JP7800673B2 - Winning probability acquisition device, method, and program - Google Patents

Description

Article 30, paragraph 2 of the Patent Act applies. (1) Date of website publication: March 10, 2022 Website address: https://www.fukui-c.ed.jp/▲～▼fec/wp-content/uploads/2022/03/2021_03NTT%E8%97%A4%E7%94%B0%E6%B0%8F%E7%89%B9%E5%88%A5%E5%AF%84%E7%A8%BF.pdf

The disclosed technology relates to a technology for obtaining the probability of word acquisition.

The total number of words a person knows is called their vocabulary size. A vocabulary size estimation test is a test that accurately estimates that vocabulary size in a short amount of time (see, for example, Non-Patent Document 1). An overview of the estimation procedure is provided below.

(1) The word list in the word familiarity database is sorted in order of familiarity, and test words are selected so that they are spaced at approximately regular intervals (for example, one word is selected for every 1,000 words). Familiarity (word familiarity) is a numerical representation of how familiar a word is. The higher the familiarity, the more familiar the word.

(2) Present a test word to the user and ask them to respond whether or not they know the word.

(3) A logistic regression analysis is performed to best explain these combinations of test words and answers. In this logistic regression analysis, the independent variable x is the total number of words in the word familiarity DB that have a familiarity level equal to or higher than that of each test word, and the dependent variable y is the probability (e.g., 0 or 1) that the user will answer that they know each word. As a result of the logistic regression analysis, a logistic model (or logistic regression equation) is obtained. An example of a logistic model is shown in Figure 12.

(4) In the logistic model obtained, the value of x corresponding to y = 0.5 is calculated and used as the estimated vocabulary size. Note that the estimated vocabulary size refers to the value estimated to be the user's vocabulary size.

This method uses a word familiarity database to accurately estimate a user's vocabulary size simply by testing whether or not they know the selected test words.

Kobayashi Tetsuo, Amano Shigeaki, Masataka Nobuo, "The Current State and Future of the Mobile Society," 2007, NTT Publishing, pp. 127-128.

No technology has previously been proposed that uses the generated logistic model to obtain the probability that a person has acquired each word contained in a text.

The disclosed technology aims to obtain the probability that a person has acquired each word contained in a text.

One aspect of the disclosed technology is an acquisition probability acquisition device, where intimacy is an index representing familiarity with a word, and the device comprises: a memory unit in which a word intimacy DB storing a plurality of words and a plurality of intimacies corresponding to each of the plurality of words is stored; an intimacy acquisition unit that acquires the intimacy corresponding to each word included in input text from the word intimacy DB stored in the memory unit; a model memory unit that stores a model representing the relationship between a value based on the intimacy corresponding to each word and a value based on the probability that a certain person has acquired each word; and an acquisition probability acquisition unit that acquires the acquisition probability, which is the probability that a certain person has acquired each word, using at least the acquired intimacy corresponding to each word and the model stored in the model memory unit.

The disclosed technology makes it possible to obtain the probability that a person has acquired each word contained in a text.

FIG. 1 is a diagram illustrating an example of the functional configuration of a model generation device and a word selection device. FIG. 2 is a diagram showing an example of the processing procedure of the model generation method and the word selection method. FIG. 3 is a diagram showing an example of a logistic regression model. FIG. 4 is a diagram illustrating an example of the functional configuration of the winning probability acquisition device. FIG. 5 is a diagram showing an example of a processing procedure of the winning probability obtaining method. FIG. 6 is a diagram for explaining an example of generation of acquired word information. FIG. 7 is a diagram illustrating an example of the functional configuration of the device for extracting recommended words to learn. FIG. 8 is a diagram showing an example of a processing procedure of the method for extracting recommended words to learn. FIG. 9 is a diagram showing examples of recommended words to study. FIG. 10 is a diagram illustrating an example of the functional configuration of a computer. FIG. 11 is a diagram showing an example of the correspondence between the familiarity and the number of words. FIG. 12 is a diagram for explaining the background art.

Below, embodiments of the disclosed technology are described with reference to the drawings.

[First embodiment]
First, a first embodiment will be described. The first embodiment is a model generation device and method, and a word generation device and method.

As illustrated in FIG. 1, the model generation device 1 of this embodiment includes a memory unit 11, a word selection unit 12, a presentation unit 13, a response reception unit 14, a model generation unit 15, and a vocabulary size estimation unit 16. The model generation device 1 does not necessarily have to include the word selection unit 12, the presentation unit 13, the response reception unit 14, the memory unit 11, and the vocabulary size estimation unit 16.

As shown by the dashed lines in Figure 1, the word generation device A1 is composed of a memory unit 11 and a word selection unit 12. The word generation device A1 may also include a presentation unit 13 and an answer acceptance unit 14.

<Storage unit 11>
A familiarity database (DB) is stored in advance in the storage unit 11. The word familiarity DB is a database that stores pairs of M words (a plurality of words) and a predetermined familiarity (word familiarity) for each of the words. In other words, a word familiarity DB that stores a plurality of words and a plurality of familiarities corresponding to each of the plurality of words is stored.

The M words in the word familiarity database are ranked in order based on familiarity (e.g., by familiarity). M is an integer equal to or greater than 2, representing the number of words contained in the word familiarity database. While there are no limitations on the value of M, it is desirable to set M to 70,000 or greater when measuring the vocabulary size of a native language, and 10,000 or greater when measuring the vocabulary size of a second language (e.g., English for native Japanese speakers). Because the average Japanese adult vocabulary size is estimated to be approximately 40,000 to 50,000 words, approximately 70,000 words can cover the vocabulary of most people, taking into account individual differences. On the other hand, second languages often have a smaller vocabulary than their native language, and a word count smaller than the M for the native language is likely to cover the vocabulary of most people. However, vocabulary size can vary significantly depending on how vocabulary is counted, including orthographic variations and the handling of derivative words. Therefore, depending on how vocabulary is counted, M may need to be 100,000 or greater for the native language. The estimated vocabulary size is limited to the number of words contained in the reference word familiarity database. Therefore, when vocabulary estimation is also performed for people with large vocabulary sizes who may be outliers, it is desirable to set the value of M to a larger value.

Familiarity (word familiarity) is an index that represents familiarity with a word. Examples of indices that represent familiarity with a word include an index that represents the familiarity of a word (for example, the numerical representation of word familiarity introduced in Non-Patent Document 1), an index that represents how often a word is seen or heard, an index that represents how well a word is known, an index that represents how well a word can be written, and an index that represents how well a word can be spoken using the word.

For example, the higher the intimacy level of a word, the more intimate the word. In this embodiment, the larger the number representing the intimacy level, the higher the intimacy level. However, this does not limit the present invention.

The memory unit 11 receives read requests from the word selection unit 12 and the model generation unit 15 as input, and outputs the words that correspond to the requests and the familiarity of those words.

<Word Selection Unit 12>
Input: Question generation request from the user or the system Output: N test words to be used in the vocabulary size estimation test When the word selection unit 12 receives a question generation request from the user or the system, it selects and outputs a plurality of test words w(1), ..., w(N) to be used in the vocabulary size estimation test from a plurality of ordered words contained in the word familiarity DB in the memory unit 11.

For example, the word selection unit 12 uses a word intimacy DB stored in the memory unit 11 to select multiple test words w(1), ..., w(N) from multiple words so that the intervals between the intimacy levels corresponding to the test words are constant (step S12).

For example, the word selection unit 12 selects N words evenly from all words included in the word familiarity DB of the memory unit 11 so that the familiarity of the selected words is at approximately regular intervals, and outputs the selected N words as test words w(1), ..., w(N).

For example, the word selection unit 12 selects words so that the intimacy interval is 0.1. For example, the word selection unit 12 may select a total of 61 words, including word w(1) with an intimacy level of 1, word w(2) with an intimacy level of 1.1, ..., word w(60) with an intimacy level of 6.9, and word w(61) with an intimacy level of 7.

The familiarity of the test words w(1), ..., w(N) does not necessarily have to be at regular intervals; it is sufficient to select them evenly. If past surveys predict a familiarity level near the boundary between whether the user knows them or not, you can select more words near the familiarity level you want to focus on. In other words, the familiarity values for the series of test words w(1), ..., w(N) may vary in density.

There is no limitation on the order in which the test words w(1), ..., w(N) are output from the word selection unit 12, but the word selection unit 12 outputs the test words w(1), ..., w(N) in order of increasing familiarity, for example.

The number of test words N may be specified in the question generation request or may be predetermined. There is no limit to the value of N, but a value of approximately 50≦N≦100 is desirable. To ensure sufficient estimation, it is desirable for N≧25. A larger N allows for more accurate estimation, but increases the burden on the user (subject) (step S12).

To reduce the burden on users and improve accuracy, tests of 50 words each can be administered multiple times (e.g., three times), the vocabulary size can be estimated for each test, and the answers from multiple tests can be re-estimated together. In this case, fewer words can be tested each time, reducing the burden on users, and allowing users to see the results for each test helps maintain their motivation to answer. Furthermore, combining the words from multiple tests to perform a final vocabulary size estimation can improve estimation accuracy.

By selecting multiple test words so that the intervals between the familiarity levels corresponding to the test words are uniform, the variation in familiarity can be reduced, making it easier for the logistic curve to converge.

<Presentation section 13>
Input: N test words Output: instruction sentence and N test words The presentation unit 13 receives the N test words w(1), ..., w(N) output from the word selection unit 12. The presentation unit 13 presents the test words w(1), ..., w(N) to the user 100 (subject) in accordance with a pre-set display format (step S13).

For example, the presentation unit 13 presents to the user 100 a predetermined instruction sentence prompting the user 100 to input an answer regarding the user's knowledge of the test words, and N test words w(1), ..., w(N) in a format for a vocabulary size estimation test, in accordance with a pre-set display format.

There are no limitations on the presentation format, and this information may be presented as visual information such as text or images, as auditory information such as audio, or as tactile information such as Braille.

For example, the presentation unit 13 may electronically display the instruction sentences and test words on the display screen of a terminal device such as a personal computer (PC), tablet, or smartphone. In other words, the presentation unit 13 may generate screen information for presentation on a display or the like and output it to the display.

Alternatively, the presentation unit 13 may be a printing device, and the instructions and test words may be printed and output on paper or the like. Alternatively, the presentation unit 13 may be a speaker on a terminal device, and the instructions and test words may be output aloud. Alternatively, the presentation unit 13 may be a Braille display, and the instructions and test words may be displayed in Braille.

User 100's response regarding their knowledge of the test words may indicate either "knowing" or "not knowing" the test words (answering that they know or do not know the test words in each rank), or it may indicate one of three or more options including "knowing" and "not knowing." Examples of options other than "knowing" and "not knowing" include "not confident (about knowing)" and "knowing the word, but not knowing its meaning." However, even if user 100 is asked to respond with three or more options including "knowing" and "not knowing," the accuracy of vocabulary size estimation may not be improved compared to when only "knowing" or "not knowing." For example, if user 100 is asked to choose from three options, "knowing," "not knowing," and "not confident," whether "not confident" is selected depends on user 100's personality. In such cases, increasing the number of options does not improve the accuracy of vocabulary size estimation. Therefore, it is usually preferable to have user 100 answer with two choices, such as "knowing" or "not knowing."

However, rather than asking whether they "know" or "don't know," it would also be better to ask for answers from the perspective of whether they "can create example sentences (using the test words)" or "cannot create example sentences," or whether they "can explain the meaning (of the test words)" or "cannot explain the meaning." By clarifying the perspective, the estimated vocabulary size will change. For example, if the question is whether the user "can create example sentences," the vocabulary size that the user believes they can use will be estimated.

Below, we will explain an example in which user 100 is asked to answer the test word by either "knowing" or "not knowing."

Also, for example, the test words are presented in order of increasing familiarity, but the order of presentation is not limited to this, and the test words may be presented in a random order.

<Answer Receiving Unit 14>
Input: User's answer regarding knowledge of test words Output: User's answer regarding knowledge of test words After being presented with the instruction sentence and test words, the user 100 inputs the answer regarding the user's 100 knowledge of the test words into the answer receiving unit 14 (step S14).

For example, the answer receiving unit 14 may be a touch panel of a terminal device such as a PC, tablet, or smartphone, and the user 100 inputs the answer into the touch panel. The answer receiving unit 14 may also be a microphone of the terminal device, in which case the user 100 inputs the answer by voice into the microphone.

The user 100 may input an answer into the answer receiving unit 14 by clicking with a mouse or the like.

The answer accepting unit 14 accepts answers regarding knowledge of the input test words (for example, an answer that the test word is known or an answer that the test word is not known) and outputs the answers as electronic data. The answer accepting unit 14 may output answers for each test word, may output answers for one test together, or may output answers for multiple tests together.

For example, if the answer receiving unit 14 receives an answer from the user 100 that the user 100 knows the test word, it assigns a numerical value of 1 to the answer regarding knowledge of that test word. On the other hand, if the answer receiving unit 14 receives an answer from the user 100 that the user 100 does not know the test word, it assigns a numerical value of 0 to the answer regarding knowledge of that test word. These numerical values are output to the model generation unit 15.

<Model Generation Unit 15>
Input: Answers about knowledge of test words from user Output: Model The answers about knowledge of test words from user 100 output from the answer receiving unit 14 are input to the model generating unit 15 .

The model generation unit 15 uses the answers regarding knowledge of the test words and the word familiarity DB stored in the memory unit 11 to obtain a model that represents the relationship between a value based on the familiarity corresponding to the test word and a value based on the probability that the user 100 will answer that they know the test word (step S15). The obtained model is output to the vocabulary size estimation unit 16.

The value based on the intimacy corresponding to the test word may be the intimacy corresponding to the test word itself, or may be a non-monotonic decreasing function value of the intimacy corresponding to the test word (e.g., a monotonically increasing function value). For simplicity of explanation, the following will illustrate the case where the value based on the intimacy corresponding to the test word is the intimacy corresponding to the test word itself.

The value based on the probability that user 100 will answer that he knows the test word may be the probability that user 100 will answer that he knows the test word itself, or may be a non-monotonic decreasing function value (e.g., a monotonically increasing function value) of the probability that user 100 will answer that he knows the test word. For simplicity of explanation, the following will illustrate a case where the value based on the probability that user 100 will answer that he knows the test word is the probability that user 100 will answer that he knows the test word itself.

Although there are no limitations on the model, one example of a model is a logistic regression model (logistic model). For simplicity of explanation, the following example illustrates a logistic curve y = f(x, Ψ) where the degree of familiarity corresponding to the test word is the independent variable x and the probability that the user 100 will respond that they know each word is the dependent variable y. Ψ is a model parameter.

The model generation unit 15 refers to the word familiarity DB stored in the memory unit 11 to obtain the familiarity corresponding to the test word w(n) that the user 100 answered that they know, and defines the obtained familiarity as x(n). This familiarity x(n) is the familiarity corresponding to the test word w(n).

For a test word w(n) that user 100 answers as knowing, the model generation unit 15 sets a point (x, y) = (x(n), 1) where the probability y that user 100 answers as knowing the test word w(n) is 1 (i.e., 100%) and the familiarity corresponding to the test word w(n) is x(n).

In addition, for a test word w(n) that the user 100 answers that they do not know (or does not answer that they know), the model generation unit 15 sets a point (x, y) = (x(n), 0) where the probability y that the user 100 answers that they know the test word w(n) is 0 (i.e., 0%) and the familiarity corresponding to the test word w(n) at that time is x(n).

The model generation unit 15 fits a logistic curve to each point (x, y) = (x(n), 1) or (x(n), 0) where n = 1, ..., N, and obtains, as a model, a logistic curve y = f(x, Ψ) that minimizes the error. In other words, the model generation unit 15 obtains, as a model, a logistic curve y = f(x, Ψ) that minimizes the error for each point (x, y) = (x(n), 1) or (x(n), 0) where n = 1, ..., N.

Figure 3 shows an example of a model of the logistic curve y = f(x, Ψ). In Figure 3, the horizontal axis represents familiarity, and the vertical axis represents the probability (y) of responding that the user knows the word. The circles represent the point (x, y) = (x(n), 1) for the test word w(n) that user 100 responds that they know, and the point (x, y) = (x(n), 0) for the test word w(n) that user 100 responds that they do not know (or do not respond that they know). "AIC" in Figure 3 represents the Akaike Information Criterion, and the smaller the value, the better the fit of the model. "n" in Figure 3 represents the number of test words.

Here, generation can also be rephrased as creation or construction. Therefore, the model generation unit 15 may be a model creation unit 15 or a model construction unit 15. Furthermore, a model may be created or constructed.

<Vocabulary size estimation unit 16>
Input: Model Output: Vocabulary size of user 100 The vocabulary size estimation unit 16 estimates the vocabulary size of user 100 based on the model (step S16).

Below, estimation methods 1 to 3 are explained as examples of methods for estimating the vocabulary size of user 100 by the vocabulary size estimation unit 16.

(Estimation method 1)
The vocabulary size estimation unit 16 obtains a predetermined value of familiarity, which is the familiarity when the value based on the probability that the user 100 will answer that they know the word in the model is a predetermined value or close to the predetermined value. Examples of the predetermined value are 0.5 or 0.8. Of course, the predetermined value may be any other value greater than 0 and less than 1.

Then, the vocabulary size estimation unit 16 refers to the word intimacy DB stored in the memory unit 11 to obtain the number of words with an intimacy level equal to or greater than a predetermined value of acquired intimacy, and sets the obtained number as the vocabulary size of the user 100.

(Estimation method 2)
The vocabulary size estimation unit 16 refers to the model and the word familiarity DB stored in the storage unit 11, and obtains an output value y(m) when familiarity x(m) corresponding to word w(m) included in the word familiarity DB is input to the model. In other words, the vocabulary size estimation unit 16 calculates the value of y corresponding to familiarity x(m) corresponding to word w(m) in the model, and sets the calculated value as the output value y(m). The vocabulary size estimation unit 16 performs this process for each word w(m) (m = 1, ..., M) included in the word familiarity DB, thereby obtaining the output value y(m) (m = 1, ..., M).

Then, the vocabulary size estimation unit 16 calculates Σ _m=1 ^M y(m) and sets this calculated value as the vocabulary size of the user 100 .

In this case, if the word w(m) is a test word and an answer regarding knowledge of the test word w(m) has been obtained, the vocabulary size estimation unit 16 may estimate the vocabulary size of the user 100 by taking into account the answer regarding knowledge of the test word w(m).

For example, the vocabulary size estimation unit 16 sets y(m) = 1 if the answer regarding knowledge of the test word w(m) is "knowledge," and sets y(m) = 0 if the answer regarding knowledge of the test word w(m) is "not knowledge." For y(m) of words other than the test word, the output value y(m) obtained from the model as described above is used.

Then, the vocabulary size estimation unit 16 uses these y(m) to calculate Σ _m=1 ^M y(m), and sets this calculated value as the vocabulary size of the user 100 .

By taking into account answers regarding knowledge of test words, a more appropriate vocabulary size can be estimated.

By estimating the vocabulary size based on a logistic model estimated from y, the probability that user 100 will respond that they know the test word, and x, the familiarity of the test word, the model converges more easily than when x is used directly, allowing for a more robust estimation of the vocabulary size. Furthermore, even if the distribution of word counts corresponding to each familiarity level differs significantly, sudden changes in the estimated vocabulary size can be suppressed.

(Estimation method 3)
The vocabulary size estimation unit 16 references the model and the word familiarity DB stored in the storage unit 11 to obtain an output value y(i) when a familiarity x(i) included in the word familiarity DB is input to the model. In other words, the vocabulary size estimation unit 16 calculates the value of y corresponding to the familiarity x(i) in the model and sets the calculated value as the output value y(i). The vocabulary size estimation unit 16 also references the word familiarity DB stored in the storage unit 11 to obtain the number n(i) of words included in the word familiarity DB that correspond to the familiarity x(i). The vocabulary size estimation unit 16 performs these processes on each familiarity x(i) (i = 1, ..., I) included in the word familiarity DB to obtain the output value y(i) (i = 1, ..., I) and the number of words n(i) (i = 1, ..., I), where I is the number of familiarity types.

Then, the vocabulary size estimation unit 16 calculates Σ _i=1 ^I y(i)×n(i) and sets this calculated value as the vocabulary size of the user 100 .

If the degree of intimacy is the same, the corresponding y value will be the same. Also, there may be words with the same degree of intimacy. For this reason, by performing calculations for each degree of intimacy, as in estimation method 3 rather than estimation method 2, the vocabulary size estimation calculation can be performed more quickly.

<Modification of the first embodiment>
The word selection unit 12 may simply select a plurality of test words w(1), . . . , w(N) from a plurality of words, rather than selecting test words such that the intervals between the familiarity scores corresponding to the test words are constant.

In addition, the model generation unit 15 may hypothesize answers regarding knowledge of non-presented words and obtain a model representing the relationship between values based on the familiarity corresponding to the test words and non-presented words and the probability or hypothetical value that the user 100 will answer that he or she knows the test words and non-presented words.

Here, non-presented words are words other than the multiple test words among the multiple words. To facilitate convergence of the logistic model, answers for non-presented words that were not used as test words are assumed and used to create the model. Words close to the upper limit of familiarity are words that many people know, and words close to the lower limit are words that many people do not know. Therefore, if user 100 answers that they know the test word with the highest familiarity, it is assumed that they also know non-presented words with a familiarity level higher than that. Conversely, if the user answers that they do not know the test word with the lowest familiarity, it is assumed that they also do not know non-presented words with a familiarity level lower than that.

In other words, assuming that non-presented words are presented to user 100, user 100's response regarding knowledge of the non-presented words will be that he knows the words with a higher familiarity than the maximum familiarity value of the test words, and that he does not know the words with a lower familiarity than the minimum familiarity value of the test words.

For example, if a user answers that they know a test word with a familiarity level of 6.5, it is assumed that they also know non-presented words with familiarity levels of 6.7 and 6.9. Furthermore, if a user answers that they do not know a test word with a familiarity level of 2, it is assumed that they do not know non-presented words with familiarity levels of 1.8 and 1.6.

In this way, by adding non-presented words, which are words that were not presented to user 100, and answers regarding knowledge of non-presented words, and estimating the model, the model is more likely to converge, and a more appropriate model can be generated. This makes it easier for the model to converge, and a more appropriate model can be generated, even if, for example, user 100 answers that he knows most of the test words, or even if user 100 answers that he does not know most of the test words.

[Second embodiment]
A second embodiment will now be described. The second embodiment is an acquisition probability acquisition device and method.

The following will focus on the differences between the first embodiment and the modified version of the first embodiment. Explanations of matters that have already been explained may be omitted.

As illustrated in Figure 4, the acquisition probability acquisition device 2 of this embodiment includes a memory unit 11, a model memory unit 21, a word extraction unit 22, an intimacy acquisition unit 23, an acquisition probability acquisition unit 24, and an acquired word information generation unit 25. The acquisition probability acquisition device 2 does not necessarily have to include the word extraction unit 22 and the acquired word information generation unit 25.

<Storage unit 11>
The storage unit 11 is the same as the storage unit 11 in the first embodiment.

The memory unit 11 stores a word intimacy DB that stores multiple words and multiple intimacy levels corresponding to each of the multiple words. Here, intimacy is an index that represents the familiarity with a word.

<Model storage unit 21>
The model storage unit 21 stores a model that represents the relationship between a value based on the familiarity corresponding to each word and a value based on the probability that a certain person has acquired each word. Here, the "certain person" refers to a person who acquires the acquisition probability. The "certain person" may be a user 100.

Here, acquiring a word means, in other words, knowing a word, being able to use a word, understanding a word, or being able to explain a word.

This example model is a model generated by the model generation device 1 of the first embodiment and a variant of the first embodiment.

As shown by the dashed line in Figure 4, the acquisition probability acquisition device 2 may further include a model generation device 1 for generating a model to be stored in the model memory unit 21.

That is, the acquisition probability acquisition device 2 may further include (1) a word selection unit 12 that selects multiple test words from multiple words, (2) a presentation unit 13 that presents the test words to the user, (3) an answer acceptance unit 14 that accepts the user's answers regarding their knowledge of the test words, and (4) a model generation unit 15 that uses the answers regarding their knowledge of the test words and a word familiarity DB stored in the memory unit 11 to obtain a model that represents the relationship between a value based on the familiarity corresponding to the test word and a value based on the probability that the user will answer that they know the test word, and sets the obtained model as the model stored in the model memory unit.

<Word Extraction Unit 22>
Input: Text Output: Words The word extraction unit 22 extracts each word contained in the input text (step S22).

Each extracted word is output to the intimacy acquisition unit 23.

The text input to the word extraction unit 22 may be any text that is readable by the word extraction unit 22, which is an information processing device. Examples of text include texts published in books such as textbooks and novels, newspapers and magazines, and web pages.

The word extraction unit 22 extracts each word contained in the text, for example, by performing morphological analysis on the input text.

<Familiarity acquisition unit 23>
Input: Word Output: Word, Intimacy Each word extracted by the word extraction unit 22 is input to the intimacy acquisition unit 23. The intimacy acquisition unit 23 acquires the intimacy corresponding to each word from the word intimacy DB stored in the storage unit 11 (step S23).

If the acquisition probability acquisition device 2 does not have a word extraction unit 22, each word contained in the text is input. In this case, the intimacy acquisition unit 23 acquires the intimacy corresponding to each word contained in the text from the word intimacy DB stored in the memory unit 11 (step S23).

Each word and the corresponding familiarity are output to the acquisition probability acquisition unit 24.

Note that the intimacy acquisition unit 23 and the word extraction unit 22 may not acquire intimacy for words that are function words such as proper nouns, numerals, particles, etc. In other words, the word extraction unit 22 may acquire intimacy only for words that are content words.

Function words such as numerals and particles are words that many people know. For this reason, by obtaining the familiarity of these function words, in other words, by processing these function words, the proportion of estimated acquired words in the text calculated by the acquired word information generation unit 25 can be increased. Conversely, by not obtaining the familiarity of these function words, in other words, by not processing these function words, the proportion of estimated acquired words in the text calculated by the acquired word information generation unit 25 can be lowered.

The intimacy acquisition unit 23 may also ignore words that are not included in the word intimacy DB without acquiring the intimacy. This allows the acquisition probability acquisition process to be performed appropriately even if the morphological analysis is incorrect.

<Acquisition Probability Acquisition Unit 24>
Input: word, intimacy Output: word, acquisition probability The acquisition probability acquisition unit 24 acquires the acquisition probability, which is the probability that a person has acquired each word, using at least the intimacy corresponding to each word and the model stored in the model memory unit 21 (step S24).

The acquisition probability acquisition unit 24 obtains an output value when the familiarity corresponding to each word is input into the model, and sets the obtained output value as the acquisition probability corresponding to each word. In other words, the acquisition probability acquisition unit 24 calculates the value of y corresponding to the familiarity x corresponding to each word in the model, and sets the calculated value as the acquisition probability corresponding to each word.

If the model stored in the model memory unit 21 is a logistic curve y = f(x, Ψ) in which the independent variable x is the familiarity corresponding to a word and the dependent variable y is the probability that a person will respond that they know each word, the acquisition probability acquisition unit 24 calculates the value of y = f(x, Ψ) corresponding to the familiarity x corresponding to each word, and sets the calculated value as the acquisition probability corresponding to each word.

The acquisition probability acquisition unit 24 may acquire the acquisition probability by taking into account parts of speech, word length, etc. For example, the acquisition probability acquisition unit 24 may also acquire the acquisition probability by using parts of speech, word length, etc. as explanatory variables.

Each word and its corresponding acquisition probability are output to the acquired word information generation unit 25.

<Acquired word information generation unit 25>
Input: Word, acquisition probability Output: Acquired word information The acquired word information generating unit 25 generates acquired word information, which is information related to the acquisition of words included in the text, using the acquisition probability corresponding to each word (step S25).

Examples of acquired word information include at least one of estimated acquired words in the text, the number of estimated acquired words in the text, and the percentage of estimated acquired words in the text.

Below, we will explain examples of how to calculate the estimated acquired words in the text, the number of estimated acquired words in the text, and the proportion of estimated acquired words in the text.

(Estimated acquired words in the text)
First, the acquired word information generation unit 25 estimates the vocabulary size of a person. The vocabulary size can be estimated by the method described in the vocabulary size estimation unit 16 of the first embodiment. To estimate the vocabulary size, the word familiarity DB from the storage unit 11 and the model from the model storage unit 21 may be input to the acquired word information generation unit 25, as shown by the dashed dotted line in FIG. 4 .

Next, the acquired word information generation unit 25 obtains GOISU(k), the number of words with an intimacy level equal to or higher than the intimacy level corresponding to each input word w(k). To obtain GOISU(k), the word intimacy DB may be input from the memory unit 11 to the acquired word information generation unit 25, as shown by the dashed line in Figure 4.

The acquired word information generation unit 25 then determines words in the text whose GOISU(k) is equal to or less than the person's vocabulary size as estimated acquired words. Generally, the higher the familiarity of a word, the smaller the GOISU(k). Therefore, it can be assumed that a person knows words with a GOISU(k) equal to or less than the person's vocabulary size.

Figure 6 shows an example of GOISU(k).

(estimated number of words acquired in the text)
First, the acquired word information generation unit 25 estimates the vocabulary size of a person. The vocabulary size can be estimated by the method described in the vocabulary size estimation unit 16 of the first embodiment. To estimate the vocabulary size, the word familiarity DB from the storage unit 11 and the model from the model storage unit 21 may be input to the acquired word information generation unit 25, as shown by the dashed dotted line in FIG. 4 .

Next, the acquired word information generation unit 25 obtains GOISU(k), the number of words with an intimacy level equal to or higher than the intimacy level corresponding to each input word w(k). To obtain GOISU(k), the word intimacy DB may be input from the memory unit 11 to the acquired word information generation unit 25, as shown by the dashed line in Figure 4.

Then, the acquired word information generation unit 25 determines the number of words whose GOISU(k) is less than or equal to the vocabulary size of a person as the number of estimated acquired words in the text.

(Percentage of estimated acquired words in the text)
The acquired word information generating unit 25 calculates a value determined by, for example, the following formula (1) or (2), and sets the calculated value as the proportion of estimated acquired words in the text.

(Σ _k=1 ^K y(k) FREQ(k))/Σ _k=1 ^K FREQ(k)…(1)
(Σ _k=1 ^K y(k) DIFF(k))/Σ _k=1 ^K DIFF(k)…(2)
Here, FREQ(k) is the number of times that word w(k) appears in the text. If the text is divided into multiple parts, DIFF(k) is the number of parts in which word w(k) appears. Examples of parts are predetermined units that make up the text, such as a unit, chapter, or section. The entire text may also be used as the unit. K is the total number of words included in the text and for which the acquisition probability has been acquired by the acquisition probability acquisition unit 24.

The acquired word information generation unit 25 counts FREQ(k) and DIFF(k) based on the input words. The acquired word information generation unit 25 uses the FREQ(k) and DIFF(k) obtained by counting to calculate the value determined by equation (1) or equation (2).

Figure 6 shows an example of FREQ(k) and DIFF(k).

Generally, the more familiar a word is, the more frequently it appears, and the less familiar a word is, the less frequently it appears. Therefore, rare words will appear less frequently in a text than familiar words will appear in a text.

For this reason, the proportion of estimated acquired words in a text calculated using formula (1) using FREQ(k) is expected to be higher than the proportion of estimated acquired words in a text calculated using formula (2) using DIFF(k). Whether formula (1) or (2) is used is determined appropriately depending on factors such as the type of information needed as acquired word information.

The acquired word information generation unit 25 may use the number of estimated acquired words in the text/K as the proportion of estimated acquired words in the text. The number of estimated acquired words in the text can be obtained using the method described in (Number of estimated acquired words in the text).

[Third embodiment]
A third embodiment will now be described. The third embodiment is an apparatus and method for extracting recommended words to study.

The following will focus on the differences between the first embodiment and the modified version of the first embodiment. Explanations of matters that have already been explained may be omitted.

As illustrated in Figure 7, the learning recommendation word extraction device 3 of this embodiment includes a memory unit 11, a model memory unit 31, an acquisition probability acquisition unit 32, and a learning recommendation word extraction unit 33.

<Storage unit 11>
The storage unit 11 is the same as the storage unit 11 in the first embodiment.

The memory unit 11 stores a word intimacy DB that stores multiple words and multiple intimacy levels corresponding to each of the multiple words. Here, intimacy is an index that represents the familiarity with a word.

<Model storage unit 31>
The model storage unit 31 stores a model that represents the relationship between a value based on the familiarity corresponding to each word and a value based on the probability that a certain person has acquired each word. Here, the "certain person" refers to a person from whom recommended words for learning are extracted. The "certain person" may be a user 100.

This example model is a model generated by the model generation device 1 of the first embodiment and a variant of the first embodiment.

As shown by the dashed line in Figure 6, the learning recommendation word extraction device 3 may further include a model generation device 1 for generating models to be stored in the model memory unit 31.

That is, the learning recommendation word extraction device 3 may further include (1) a word selection unit 12 that selects multiple test words from multiple words, (2) a presentation unit 13 that presents the test words to the user, (3) an answer acceptance unit 14 that accepts the user's answers regarding their knowledge of the test words, and (4) a model generation unit 15 that uses the answers regarding their knowledge of the test words and a word familiarity DB stored in the memory unit 11 to obtain a model that represents the relationship between a value based on the familiarity corresponding to the test word and a value based on the probability that the user will answer that they know the test word, and sets the obtained model as the model stored in the model memory unit.

<Acquisition Probability Acquisition Unit 32>
Input: word Output: word, acquisition probability A word set consisting of a plurality of words that are candidates for recommended words to learn is input to the acquisition probability acquisition unit 32 .

The acquisition probability acquisition unit 32 uses at least the word familiarity DB stored in the memory unit 11 and the model stored in the model memory unit 31 to acquire the acquisition probability, which is the probability that a person has acquired each word included in the input word set (step S32).

The acquisition probability acquisition unit 32 obtains an output value when the familiarity corresponding to each word is input into the model, and sets the obtained output value as the acquisition probability corresponding to each word. In other words, the acquisition probability acquisition unit 32 calculates the value of y corresponding to the familiarity x corresponding to each word in the model, and sets the calculated value as the acquisition probability corresponding to each word.

If the model stored in the model memory unit 31 is a logistic curve y = f(x, Ψ) in which the independent variable x is the familiarity corresponding to a word and the dependent variable y is the probability that a person will respond that they know each word, the acquisition probability acquisition unit 32 calculates the value of y = f(x, Ψ) corresponding to the familiarity x corresponding to each word, and sets the calculated value as the acquisition probability corresponding to each word.

The acquisition probability acquisition unit 32 may acquire the acquisition probability by taking into account parts of speech, word length, etc. For example, the acquisition probability acquisition unit 32 may acquire the acquisition probability by using parts of speech, word length, etc. as explanatory variables.

Each word and its corresponding acquisition probability are output to the acquired word information generation unit 25.

<Recommended study word extraction unit 33>
Input: word, acquisition probability Output: recommended words to learn The recommended words to learn extraction unit 33 extracts recommended words to learn from the word set based on the acquired acquisition probability (step S33).

For example, the study recommendation word extraction unit 33 may extract words whose acquisition probability is close to a predetermined probability as study recommendation words.

The predetermined probability is a number greater than 0 and less than 1. An example of a predetermined probability is 0.5.

The recommended learning word extraction unit 33 may extract a predetermined number of words that are close to a predetermined probability as recommended learning words.

If the specified probability is 0.5 and the specified number is 7, for example, the seven words shown in Figure 9 are extracted as recommended words to learn. In Figure 9, ENTRY is the notation of the word, PSY is the familiarity, Prob is the probability of acquisition, YN is information about the answer from user 100 if they know or do not know these words, and Distance50 is the magnitude of the difference between the specified probability of 0.5 in this case and Prob.

In this example, since no answer was received from user 100 indicating that they knew or did not know the word, a "-" is displayed in YN. If user 100 responded that they knew the word, a "1" is displayed in YN, and if user 100 responded that they did not know the word, a "0" is displayed in YN.

Suggested words to learn are presented to the person for whom the suggested words to learn are to be extracted. The suggested words to learn may be presented to the person for whom the suggested words to learn are to be extracted in the form of a table as shown in Figure 9.

The study recommendation word extraction unit 33 may extract words that fall within a specified range that includes a specified probability as study recommendation words.

The study recommendation word extraction unit 33 may extract, as study recommendation words, words of a predetermined part of speech whose acquired probability is close to a predetermined probability. Examples of predetermined parts of speech are verbs, nouns, and adjectives. The predetermined parts of speech may be two or more parts of speech. In this case, the study recommendation word extraction unit 33 may extract, as study recommendation words, words whose acquired probability is close to a predetermined probability from each of words of two or more parts of speech.

The part-of-speech information may be stored in a word familiarity DB. In this case, the learning recommendation word extraction unit 33 can refer to the word familiarity DB to obtain the part-of-speech of the word and perform the above processing.

The learning recommendation word extraction unit 33 may refer to a dictionary that stores words and their parts of speech stored in a memory unit not shown in the figure, obtain the word's part of speech, and perform the above processing.

<Modification of the third embodiment>
The word set consisting of multiple words that are candidates for recommended words to be studied and input to the acquisition probability acquisition unit 32 may be words contained in a predetermined text. For this purpose, the recommended word extraction device 3 may be equipped with a word extraction unit 34 described below.

<Word Extraction Unit 34>
Input: Text Output: Words The word extraction unit 34 extracts each word contained in the input text (step S34).

Each extracted word is output to the acquisition probability acquisition unit 32 as a set of words that are candidates for recommended learning words.

The text input to the word extraction unit 34 may be any text that is readable by the word extraction unit 22, which is an information processing device. Examples of text include texts published in books such as textbooks and novels, newspapers and magazines, and web pages.

The word extraction unit 34 extracts each word contained in the text, for example, by performing morphological analysis on the input text.

[Variations]
The present disclosure is not limited to the above-described embodiments, and various modifications and applications are possible within the scope of the gist of the present disclosure.

The various processes described in the embodiments may not only be performed chronologically in the order described, but may also be performed in parallel or individually depending on the processing capacity of the device performing the processes or as needed.

For example, data exchange between the components of the model generation device 1, acquisition probability acquisition device 2, and learning recommendation word extraction device 3 may be performed directly or via a memory unit not shown.

[Programs, recording media]
The processing of each unit of each of the above-mentioned devices may be realized by a computer, in which case the processing content of the functions that each device should have is described by a program. Then, by loading this program into storage unit 1020 of computer 1000 shown in Figure 10 and operating arithmetic processing unit 1010, input unit 1030, output unit 1040, display unit 1060, etc., various processing functions of each of the above-mentioned devices are realized on the computer.

The program describing this processing content can be recorded on a computer-readable recording medium. A computer-readable recording medium is, for example, a non-transitory recording medium, specifically a magnetic recording device, optical disk, etc.

This program may be distributed, for example, by selling, transferring, or lending portable recording media such as DVDs or CD-ROMs on which the program is recorded. Furthermore, this program may be distributed by storing it in a storage device of a server computer and transferring it from the server computer to other computers via a network.

A computer executing such a program, for example, first stores the program recorded on a portable recording medium or transferred from a server computer in its own non-transitory storage device, auxiliary storage unit 1050. Then, when executing a process, the computer loads the program stored in its own non-transitory storage device, auxiliary storage unit 1050, into storage unit 1020 and executes processing in accordance with the loaded program. Alternatively, as an alternative execution form of this program, the computer may load the program directly from a portable recording medium into storage unit 1020 and execute processing in accordance with the program. Furthermore, each time a program is transferred from a server computer to this computer, the computer may execute processing in accordance with the received program. Alternatively, the above-described processing may be executed using a so-called ASP (Application Service Provider) type service, which realizes processing functions simply by issuing execution instructions and obtaining results, without transferring the program from the server computer to this computer. In this embodiment, the program includes information used for processing by an electronic computer that is equivalent to a program (such as data that is not a direct instruction to a computer but has properties that dictate computer processing).

In addition, in this form, the device is configured by executing a specified program on a computer, but at least some of the processing content may also be realized in hardware.

For example, the word selection unit 12, presentation unit 13, answer acceptance unit 14, model generation unit 15, vocabulary size estimation unit 16, word extraction unit 22, intimacy acquisition unit 23, acquisition probability acquisition unit 24, acquired word information generation unit 25, acquisition probability acquisition unit 32, recommended learning word extraction unit 33, and word extraction unit 34 may be configured by a processing circuit.

Furthermore, the memory unit 11, the model memory unit 21, and the model memory unit 31 may be configured as memory.

The following notes are further disclosed regarding the above embodiments.

(Additional note 1)
Memory and
at least one processor coupled to said memory;
Including,
The memory stores a word familiarity DB in which a plurality of words and a plurality of familiarity degrees respectively corresponding to the plurality of words are stored, the familiarity being an index representing familiarity with a word;
The processor:
selecting a plurality of test words from the plurality of words using a word familiarity DB stored in the memory so that the familiarity intervals corresponding to the test words are constant;
Word selection device.

(Additional note 2)
A non-transitory storage medium storing a program executable by a computer to perform a word selection process,
The word selection process includes:
The familiarity is an index representing familiarity with a word, and a word familiarity DB storing a plurality of words and a plurality of familiarity degrees corresponding to the plurality of words is used to select a plurality of test words from the plurality of words so that the intervals between the familiarity degrees corresponding to the test words are constant.
Non-transitory storage medium.

(Additional note 3)
Memory and
at least one processor coupled to said memory;
Including,
The memory stores a word familiarity DB in which a plurality of words and a plurality of familiarity degrees respectively corresponding to the plurality of words are stored, the familiarity being an index representing familiarity with a word;
The processor:
a method for inputting a plurality of test words and answers regarding knowledge of the test words of a user to whom the plurality of test words have been presented, and obtaining a model representing the relationship between a value based on the familiarity corresponding to the test words and a value based on the probability that the user will answer that they know the test words, using the answers regarding knowledge of the test words and a word familiarity DB stored in the memory;
Model generation device.

(Additional note 4)
A non-transitory storage medium storing a program executable by a computer to perform a model generation process,
The model generation process includes:
a method for inputting a plurality of test words and answers of a user to which the plurality of test words have been presented, the method using the answers to the test word knowledge and a word familiarity DB to obtain a model representing the relationship between a value based on familiarity corresponding to the test words and a value based on the probability that the user will answer that they know the test words;
The intimacy is an index representing familiarity with a word, and the word intimacy DB stores a plurality of words and a plurality of intimacies corresponding to the plurality of words, respectively.
Non-transitory storage medium.

(Additional note 5)
Memory and
at least one processor coupled to said memory;
Including,
The memory includes:
The intimacy is an index representing the familiarity of a word, and a word intimacy DB storing a plurality of words and a plurality of intimacies corresponding to the plurality of words;
a model representing the relationship between a value based on the familiarity corresponding to each word and a value based on the probability that a person has acquired each word;
is stored,
The processor:
The familiarity corresponding to each word included in the input text is obtained from the word familiarity DB stored in the memory;
acquiring an acquisition probability, which is the probability that the certain person has acquired each of the words, by using at least the familiarity corresponding to each of the acquired words and the model stored in the memory;
Acquisition probability acquisition device.

(Additional note 6)
A non-transitory storage medium storing a program executable by a computer to execute an acquisition probability acquisition process,
The acquisition probability acquisition process includes:
The intimacy is an index representing the familiarity of a word, and the intimacy corresponding to each word included in the input text is obtained from a word intimacy DB that stores a plurality of words and a plurality of intimacies corresponding to each of the plurality of words;
acquiring an acquisition probability, which is the probability that each word is acquired by the certain person, by using at least a model representing the relationship between a value based on the familiarity corresponding to each word and a value based on the probability that each word is acquired by the certain person, and the familiarity corresponding to each acquired word;
Non-transitory storage medium.

(Supplementary Note 7)
Memory and
at least one processor coupled to said memory;
Including,
The memory includes:
The intimacy is an index representing the familiarity of a word, and a word intimacy DB storing a plurality of words and a plurality of intimacies corresponding to the plurality of words;
a model representing the relationship between a value based on the familiarity corresponding to each word and a value based on the probability that a person has acquired each word;
is stored,
The processor:
using at least the word familiarity DB stored in the memory and the model stored in the memory, to obtain an acquisition probability that is the probability that the certain person has acquired each word included in the input word set;
extracting learning recommendation words from the word set based on the acquired acquisition probability;
A device for extracting recommended learning words.

(Supplementary Note 8)
A non-transitory storage medium storing a program executable by a computer to execute a recommended learning word extraction process,
The recommended learning word extraction process includes:
The intimacy is an index representing the familiarity with a word, and the acquisition probability, which is the probability that a certain person has acquired each word included in the input word set, is obtained by using at least a word intimacy DB that stores a plurality of words and a plurality of intimacies corresponding to each of the plurality of words, and a model that represents the relationship between a value based on the intimacy corresponding to each word and a value based on the probability that a certain person has acquired each word,
extracting learning recommendation words from the word set based on the acquired acquisition probability;
Non-transitory storage medium.

All publications, patent applications, and technical standards mentioned in this specification are incorporated by reference into this specification to the same extent as if each individual publication, patent application, and technical standard was specifically and individually indicated to be incorporated by reference.

Claims (5)

a storage unit in which a word familiarity DB is stored, the familiarity being an index representing familiarity with a word, and in which a plurality of words and a plurality of familiarity degrees respectively corresponding to the plurality of words are stored;
a familiarity acquiring unit that acquires familiarity corresponding to each word included in the input text from a word familiarity DB stored in the storage unit;
a model storage unit that stores a model representing the relationship between a value based on the familiarity corresponding to each word and a value based on the probability that a person has acquired each word;
an acquisition probability acquisition unit that acquires an acquisition probability, which is the probability that each word is acquired by the certain person, by using at least the familiarity corresponding to each of the acquired words and a model stored in the model storage unit;
An acquisition probability acquisition device including: 2. The winning probability acquisition device of claim 1,
The acquisition probability acquisition device further includes: (1) a word selection unit that selects a plurality of test words from a plurality of words; (2) a presentation unit that presents the test words to a user; (3) an answer acceptance unit that accepts the user's answer regarding knowledge of the test words; and (4) a model generation unit that uses the answer regarding knowledge of the test words and a word familiarity DB stored in the storage unit to obtain a model that represents the relationship between a value based on familiarity corresponding to the test words and a value based on the probability that the user will answer that he or she knows the test word, and sets the obtained model as the model stored in the model storage unit.
Acquisition probability acquisition device. 3. The winning probability acquisition device according to claim 1 or 2,
and further comprising an acquired word information generating unit that generates acquired word information, which is information regarding the acquisition of words included in the text, using the acquired acquisition probability corresponding to each of the words.
Acquisition probability acquisition device. an intimacy acquisition step in which an intimacy acquisition unit acquires an intimacy corresponding to each word included in the input text from a word intimacy DB storing a plurality of words and a plurality of intimacies corresponding to each of the plurality of words, the intimacy being an index representing intimacy with respect to a word;
an acquisition probability acquisition step in which an acquisition probability acquisition unit acquires an acquisition probability, which is the probability that each word is acquired by the certain person, using at least a model representing the relationship between a value based on the familiarity corresponding to each word and a value based on the probability that each word is acquired by the certain person, and the familiarity corresponding to each word acquired;
Including how to obtain the probability of winning. A program for causing a computer to function as each part of the winning probability acquisition device of claim 1.