JP7139271B2 - Information processing device, information processing method, and program - Google Patents

Description

The present invention relates to an information processing device, an information processing method, and a program.

Research has been conducted to replace the content of a sentence with an expression that can be used for machine learning by using a specific expression (for example, a proper noun) included in the sentence as a feature amount. In relation to this, the input text is segmented into phrases by morphological analysis, and impression elements and scores are calculated for each phrase segmented from the input text using an impression dictionary in which impression elements and their scores are associated with phrases in advance. Generate a list of associated phrases, calculate the likelihood of each impression element from the entire phrase of the input text, calculate the objective impression by associating the likelihood as a score of each impression element, and convert it to the subjective impression. There is known a technique of calculating impression difference information by comparing a score of an impression element based on an objective impression with a score of the impression element based on an objective impression (for example, see Patent Document 1).

JP 2017-84015 A

In the world, new words that have not been used until now, such as neologisms and coined words, may become popular. For example, if new content with a unique title is released and the content becomes a hot topic among people, the unique title will become popular as a new word. However, with conventional techniques, it is often difficult to frequently update dictionaries according to fashion, and furthermore, it is often difficult to find out what kind of sentences should be registered as named entities in dictionaries. was not For this reason, the conventional technology may not be able to accurately extract named entities from documents.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and aims to provide an information processing apparatus, an information processing method, and a program capable of extracting a named entity from a document with high accuracy.

One aspect of the present invention includes a dividing unit that divides a sentence into character strings containing at least one character, and a score for each of the character strings divided by the dividing unit based on a plurality of queries input by a user. An information processing device comprising: a calculating unit for calculating; and an extracting unit for extracting a named entity from the sentence based on the score calculated by the calculating unit.

According to one aspect of the present invention, a named entity can be extracted from a document with high accuracy.

1 is a diagram showing an example of an information processing system 1 including an information processing device 100 according to a first embodiment; FIG. It is a figure showing an example of composition of information processor 100 in a 1st embodiment. 4 is a diagram showing an example of a search log 132; FIG. 4 is a flow chart showing a series of processes of the control unit 110 in the first embodiment; FIG. 4 is a diagram schematically showing a method of extracting a named entity from three-letter text; FIG. 4 is a diagram schematically showing a method of extracting a named entity from three-letter text; FIG. 4 is a diagram schematically showing a method of extracting a named entity from three-letter text; FIG. 4 is a diagram schematically showing a method of extracting a named entity from three-letter text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram schematically showing a method of extracting named entities from four-character text; FIG. 4 is a diagram for explaining a method of determining a pattern for score calculation; FIG. 4 is a diagram for explaining a method of determining a pattern for score calculation; FIG. 4 is a diagram for explaining a method of determining a pattern for score calculation;

Hereinafter, an information processing apparatus, an information processing method, and a program to which the present invention is applied will be described with reference to the drawings.

[Overview]
An information processing device is realized by one or more processors. The information processing device divides a text into character strings each including at least one character, and calculates a score for each character string based on a search log that is a history of multiple queries input by the user. Then, the information processing device extracts a named entity from the sentence based on the calculated score. This makes it possible to extract named entities from documents with high accuracy. As a result, for example, it is possible to obtain a distributed representation that accurately expresses the content of the sentence.

<First embodiment>
[overall structure]
FIG. 1 is a diagram showing an example of an information processing system 1 including an information processing device 100 according to the first embodiment. The information processing system 1 according to the first embodiment includes, for example, one or more terminal devices 10, a service providing device 20, and an information processing device 100. FIG. Some or all of these devices are connected to each other via a network NW. Note that some of these devices may be included in other devices as virtual devices. Alternatively, a virtual machine in which some or all of the functions of the information processing apparatus 100 are implemented by the functions of the service providing apparatus 20 may be used.

Each device shown in FIG. 1 transmits and receives various information via the network NW. The network NW includes, for example, wireless base stations, Wi-Fi access points, communication lines, providers, the Internet, and the like. Note that not all combinations of the devices shown in FIG. 1 need to be able to communicate with each other, and the network NW may partially include a local network.

The terminal device 10 is a terminal device including an input device, a display device, a communication device, a storage device, and an arithmetic device, such as a mobile phone such as a smart phone, a tablet terminal, and various personal computers. The communication device includes a network card such as a NIC (Network Interface Card), a wireless communication module, and the like. In the terminal device 10 , a UA (User Agent) such as a web browser or an application program is activated, and a request according to the user's input is sent to the service providing device 20 . Also, the terminal device 10 on which the UA is activated causes the display device to display various images based on the information acquired from the service providing device 20 .

The service providing device 20 is, for example, a web server that provides a web page to the terminal device 10 in response to a request from a web browser activated as a UA. The web pages may be, for example, web pages that constitute various websites such as shopping sites, auction sites, and flea market sites. The service providing apparatus 20 may also provide the terminal apparatus 10 with web pages that provide various services such as search sites, SNS (Social Networking Service), and mail services. Further, the service providing device 20 is an application server that provides the same services as various websites such as a sales site by providing content to the terminal device 10 in response to a request from an application activated as a UA. good too.

The information processing apparatus 100 acquires a search log from the service providing apparatus 20, and uses the search log to extract named entities from sentences. The named entity according to the present embodiment includes, for example, not only one word such as a noun, but also one phrase in which a noun and a noun are connected by another part of speech (for example, a particle), It contains one sentence including various parts of speech such as nouns, verbs, particles, and auxiliary verbs. In other words, any long sentence can be a named entity as long as it is a word used by humans as a unique entity.

[Configuration of information processing device]
FIG. 2 is a diagram showing an example of the configuration of the information processing apparatus 100 according to the first embodiment. As illustrated, the information processing apparatus 100 includes, for example, a communication unit 102, a control unit 110, and a storage unit .

The communication unit 102 includes, for example, a communication interface such as a NIC (Network Interface Card) and a DMA (Direct Memory Access) controller. The communication unit 102 communicates with the service providing apparatus 20 and other web servers via the network NW.

The control unit 110 includes, for example, an acquisition unit 112, a text division unit 114, a phrase score calculation unit 116, and a named entity extraction unit 118. The components of the control unit 110 are realized by executing a program stored in the storage unit 130 by a processor such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). Some or all of the components of the control unit 110 are implemented by hardware (circuitry) such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), and FPGA (Field-Programmable Gate Array). may be realized by cooperation of software and hardware.

The storage unit 130 is implemented by, for example, a HDD (Hard Disc Drive), flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), ROM (Read Only Memory), RAM (Random Access Memory), and the like. The storage unit 130 stores a search log 132 in addition to various programs such as firmware and application programs.

FIG. 3 is a diagram showing an example of the search log 132. As shown in FIG. As shown in the illustrated example, the search log 132 is historical information in which each query input by the user to the search engine and the number of times the query is input are associated with each aggregated period. The number of inputs may be, for example, the number of unique browser cookies. In this case, even if the same query is input many times via the same browser, the number of input times of the query is counted as one.

[Processing flow]
A series of processes performed by the control unit 110 according to the first embodiment will be described below using a flowchart. FIG. 4 is a flow chart showing a series of processes of the control unit 110 in the first embodiment. The processing of this flowchart may be performed repeatedly at a predetermined cycle, for example.

First, the acquisition unit 112 acquires text, which is one of contents, from the service providing device 20 via the communication unit 102 (S100). For example, text is text that users can search using a search engine. Articles posted on general sites (unofficial sites) operated by third parties unrelated to , organizations, celebrities, etc.

Next, the acquisition unit 112 acquires the search log 132 from the service providing device 20 via the communication unit 102 (S102). For example, the acquisition unit 112 acquires the search log 132 including the history of queries input by the user in the last several months.

Next, the text division unit 114 divides the text acquired by the acquisition unit 112 into one or more phrases (S104). A phrase according to the present embodiment may be a single character such as particles such as "wa", "ga", and "wo", or may be a character string including a plurality of characters. That is, in the description of the present embodiment, the term phrase is not used in the dictionary sense of a group of words, but rather as a single character, a single word, a single phrase, or a single sentence. , used in a slightly broader sense.

For example, if the number of characters included in the text is N, the text dividing unit 114 divides the text according to a combination pattern of 2<(N-1)> phrases. For example, text is split in one pattern if N=1, split in two patterns if N=2, split in four patterns if N=3, and split in four patterns if N=4. If so, it is divided into 8 different patterns.

Next, based on the search log 132, the phrase score calculation unit 116 calculates, for each phrase divided by the text division unit 114, an index value (hereinafter referred to as phrase score) is calculated (S106). For example, the phrase score calculator 116 calculates the phrase score based on Equation (1).

In the formula, S represents the phrase score, β represents the number of times a query was entered that matches the phrase to be scored (the number of searches using that query), and α is a real number greater than 1 (e.g., 10, etc.). ), and L represents the length of the phrase to be scored, that is, the number of characters included in the phrase.

For example, if the phrase for score calculation matches any of the queries included in the search log 132, that is, if the phrase for score calculation is input as a query one or more times, the phrase score calculation unit 116 increases the phrase score S of the target phrase as the number of query inputs β and the phrase length L increases; Decrease the phrase score S. Note that if the phrase for score calculation does not match any of the queries included in the search log 132, that is, if the phrase for score calculation is not input as a query, the phrase score calculation unit 116 Set the phrase score S of the phrase to 0.

Next, the named entity extractor 118 extracts named entities from the text based on the phrase score calculated for each phrase by the phrase score calculator 116 (S108). This completes the processing of this flowchart.

5 to 8 are diagrams schematically showing methods of extracting named entities from three-letter text. These figures schematically show that the three-letter text "ABC" is divided into one or more phrases. Therefore, the text is split in 4 (2 ² ) patterns.

For example, FIG. 5 shows a pattern 1 that divides one text "ABC" into a one-letter phrase "A", a one-letter phrase "B", and a one-letter phrase "C". represent. In the case of pattern 1, the phrase score calculation unit 116 calculates the phrase score S _A for the phrase “A”, the phrase score S _B for the phrase “B”, and the phrase score S B for the phrase “C”. calculates the phrase score _SC .

FIG. 6 shows pattern 2 for dividing a single text "ABC" into a two-letter phrase "AB" and a one-letter phrase "C". The phrase score calculator 116 calculates the phrase score _{S AB} for the phrase “AB” and the phrase score SC for the phrase “C”.

FIG. 7 shows a pattern 3 in which a single text "ABC" is divided into a one-letter phrase "A" and a two-letter phrase "BC". The phrase score calculator 116 calculates the phrase score S _A for the phrase “A” and the phrase score S _BC for the phrase “BC”.

FIG. 8 shows a pattern 4 in which one text "ABC" is used as it is as one phrase. The phrase score calculator 116 calculates the phrase score S _ABC for the phrase "ABC".

After calculating the phrase score S of each phrase for each pattern as described above, the phrase score calculator 116 calculates the sum of the phrase scores S for each pattern. In the pattern 1 illustrated in FIG. 5, the sum of the phrase scores S is (S _A +S _B +S _C ), and in the pattern 2 illustrated in FIG. 6, the sum of the phrase scores S is (S _AB +S _C ), In pattern 3 illustrated in FIG. 7, the sum of phrase scores S is (S _A +S _BC ), and in pattern 4 illustrated in FIG. 8, the sum of phrase scores S is (S _ABC ).

The named entity extraction unit 118 selects the pattern that maximizes the sum of the phrase scores S from these four patterns, and extracts the phrase represented by that pattern as a named entity. For example, when the sum S _ABC of the phrase scores S of pattern 4 is the maximum, the named entity extraction unit 118 extracts one phrase “ABC” as a named entity. Also, for example, when the sum of the phrase scores S of pattern 2 ( _{S AB} +SC ) is the maximum, the named entity extraction unit 118 extracts the phrase “AB” and the phrase “C” as named entities. .

9 to 16 are diagrams schematically showing methods of extracting named entities from four-character text. These figures schematically show that the four-letter text "ABCD" is divided into one or more phrases. Therefore, the text is split in 8 (2 ³ ) patterns.

For example, FIG. 9 shows a pattern 1-1 that divides one text "ABCD" into phrases for each character. FIG. 10 shows a pattern 1-2 for dividing one text "ABCD" into three phrases "A", "B" and "AB". FIG. 11 shows a pattern 2-1 for dividing one text "ABCD" into three phrases "AB", "C" and "D". FIG. 12 shows a pattern 2-2 for dividing one text "ABCD" into two phrases "AB" and "CD". FIG. 13 shows a pattern 3-1 for dividing one text "ABCD" into three phrases "A", "BC" and "D". FIG. 14 shows a pattern 3-2 for dividing one text "ABCD" into two phrases "A" and "BCD". FIG. 15 shows a pattern 4-1 for dividing one text "ABCD" into two phrases "ABC" and "D". FIG. 16 shows a pattern 4-2 in which one text "ABCD" is used as it is as one phrase. In the same manner as described above, after calculating the phrase score S of each phrase for each pattern, the phrase score calculator 116 calculates the sum of the phrase scores S for each pattern. Then, the named entity extracting unit 118 selects the pattern that maximizes the sum of the phrase scores S from these eight patterns, and extracts the phrase represented by that pattern as a named entity.

According to the first embodiment described above, the text is divided into one or a plurality of phrases according to a number combination pattern based on the number of characters N of the text, and the phrase score is calculated for each divided phrase in each pattern. Then, the sum of phrase scores is calculated for each pattern, and the phrase with the maximum calculated sum is extracted as a named entity. This makes it possible to extract named entities from documents with high accuracy.

Traditionally, named entities have been extracted from text using dictionaries in which named entities have been registered in advance. be. However, it is practically difficult to update the dictionary on a daily basis. In addition, niche terms that have begun to be used as new words in some communities are difficult to register as named entities in dictionaries.

Therefore, for example, if the title of the new content is a sentence such as "Official guidebook for XX, how to walk from ◇◇ to △△", a dictionary is applied to the text containing this title to extract the named entity. In this case, multiple words such as ``○○○'', ``official'', ``guidebook'', ``◇◇'', ``△△'', and ``way of walking'' are extracted as named entities, and instead of extracting ``○ 〇〇Official guidebook ・How to walk from ◇◇ to △△” will not be extracted as a named entity.

On the other hand, considering the nature of the query, if the title of some new content is a long title like a sentence, the user will copy the character string representing the title from the official website or a third-party website, etc. It is assumed that the copied character string is pasted in the input field of the search site. In this case, queries that are the same word for word as the title of the content, which is the named entity, are collected as the search log 132 . In particular, a very recent search log 132, such as the last few months, is likely to include new words that are currently popular as queries. Therefore, in this embodiment, by comparing the phrases obtained by dividing the text with the query in the search log 132, the named entity can be accurately extracted from the text even if the named entity is long or brand new. be able to.

It is also conceivable to extract a part of text surrounded by punctuation marks (descriptive symbols) such as brackets, apostrophes, and primes as named entities. However, in this method, there are cases where a person's dialogue or quoted sentence is extracted as a named entity, and it is not possible to distinguish whether it is a named entity used as a noun or simply a dialogue or quoted sentence. Can not. In addition, although character names and personal names in content are named entities, they cannot be extracted from text because they are not usually enclosed in parentheses.

On the other hand, in the present embodiment, it is possible to extract named entities without relying on punctuation. In addition, movie and book titles often have subtitles, and the subtitles may be surrounded by punctuation marks. Even if a named entity is surrounded by punctuation, if the user inputs a named entity surrounded by punctuation as a query, the method of the present embodiment also extracts named entities that include the punctuation. be able to.

In addition, when simply comparing the phrases of the text with the queries in the search log 132, relatively short phrases such as "wa", "o", "desu", and "masu" tend to appear in the text. Phrase coincidentally matches the query, and the phrase score S tends to be large. On the other hand, in the present embodiment, the phrase score S is the product of the number of inputs β and an arbitrary cardinal number α, where the exponent is the phrase length L. is large, the phrase score S is increased, and even for a phrase with a large number of inputs β, the phrase score S can be decreased if the phrase length L is small. As a result, it is possible to accurately extract phrases and sentences in which a plurality of nouns are connected by particles and the like as a single named entity while suppressing the extraction of particles and the like as named entities.

<Second embodiment>
A second embodiment will be described below. In the above-described first embodiment, the sum of phrase scores is calculated for all patterns that are combinations of phrases, and the phrase with the maximum sum is extracted as a unique expression. In contrast, in the second embodiment, instead of calculating the sum of phrase scores for all patterns, the sum of phrase scores is calculated after rationally determining patterns to be verified. It differs from the embodiment. In the following, differences from the first embodiment will be mainly described, and descriptions of common points with the first embodiment will be omitted. In addition, in description of 2nd Embodiment, the same code|symbol is attached|subjected and demonstrated about the same part as 1st Embodiment.

17 to 19 are diagrams for explaining a method of determining patterns for score calculation. For example, when a four-character text "ABCD" is given, the text dividing unit 114 according to the second embodiment divides the first character "A" at the beginning of the text, followed by the second character "A", as illustrated in FIG. Split between the two letters 'B' to generate the phrase 'A' and the phrase 'B', and do not split between these letters to generate the two letter phrase 'AB'. . The phrase score calculation unit 116 according to the second embodiment uses the search log 132 to calculate the sum (S _A +S _B ) of the phrase score S _A of the phrase “A” and the phrase score S _B of the phrase “B”. At the same time, the phrase score SAB of the phrase " _AB " is calculated. The text dividing unit 114 compares these and eliminates the pattern derived from the pattern with the smaller score from the pattern candidates to be verified next. The first character 'A' is an example of a 'first character string', the second character 'B' is an example of a 'second character string', and the phrase 'AB' is an example of a 'third character string'. is an example.

In the example of FIG. 17, S _AB is larger than (S _A +S _B ). In this case, there are four types of combinations of phrases including the third character "C" following the second character: pattern 1, pattern 2, pattern 3, and pattern 4 described above. Of these four patterns, at least pattern 1 and pattern 2 are the same as before the inclusion of the third character "C". For example, the sum of the phrase scores for pattern 1 is (S _A + S _B + S _C ), and the sum of the phrase _scores for pattern 2 is (S _AB + _{S C} ). If the result is that _SAB is larger, the magnitude relation of the sum of the phrase scores of pattern 1 and pattern 2 does not change. Therefore, the phrase score calculation unit 116 does not calculate the score for pattern 1, which divides the text between the first character "A" at the beginning of the text and the second character "B" following it.

Next, the text dividing unit 114 compares the scores of the remaining three patterns 3, as illustrated in FIG. In the example of FIG. 18, pattern 4 has the highest phrase score S _ABC . Therefore, the phrase score calculation unit 116 determines the pattern 3-1 that divides the first letter "A" of the text and the following three-letter combination "BCD" and the first two-letter combination "AB" of the text. and the following two-letter combination "CD", and pattern 4-1 splitting between the first three-letter combination "ABC" of the text followed by the single letter "D". , the score is calculated only for a total of 4 patterns, pattern 4-2, in which the text is not divided into a single phrase. Thus, in our method, when considering the last k-th character of a string, we compare k patterns.

Next, the phrase score calculator 116 calculates the sum of the phrase scores of the four patterns (3-1, 2-2, 4-1, -2). For example, when the sum (S _ABCD +S _D ) of the phrase score of pattern 4-1 is the largest, the named entity extraction unit 118 extracts the phrase "ABC" and the phrase "D" as named entities. In this way, the characters are sequentially combined from the first character of the text, and the combination pattern is selected according to the score of each combination candidate, whereby the optimum combination can be searched for.

Note that when the phrase score of a certain pattern is 0, derived patterns obtained by adding one character to that pattern need not be taken into consideration thereafter. A phrase score of 0 means that the number of input times β of queries that match the phrase is zero. In other words, it means that no user has ever input the phrase represented by the pattern as a query, and the probability that the phrase of the pattern is a named entity is extremely low.

According to the second embodiment described above, the characters are combined in order from the beginning of the text, the phrase score of each pattern is compared at that point, and the pattern with the smaller phrase score is excluded from subsequent processing.

For example, in content such as movies, dramas, and animations, some titles are proper nouns in which particles such as "no" and "to" are sandwiched between a certain word A and a certain word B. Specifically, the titles are "XX and □□" and "XX's △△". When the phrase score is obtained for all patterns of such content titles, the score is calculated for phrases in which particles such as "no" and "to" appear at the beginning of nouns. However, in the real world, it is extremely rare for a particle to appear at the beginning of a phrase, and it can be assumed that the phrase itself does not exist in the world. Therefore, by combining the characters in order from the beginning of the text, comparing the phrase scores of each pattern at that point, and excluding patterns with lower phrase scores from subsequent processing, we can efficiently extract proper nouns. be able to.

As in the first embodiment described above, when all patterns are divided, the number of score calculations increases to 2 ^N−1 , where N is the number of characters included in the text. On the other hand, in the second embodiment, since the number of patterns is reduced in consideration of the continuity of characters, the number of score calculations can be suppressed to ^N2 times.

<Hardware configuration>
The information processing apparatus 100 of the embodiment described above is realized by, for example, a hardware configuration as shown in FIG. 19 . FIG. 19 is a diagram illustrating an example of the hardware configuration of the information processing apparatus 100 according to the embodiment.

Information processing apparatus 100 includes NIC 100-1, CPU 100-2, RAM 100-3, ROM 100-4, secondary storage device 100-5 such as flash memory and HDD, and drive device 100-6 connected to an internal bus or dedicated communication line. are interconnected by A portable storage medium such as an optical disc is mounted in the drive device 100-6. A program stored in the secondary storage device 100-5 or a portable storage medium attached to the drive device 100-6 is developed in the RAM 100-3 by a DMA controller (not shown) or the like, and executed by the CPU 100-2. Thus, the control unit 110 is realized. The program referred to by control unit 110 may be downloaded from another device via network NW.

As described above, the mode for carrying out the present invention has been described using the embodiments, but the present invention is not limited to such embodiments at all, and various modifications and replacements can be made without departing from the scope of the present invention. can be added.

Reference Signs List 1 information processing system 10 terminal device 20 service providing device 100 information processing device 102 communication unit 110 control unit 112 acquisition unit 114 text division unit 116 phrase score calculation unit , 118... Named entity extraction unit, 130... Storage unit

Claims (5)

a dividing unit that divides the text into character strings each including at least one character in a number of patterns corresponding to the number of characters included in the text ;
Based on a plurality of queries input by the user, for each pattern, calculate the score of each character string divided from the sentence by the dividing unit, and calculate the sum of all the scores of the character strings for each pattern. a calculating unit that calculates;
A pattern that maximizes the sum of the scores calculated by the calculator is selected from the patterns corresponding to the number of characters, and the character string in the pattern that maximizes the sum of the scores is extracted as a unique expression. and an extractor for
The calculation unit increases the score if the length of the character string is large even for the character string that matches the query with a small number of inputs, and the character string that matches the query with a large number of inputs. reduce the score if the length of the string is small, even if
Information processing equipment. The calculation unit calculates the score based on the number of times the query is input that matches the character string and the length of the character string.
The information processing device according to claim 1 . The dividing unit divides the sentence at a first position,
The calculation unit calculates a score of a first character string containing at least one character appearing immediately before the first position and a score of a second character string containing at least one character appearing immediately after the first position. , calculating a score of a third character string that combines the first character string and the second character string,
The extraction unit compares the sum of the score of the first character string and the score of the second character string with the score of the third character string, and selects the character string with the smaller score as the unique expression Exclude from the extraction target of
The information processing apparatus according to claim 1 or 2 . the computer
dividing the text into character strings containing at least one character in a number of patterns corresponding to the number of characters included in the text ;
Based on a plurality of queries input by the user, for each pattern, calculate the score of each character string divided from the sentence ,
calculating the sum of all the scores of the character strings for each of the patterns;
Selecting a pattern that maximizes the sum of the calculated scores from the number of patterns corresponding to the number of characters ,
extracting the character string in the pattern that maximizes the sum of the scores as a named entity ;
Even if the character string matches the query with a small number of inputs, if the length of the character string is large, the score is increased,
Even if the character string matches the query with a large number of inputs, if the length of the character string is small, the score is reduced.
Information processing methods. to the computer,
A process of dividing the text into character strings containing at least one character in a number of patterns corresponding to the number of characters included in the text ;
A process of calculating the score of each character string divided from the sentence for each pattern based on a plurality of queries input by the user;
A process of calculating the sum of all the scores of the character strings for each of the patterns;
A process of selecting a pattern that maximizes the sum of the calculated scores from the number of patterns corresponding to the number of characters ;
A process of extracting the character string in the pattern that maximizes the sum of the scores as a named entity;
A process of increasing the score if the length of the character string is large even for the character string that matches the query with a small number of inputs;
A process of reducing the score if the length of the character string is small even for the character string that matches the query with the large number of inputs;
program to run the

Images