JPH08180058A - Compound word retrieval device - Google Patents

Abstract

PURPOSE: To provide a word retrieval device for outputting the plural results of compound word retrieval in an order from the most certain one by a prescribed procedure when more than two words for constituting a compound word in a dictionary match with the words provided in an input sentence. CONSTITUTION: A word retrieval part 105 divides the words from an inputted sentence and further, obtains a compound word number corresponding to the compound word provided with the words. When the same compound number is provided in more than two words, a compound word retrieval part 107 outputs it to a compound word tentative storage part 108 as a compound word candidate. A compound word aligning part 109 rearranges the compound words stored in the compound word tentative storage part 108 in the order of certainty by the prescribed procedure and an output part 110 outputs the compound words and compound word translations in the order.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compound word search apparatus for searching compound words composed of a plurality of words in a character string.

[0002]

[Prior Art] When translating an English sentence into a Japanese sentence,
Proper translation of collocation has become an important issue. here,
A collocation is an idiom or idiom composed of two or more words. The following is an example of a compound word search device. The first conventional technique is disclosed in Japanese Patent Laid-Open No. 63-89975.
It was published in the issue. This is to prioritize a plurality of retrieved collocations according to a certain rule when they share the same word. The second conventional technique is disclosed in Japanese Patent Laid-Open No. 1-266671. This aims at speeding up by narrowing down the candidates at the time of search by classifying the collocations by paying attention to the leading letters and pronunciations of the words forming the collocations.

[0003]

By the way, in the above two collocation retrieval devices, in order to be able to retrieve a certain collocation, all the words constituting the collocation registered in the dictionary are included in the input sentence. Need to be. So, for example, if the lexical word “in one's idiomatic fashion” is registered in the dictionary,
When it says "in his idiomatic fashion,""one's
This compound cannot be searched because "" and "his" do not match. Furthermore, due to the nature of the example sentence, bilingual example sentences in commonly published dictionaries contain only a part of sentence components such as compound words. The sentence is not a unit, but a sentence.Therefore, when performing a collocation search by referring to these bilingual example sentences, the above-mentioned device is almost useless and remarkably inconvenient. Become.

Further, in order to perform a collocation search by matching partial words with reference to a bilingual example sentence, it is indispensable to provide a method for selecting an optimum candidate at high speed, but such a method is currently available. Absent. In view of the above-mentioned problems, the present invention can efficiently use a compound word as a translated word when two or more words forming a compound word registered in a dictionary match a word included in an input sentence. The purpose is to provide a device.

[0005]

In order to solve the above problems, in the invention according to claim 1, there is provided a compound word search device for searching a compound word composed of a plurality of words in an input sentence, for inputting a sentence. The input means, the sentence temporary storage means for temporarily storing the sentence input by the input means, the basic word of the headword, and the identifier of the compound word including the headword and its inflected word are stored in advance in association with each other. Stored in the sentence temporary storage means, and a word storage means in which the word storage means, an identifier for identifying each collocation, a collocation specified by the identifier, and a translation thereof are stored in advance in association with each other. A sentence that is divided into words, and when the divided words are not in the basic form, the word is divided into basic words, and the word obtained by the word dividing means matches the headword stored in the word storage means. If the word detection means detects the corresponding collocation identifier, when the word detection means detects the collocation identifier, the identifier, the word that remains divided and the word that is the basic form or the word that remains divided And, when not detected, a word that is still divided and a word that is in the basic form or a word that is still divided and is stored in the divided order as a set, and the word temporary storage means stores the word in the divided order. A collocation search unit that searches the collocation storage unit for collocations and collocations identified by the identities when a plurality of identities match each other, and all collocations and collocations searched by the collocation search unit. Is stored for each collocation, a collocation temporary storage means, a ranking determination means for determining an optimal ranking of the collocations stored in the collocation temporary storage means in a predetermined procedure, and the ranking determination means. It is characterized by in and output means for outputting the collocation and collocation translation to higher determined ranking order.

In the invention according to claim 2, the rank determining means determines the number of exchanges between the order of words stored in the word temporary storage means and the order of words stored in the compound word temporary storage means. It is characterized by including a word replacement number counting unit that counts and increases the rank of collocations having a small number of times. In the invention according to claim 3, the rank determining means forms a compound word stored in the word temporary storage means between a first word and a last word forming a compound word stored in the compound word temporary storage means. It is characterized by including a word insertion number counting unit that counts how many words that are not inserted and increases the rank of a collocation having a small number.

In the invention according to claim 4, the rank determining means includes a word forming a compound word stored in the word temporary storage means and a word forming a compound word stored in the word temporary storage means. It is characterized by including a word matching rate calculation unit that calculates the matching rate of and increases the rank of collocations having a high matching rate. In a fifth aspect of the present invention, the rank determining means counts the number of words stored in the word temporary storage means that match a word forming one compound word stored in the multiple word temporary storage means. It is characterized by including a word matching number counting unit for increasing the rank of a collocation having a large number.

In the invention according to claim 6, the rank determining means determines the number of exchanges of the order of the words stored in the word temporary storage means and the order of the words stored in the compound word temporary storage means. When there are collocations having the same rank in the word replacement number counting unit that counts and increases the rank of collocations having a small number of times, the collocations stored in the collocation temporary storage means are configured. A word insertion number counting unit that counts how many words that do not form a collocation stored in the word temporary storage means between the first and last words are inserted, and increases the rank of collocations with a small number thereof. When the word insertion number counting unit has collocations having the same rank, the vocabulary forming the collocation stored in the word temporary storage means and the collocation stored in the collocation temporary storage means are constructed. If there is a word matching rate calculation unit that calculates the matching rate with a word and increases the rank of a word having a high matching rate, and the word matching rate calculation unit has a word with the same rank, the word temporary storage unit A word match number counting unit for counting the number of words stored in the word temporary storage unit that match a word constituting one collocation stored in the word and increasing the rank of the collocation having a large number. It has a feature.

[0009]

With the above construction, in the invention of claim 1, the word dividing means divides the sentence input from the input means stored in the sentence temporary storage means into words, and when the divided words are not in the basic form (inflection) When it is a word in the form), it is converted to the word in the basic form. The word detecting means detects an identifier of a corresponding collocation word when the obtained word matches a heading word stored in the word storing means in which only the basic word is used as a heading word in order to speed up the search. , When a collocation identifier is detected, the identifier and the word that has been divided (words in the input sentence) or a basic word in addition to that are used as a set, and when the collocation identifier is not detected, the word that remains divided or In addition to this, the basic form words are temporarily stored in the word temporary storage means in a divided order (from the beginning of the input sentence). The collocation search means determines that the input sentence includes collocations when the identifiers stored in the word temporary storage means match in a plurality of sets, and the collocations and collocations identified by the matching identifiers are collocated. The data is searched from the storage means and sequentially stored in the multiple word temporary storage means for each multiple word. The rank determining means determines an optimum rank (suitable for translation) as a collocation in an input sentence by a predetermined procedure. The output means outputs (for example, displays on the display screen) the collocations and collocations in descending order of rank.

In the invention of claim 2, in addition to the operation of the invention of claim 1, the word replacement number counting unit of the order determining means is:
In order to determine the optimum rank as a complex word in the input sentence, the word order of the complex word stored in the complex word temporary storage means and the word order stored in the word temporary storage means (the word sequence of the complex word of the complex word storage means and the complex word of the input sentence Calculate the number of replacements (word order),
Increase the word order of collocations that have a low frequency.

According to the invention of claim 3, in addition to the operation of the invention of claim 1, the word insertion number counting section of the rank determining means is:
In order to determine the optimum rank as a collocation in the input sentence, what is the word that does not form a collocation stored in the word temporary storage means between the first and last words that form the collocation stored in the collocation temporary storage means. Count whether words have been inserted, and increase the rank of collocations with smaller numbers.

In the invention of claim 4, in addition to the operation of the invention of claim 1, the word concordance rate calculation unit of the rank determining means includes:
In order to determine the optimal rank as a collocation in the input sentence, the matching rate with the words forming the collocation stored in the word temporary storage means is calculated, and the collocation with a high matching rate is ranked higher. In the invention of claim 5, in addition to the operation of the invention of claim 1, the word coincidence number counting unit of the rank determining means stores the word in the temporary word storage means in order to determine the optimum rank as a complex word in the input sentence. Counting the number of words stored in the word temporary storage means that match the words forming one compound word,
Increase the rank of collocations with a large number.

In the sixth aspect of the invention, in addition to the operation of the first aspect of the invention, when there is the same rank in determining an appropriate collocation order of the deciding means, the word replacement number counting section and the word insertion number The counting unit, the word matching rate calculating unit, and the word matching number counting unit are made to function to determine the optimum collocation order.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a compound word search device according to the present invention will be described based on embodiments. FIG. 1 is a block diagram of an embodiment of a compound word search device according to the present invention. This compound word search device includes an input unit 101, a sentence temporary storage unit 102, and a word dictionary 103.
The multilingual dictionary 104, the word search unit 105, the word temporary storage unit 106, the multiword search unit 107, the multiword temporary storage unit 108, the multiword alignment unit 109, and the output unit 110.

The input unit 101 receives a sentence input from the operator, stores the received sentence in the sentence temporary storage unit 102, and gives a start instruction to the word search unit 105. The sentence temporary storage unit 102 temporarily stores the sentence input from the input unit 101. As shown in FIG. 2, the word dictionary 103 has registered therein a word heading and a compound word number which is an identifier of a compound word including the word. Note that only basic-form (original-form) words are registered as word headings, and collocation numbers are not only collocations that include basic-form words, but collocations that include conjugated words. Here, the collocation means an idiom or an idiom composed of two or more words.

As shown in FIG. 3, the collocation dictionary 104 registers collocation numbers which are identifiers of collocations, collocations and their translations in the order of collocation numbers. The word search unit 105 includes a word division unit and a word detection unit. The word division unit is the input unit 10
When the activation instruction is received from 1, the sentence stored in the sentence temporary storage unit 102 is read and divided into words according to a predetermined procedure. Specifically, in the case of Western languages, it is divided by the space provided between characters, and in the case of Japanese, Chinese, etc., it is divided by morphological analysis, the longest matching method, and the like.

Further, the word division unit incorporates a morphological analysis dictionary (not shown), and converts the divided words into basic forms based on this dictionary. When it is transformed into the basic form, the word that is still divided and the word that is converted into the basic form are notified to the word detection unit when the word that is still divided is the basic form. Furthermore, when the reading of all the sentences stored in the sentence temporary storage unit 102 is completed, the word detection unit is notified of the completion of reading.

When the word detection unit receives a word from the word division unit, the word dictionary 103 uses the basic word as a key.
When the word is registered in the word heading, the corresponding collocation number is extracted and the word temporary storage unit 1
In 06, a word (a word that is still divided and a basic form word or a word that is still divided) and a collocation number are stored as a set and stored. When the word is not registered in the word heading, only that word (a word that is still divided and a basic form word or a word that is still divided) is stored as a set. When the notification of the completion of reading is received from the word division unit and the search of the word dictionary 103 is completed, the collocation search unit 1
A start instruction is given to 07.

The word temporary storage unit 106 stores a word and a collocation number or a word in order from the beginning of the input sentence. Upon receiving the activation instruction from the word search unit 105, the compound word search unit 107 reads the compound word numbers of each set stored in the word temporary storage unit 106 and detects the compound word numbers commonly included in the plurality of sets. Then, the collocations and collocations that match the detected collocation number are extracted from the collocation dictionary 104, and all are stored in the collocation temporary storage unit 108 in the ascending order of the collocation numbers, with the rank being "1". When the retrieval of the complex word or the like from the complex word dictionary 104 is completed, a start instruction is given to the complex word alignment unit 109.

As shown in FIG. 4, the compound word temporary storage unit 108 stores a rank 401, a compound number 402, a compound word 403, a compound translation 404, a word replacement number 405, an inserted word number 406, a dictionary matching rate 407, and a word matching number 408. It consists of a collocation table 410 having storage fields for each. The order column 401 is rewritten by the complex word alignment unit 109 after “1” is set as an initial value by the complex word search unit 107. Collocation number 40
The contents of the fields of 2, the complex word 403, and the complex word translation 404 are written by the complex word search unit 107. Specifically, for example, when the sentence shown in FIG. 5 is input from the input unit 101,
The collocation table 410 includes a rank 401, a collocation number 402, a collocation 403, and a collocation translation 404 in the collocation search unit 107.
The search result is written in each column of, and the word replacement number 405, the insertion word number 406, the dictionary matching rate 407, and the word matching number 40.
Each column of 8 is left blank. Here, the ranking is
It means the order of the most appropriate (most suitable) collocations made up of the words contained in the input sentence.

The complex word alignment unit 109 includes a replacement number counting unit, an inserted word number counting unit, a dictionary matching rate calculating unit, a word matching number counting unit, and a counter unit. Upon receiving the activation instruction from the complex word search unit 107, the replacement number counting unit sets the value of the counter M of the counter unit to the initial value “1”. It is determined whether or not there is the M-th compound word stored in the compound word temporary storage unit 108, and when there is, the compound word is extracted, and the word stored in the word temporary memory unit 106 that matches the word forming the compound word. Mark. At this time, the compound word temporary storage unit 1
A mark is added so that the order of the words forming the collocation stored in 08 can be determined. The number of words (word replacement number) in which the appearance order of the marks added to the words (word sets are stored in order from the beginning of the sentence in the word temporary storage unit 106) is counted. The counted word replacement number is stored in association with the collocation stored in the collocation temporary storage unit 108, the mark added to the word stored in the word temporary storage unit 106 is erased, and the counter M of the counter unit is also stored. The value of is incremented by "1". When it is determined that there is no M-th compound word, the compound words stored in the compound word temporary storage unit 108 are rearranged in order such that the compound word with the smaller word replacement number comes first. However, the rearrangement is performed by the compound word temporary storage unit 10.
Applies between 8 collocated collocations. When the rearrangement of collocations is completed, an activation instruction is given to the insertion word number counting unit.

Upon receiving the activation instruction from the word replacement number counting section, the inserted word number counting section sets the value of the counter M of the counter section to the initial value "1". It is determined whether or not there is the M-th compound word stored in the compound word temporary storage unit 108, and when there is, the compound word is taken out, and the word stored in the word temporary memory unit 106 that matches the word forming the compound word is determined. Add a mark.

Words stored in the word temporary storage unit 106 (limited to words that have been divided) are sequentially arranged from the beginning.
Number of unmarked words between the first marked word and the last marked word (number of inserted words)
Count. Specifically, the number of sets in which no mark is added to the word is counted. The number of inserted words counted is stored in the compound word temporary storage unit 108.
The word added to the word is stored in the word temporary storage unit 106, the mark added to the word is erased, and the value of the counter M of the counter unit is incremented by “1”.

When it is judged that there is no Mth compound word,
The collocations stored in the collocation temporary storage unit 108 are rearranged in order such that the collocation having a smaller number of word insertions comes first. However, the rearrangement is applied between the collocations having the same rank in the collocation temporary storage unit 108. When the rearrangement of collocations is completed, a start instruction is given to the dictionary matching rate calculation unit. When receiving the activation instruction from the inserted word number counting unit, the dictionary matching rate calculation unit sets the value of the counter M of the counter unit to the initial value “1”. It is determined whether or not there is the M-th compound word stored in the compound word temporary storage unit 108, and when there is, the compound word is taken out, and the word stored in the word temporary memory unit 106 that matches the word forming the compound word is determined. Add a mark.

A value (dictionary concordance rate) is calculated by dividing the number of words to which the mark is added in the word temporary storage unit 106 by the number of words constituting the compound word in the compound word temporary storage unit 108. The calculated dictionary matching rate is stored in association with the multiple words stored in the multiple word temporary storage unit 108, and the mark added to the word stored in the temporary word storage unit 106 is deleted. The value of is incremented by "1". When it is determined that there is no Mth collocation, the collocations stored in the collocation temporary storage unit 108 are rearranged so that the collocation having a high dictionary matching rate comes first. However, the rearrangement is applied between the collocations having the same rank in the collocation temporary storage unit 108. When the rearrangement of collocations is completed, a start instruction is given to the word matching number counting unit.

When receiving the activation instruction from the dictionary matching rate calculation unit, the word matching number counting unit sets the value of the counter M of the counter unit to the initial value "1". It is determined whether or not there is the M-th compound word stored in the compound word temporary storage unit 108, and when there is, the compound word is extracted, and the word stored in the word temporary memory unit 106 that matches the word forming the compound word. Mark. Count the number of words with a mark (the number of matching words). Specifically, the number of sets in which a mark is added to a word is counted. The counted number of word matches is stored corresponding to the collocation stored in the collocation temporary storage unit 108, and the word temporary storage unit 10 is stored.
The mark added to the word stored in 6 is erased, and at the same time, the value of the counter M of the counter unit is incremented by "1". When it is determined that there is no Mth compound word, the compound words stored in the compound word temporary storage unit 108 are rearranged in order such that the compound word having the largest number of word matches is first. However, the rearrangement is applied between the collocations having the same rank in the collocation temporary storage unit 108.

When the rearrangement of the complex words is completed, the output unit 11
A start instruction is given to 0. For example, the collocation unit 109
The collocation table 410 shown in FIG.
8 is stored, the word replacement number 405, the insertion word number 406, and the dictionary matching rate 4 are finally obtained as shown in FIG.
07 and a word match number 408 are filled in with numerical values, and the collocation table 6 in which collocations are rearranged from the highest ranking 401
Create 01.

When the rearrangement by each part of the continuous word alignment unit 109 is described by a continuous word number, the word replacement number counting unit gives 1-2-3-4-5-7-6, and the insertion word number counting unit 2-4-4. 5-7-1-3-6, the dictionary matching rate calculation unit becomes 2-4-7-5-1-3-6, and the word matching number counting unit 4-2 as shown in FIG.
-7-5-1-3-6.

Upon receiving the activation instruction from the complex word alignment unit 109, the output unit 110 sequentially outputs the complex words stored in the complex word temporary storage unit 108 from the complex word having the highest rank. Next, the operation of this embodiment will be described with reference to the flowcharts shown in FIGS. When the input unit 101 receives an input of a sentence, the input unit 101 temporarily stores the sentence in the sentence temporary storage unit 102 (S7).
02).

When the word search unit 105 receives the activation instruction from the input unit 101, it retrieves a word from the sentence stored in the sentence temporary storage unit 102, and a word matching the basic word is found in the word dictionary 103. (S706). When it is found in a heading word, the word (a word that is still divided and a basic word or a word that is still divided) and a collocation number of all collocations corresponding to the heading word extracted from the word dictionary 103 are set as a set. The word is temporarily stored in the word temporary storage unit 106 (S708). If the word is not a headword, only that word is stored in the word temporary storage unit 106 as a set (S710). It is determined whether there are any words remaining in the sentence temporary storage unit 102 (S71
2) If yes, return to S704, and if no, S
Move to 714.

In S714, the complex word search unit 107
A collocation number commonly included in a plurality of sets stored in the word temporary storage unit 106 is searched, and a collocation and a collocation identified by the searched collocation number are retrieved from the collocation dictionary 104,
It is stored in the collocation table of the collocation temporary storage unit 108 (S
714). Next, the word alignment unit 109 sets the value of the counter M to "1" (S802). Compound word temporary storage unit 10
The presence or absence of the Mth compound word stored in 8 is determined (S80
4), if there is, the Mth compound word is used as the compound word temporary storage unit 10
It is read from 8 (S806). The word temporary storage unit 106 that matches the marks with the sequence order of the read collocation words.
It is added to the word stored in (S808). The number of word replacements whose order is replaced with the mark added to the word temporary storage unit 106 is counted (S810). The counted word replacement number is stored in the compound word table of the compound word temporary storage unit 108 (S812). The mark in the word temporary storage unit 106 is erased (S
814), the value of the counter M is incremented by "1" (S816), and the process returns to S802.

When there is no M-th compound word in the compound word temporary storage unit 108, the compound words are ranked in order from the smallest word replacement number stored in the compound word table and the compound words are rearranged (S81).
8). It is determined whether there are collocations having the same rank (S82).
0), if there is, to S902, if not, to S112
Move to 0 respectively. S902 to S906 are the above S802.
-S806 is the same as that of S806, and the description thereof is omitted.

A mark is added to the word stored in the word temporary storage unit 106 that matches the word included in the read complex word (S908). The number of inserted words inserted between the marked words is counted (S910). The counted number of inserted words is stored in the compound word table of the compound word temporary storage unit 108 (S912). S914 and S916 are the above S8.
14 and S816, the description thereof will be omitted.

When there is no M-th compound word in the compound word temporary storage unit 108, the compound words are ranked in order from the smallest number of inserted words stored in the compound word table, and the compound words are rearranged (S91).
8). It is determined whether or not there is a compound word having the same rank (S920), and if there is, S1002 which is the same as S902 to S908.
The processes of S1008 to S1008 are performed, and if not, the process proceeds to S1120.

The dictionary matching rate is calculated with the number of words included in the collocation as the denominator and the number of words with the mark stored in the word temporary storage unit 106 as the numerator (S101).
0). The calculated dictionary matching rate is stored in the collocation table 410 of the collocation temporary storage unit 108 (S1012), and the above S9 is performed.
14, the same processing of S1014 and S1016 as S916 is performed. When there is no Mth compound word in the compound word temporary storage unit 108, the compound words are ranked in descending order of the dictionary matching rate and the compound words are rearranged (S1018). It is determined whether or not there are collocations of the same rank (S1020).
The same processes of S1102 to S1108 as those of 2 to S1008 are performed, and if not, the process proceeds to S1120.

The number of marked words stored in the word temporary storage unit 106 is counted as the number of word matches (S1).
110). The number of word matches is stored in the multiple word table of the multiple word temporary storage unit 106 (S1112).
The same processes of S1114 and S1116 as 1016 are performed. When there is no M-th compound word in the compound word temporary storage unit 108, the compound words are ranked in descending order of the number of word matches and the compound words are rearranged (S1118).

In step S1120, the output unit 110 outputs the complex word and its complex word translation stored in the complex word table 601 in descending order of rank. Next, a device for creating the word dictionary 103 used in this embodiment will be described. The word dictionary creating device has substantially the same configuration as the word searching unit 105 described in the present embodiment.

FIG. 12 shows an example of the collocation dictionary 104. As described above, the collocation dictionary 104 stores collocation numbers for identifying collocations, collocations and collocations in association with each other in advance. The word dictionary creating device reads from the first word "before" that forms the compound word identified by the compound number "11", performs morphological analysis, recognizes that "before" is a basic word, and recognizes the word dictionary. The heading word “before” and the collocation number “11” are stored in 103 in association with each other. Hereinafter, the same applies to the word "I". The word "was" is read out, and as a result of the morphological analysis, the word "b" is read.
Recognizing that “e” is a basic word, the word dictionary 103
The heading word “be” and the collocation number “11” are stored in correspondence with each other. The same process is repeated to read the word “be” having the collocation number 12, and as a result of the morphological analysis, recognize that it is the basic word and store the collocation number “12”, but the same word “be” Since the word is already stored as the index word in the word dictionary 103, the word number “12” is stored after the word number “11” already stored.
As a result of such processing, the word dictionary 103 becomes as shown in FIG.

Although the present invention has been described above based on the embodiments, it goes without saying that the present invention is not limited to the above embodiments. That is, in the above-mentioned embodiment, English-Japanese collocation translation is taken as an example, but any combination of languages can be dealt with by changing the memorized contents of the collocation dictionary or word dictionary to another language.

[0040]

As described above, according to the invention of claim 1, it is possible to use a parallel translation example sentence such as an existing dictionary in a collocation dictionary, and in a sentence in which two or more words forming the collocation are input. When included, the collocations and collocations are output in an appropriate order that can be used for translation, so that the collocations in the sentence can be used quickly and completely.

According to the second aspect of the invention, in addition to the effect of the first aspect of the invention, since the compound word in which the word order in the input sentence and the word order of the compound word in the compound word dictionary are less interchanged is preferentially output, More appropriate collocation can be used efficiently. According to the invention of claim 3, in addition to the effect of the invention of claim 1, further, the compound word inserted between the words composing the compound word in the input sentence and having less composing words is preferentially output. Therefore, more appropriate collocation can be efficiently used.

According to the invention of claim 4, in addition to the effect of the invention of claim 1, further, since the collocations having a high matching rate of the words constituting the collocation are preferentially output, a more appropriate collocation is efficiently produced. Can be used for various purposes. According to the invention of claim 5, in addition to the effect of the invention of claim 1, since a compound word having a large number of words forming the compound word is preferentially output, a more appropriate compound word can be efficiently used. it can.

According to the invention of claim 6, the same effects as those of claims 1 to 5 can be obtained.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of a compound word search device according to the present invention.

FIG. 2 is a diagram showing the contents of a word dictionary in the first embodiment.

FIG. 3 is a diagram showing the contents of a collocation dictionary in this embodiment.

FIG. 4 is a diagram showing an example of contents stored in a complex word temporary storage unit in the present embodiment.

FIG. 5 is a diagram showing an input sentence according to the present embodiment.

FIG. 6 is a diagram showing an example of stored contents of a complex word temporary storage unit in the present embodiment.

FIG. 7 is a flowchart for explaining the operation of this embodiment.

FIG. 8 is a flowchart for explaining the operation of this embodiment.

FIG. 9 is a flow chart for explaining the operation of this embodiment.

FIG. 10 is a flowchart for explaining the operation of this embodiment.

FIG. 11 is a flowchart for explaining the operation of this embodiment.

FIG. 12 is a diagram showing an example of a collocation dictionary.

13 is a diagram showing a word dictionary created from the collocation dictionary shown in FIG.

[Explanation of symbols]

 101 Input Unit 102 Sentence Temporary Storage Unit 103 Word Dictionary 104 Conjunction Dictionary 105 Word Search Unit 106 Word Temporary Storage Unit 107 Conjunction Search Unit 108 Conjunction Temporary Storage Unit 109 Conjunction Sorting Unit 110 Output Unit

Claims (6)

[Claims] 1. A compound word search device for searching a compound word composed of a plurality of words in an input sentence, comprising: input means for inputting a sentence; and sentence temporary storage means for temporarily storing the sentence input by the input means. And a word storage unit that stores in advance a headword that is a basic form and a collocation identifier that includes the headword and its inflection word in association with each other, an identifier that identifies each collocation, and an identifier that is specified by the identifier. A complex word stored in advance in association with a complex word and its translation, and the sentence stored in the sentence temporary storage means is divided into words, and when the divided word is not in the basic form, it is converted into a basic form word. A word dividing unit for converting; a word detecting unit for detecting an identifier of a corresponding collocation word when the word obtained by the word dividing unit matches a heading word stored in the word storage unit; When the means detects a collocation identifier, the identifier and the word that remains divided and the basic form or the word that remains divided are detected.If the means does not detect the word that remains divided and the basic form or word division A word temporary storage means for storing the as-set words in a divided order as a set; and a compound word identified by the identifier when the identifiers stored in the word temporary storage means match in a plurality of sets. A collocation search unit that searches collocation translations from the collocation storage unit, a collocation temporary storage unit that temporarily stores all collocations and collocations searched by the collocation search unit for each collocation, and the collocation temporary storage unit. Order determination means for determining the optimal order of stored collocations by a predetermined procedure, and output means for outputting collocations and collocations in descending order of order determined by the order determination means. Phrase search apparatus according to claim that there were example. 2. The rank determining means counts the number of times the order of words stored in the word temporary storage means and the order of words stored in the word temporary storage means are interchanged, and the number of words having a small number of times is changed. The multi-word search device according to claim 1, further comprising: a word replacement number counting unit for increasing the rank of. 3. The rank determining means determines, between the first and last words forming the collocation stored in the collocation temporary storage means, what words are not forming the collocation stored in the word temporary storage means. 2. The multi-word search device according to claim 1, further comprising a word insertion number counting unit that counts whether or not words have been inserted and increases the rank of the multi-words having the smaller number. 4. The rank determining means calculates a matching rate between a word forming a compound word stored in the word temporary storage means and a word forming a compound word stored in the word temporary storage means, The multi-word search device according to claim 1, further comprising a word concordance rate calculation unit for increasing a rank of a multi-word having a high matching rate. 5. The rank determining means counts the number of words stored in the word temporary storage means that match a word forming one compound word stored in the multiple word temporary storage means, and the number is The multi-word search device according to claim 1, further comprising a word matching number counting unit for increasing the rank of a large multi-word. 6. The rank determining means counts the number of times of exchange of the order of words stored in the word temporary storage means and the order of words stored in the multiple word temporary storage means, and the multiple words with a small number of times. And a word replacement number counting unit for increasing the rank of the word replacement number counting unit, when there is a compound word of the same rank in the word replacement number counting unit, between the first and last words constituting the compound word stored in the compound word temporary storage means. In the word insertion number counting unit, which counts the number of words that do not form a collocation stored in the word temporary storage means and increases the rank of collocations with a small number, When there are collocations having the same rank, the matching rate between the words forming the collocation stored in the word temporary storage means and the words forming the collocation stored in the collocation temporary storage means is calculated, And a word matching rate calculating unit for increasing the rank of a collocation having a high matching rate, and when there is a collocation having the same rank in the word matching rate calculating unit, one collocation stored in the collocation temporary storage means is configured. 2. The multi-word search according to claim 1, further comprising a word match number counting section for counting the number of words stored in the word temporary storage means that match the corresponding word, and increasing the rank of the multi-word having a large number. apparatus.