13. Infant vocabulary meaning category assigning means (category assigning means)
14: Unauthorized data detection means (injustice detection means)
15 ... Average acquired age generation means 16 ... Data reliability determination means (reliability determination means)
17 ... Database item generation means (generation means)
DESCRIPTION OF SYMBOLS 18 ... User interface browsing means 31 ... User interface input part 32 ... Word acquisition age generation part 33 ... Infant vocabulary meaning category provision part 34 ... Intermediate data holding part 35 ... Incorrect data detection part 36 ... Average acquisition age generation part 37 ... Data reliability Sex determination unit 38 ... Database item generation unit 310 ... Semantic category definition dictionary 311 ... Meaning word age collation definition 312 ... Noun category collation definition 313 ... Daily / greeting category collation definition 314 ... NV ratio calculation collation definition 315 ... Crossing reached 50% Age dictionary

Claims (9)

A method for creating an infant vocabulary development database using word information posted on the web through a user terminal,
An input receiving step of displaying an interface for inputting word information on the terminal and receiving the word information input by the user through the terminal;
A fraud detection step in which the fraud detection means detects fraudulent word information by applying the acquired age or semantic category of the word information received in the input reception step to a plurality of definition criteria prepared in advance, and eliminates the fraud information; ,
An average acquisition age calculation means for calculating an average acquisition age of word information recognized as effective without being excluded in the fraud detection step;
A reliability determining step for determining the reliability of the word information recognized as valid based on the average acquired age calculated in the average age calculating step;
A generating step for generating an infant vocabulary development database based on the word information whose reliability is recognized in the reliability determining step;
A method for creating an infant vocabulary development database. An acquired age generation step for calculating the acquired age from the difference between the input date and time of the word information and the date of birth input through the terminal, with respect to the word information received in the input receiving step. ,
A category assigning step in which the category assigning unit assigns the semantic category with reference to a category dictionary prepared in advance for the word information received in the input accepting step;
A user interface browsing means for displaying on the terminal an interface capable of searching and browsing the database generated in the generating step on the web;
The infant vocabulary development database creation method according to claim 1, further comprising: The fraud detection step includes a step of selecting an arbitrary number of the word information for each user in ascending order of the acquired age;
Recognizing the user's word information group as incorrect information if the acquired word information group includes meaningful word information whose acquired age is before a reference value;
Collating the semantic categories of each of the selected word information, and if there is no word information corresponding to a noun category or daily / greeting category, recognizing the user's word information group as incorrect information;
By referring to the semantic category of each selected word information, the ratio of the word information corresponding to the noun category and the word information belonging to the verb category is calculated, and the ratio is outside the range of the predetermined index. If there is a step of recognizing the user's word information group as illegal information,
The method for creating an infant vocabulary development database according to claim 1, wherein: The reliability determining step is to check the reliability of each word information recognized as valid by comparing the average acquired age with an age dictionary summarizing acquired age for each word;
If the number of word information groups recognized as valid does not exceed a preset reference number, denying the reliability of the word information groups;
The method for creating an infant vocabulary development database according to any one of claims 1 to 3. A device for creating an infant vocabulary development database using word information posted on the web through a user terminal,
An input means for displaying an interface for inputting word information on the terminal, and receiving word information input by a user through the terminal;
The fraud detection means for detecting the fraudulent word information by applying the acquisition age or semantic category of the word information received by the input means to a plurality of definition criteria prepared in advance, and removing the fraud information;
An average acquired age calculating means for calculating an average acquired age of word information recognized as effective without being excluded by the fraud detector;
Reliability determining means for judging the reliability of word information recognized as effective based on the average acquired age calculated by the average age calculating means;
Generating means for generating an infant vocabulary development database based on word information whose reliability is recognized by the reliability determining means;
An infant vocabulary development database creation device characterized by comprising: For the word information received by the input means, acquired age generation means for calculating the acquired age from the difference between the input date and time of the word information and the date of birth input through the terminal;
For the word information received by the input means, referring to a category dictionary prepared in advance, category giving means for giving the semantic category;
User interface browsing means for displaying on the terminal an interface capable of searching and browsing the database generated in the generating step on the web;
The infant vocabulary development database creation device according to claim 5, further comprising: The fraud detection means selects an arbitrary number of the word information for each user in ascending order of the acquired age;
Means for recognizing the user's word information group as fraudulent information if the acquired word information group includes meaningful word information whose acquired age is before a reference value;
Means for recognizing the user's word information group as illegal information if the semantic category of each selected word information is collated and there is no word information corresponding to a noun category or daily / greeting category;
By referring to the semantic category of each selected word information, the ratio of the word information corresponding to the noun category and the word information belonging to the verb category is calculated, and the ratio is outside the range of the predetermined index. If there is, means for recognizing the user's word information group as illegal information,
The infant vocabulary development database creation device according to any one of claims 5 and 6, characterized by comprising: The reliability determination means is a means for checking the reliability of each word information recognized as valid by checking the average acquired age for each word with a month dictionary summarizing acquired age for each word;
If the number of word information groups recognized as valid does not exceed a preset reference number, means for denying the reliability of the word information groups;
The infant vocabulary development database creation device according to any one of claims 5 to 7, further comprising:   An infant vocabulary development database creation program for causing a computer to function as each means constituting the infant vocabulary development database creation device according to any one of claims 5 to 8.

Images