JPH11238068A - Text retrieval device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent the omission of retrieval at the time of executing a fuzzy retrieval. SOLUTION: A retrieval character string developing part 10 generates a substitution character string obtained by substituting an erroneous recognition permission character '#' for a part of the character string. The number of erroneous recognition permission characters is determined in accordance with a matching degree designated by an input part 4. A matching part 12 can detect a candidate character string containing an error character by considering '#'as one arbitrary character. Then, '#' is considered to be two arbitrary characters or '##' is considered as one arbitrary character. Thus, the candidate character string generating erroneous division and erroneous connection can be detected in character recognition for generating the text of a retrieval object.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a text search apparatus for a text in which a character string may be inaccurate, and more particularly to a reduction in search omission and weighting of search results.

[0002]

2. Description of the Related Art Conventionally, there has been a text search apparatus for designating a search character string and searching for the search character string included in a document or a character string. A character string search function provided in a word processor is one example of such a function.

[0003] It is assumed that the documents and character strings to be searched are basically free from errors. In the search, only when the character string included in the search target text completely matches the search character string, it is determined that the character string of interest exists in the search target text.

On the other hand, when the text to be searched is text data read by an optical character reader (OCR), a recognition error in the reading causes an error.
The probability of incomplete text containing incorrect strings is high. Since Japanese OCR has low accuracy, it is particularly likely to do so. If a search is performed on the incomplete text based on a perfect match with the search character string as described above, a search omission may occur. That is, a character string portion that would have hit if the search target text was correctly read may not hit due to recognition error.

[0005] In order to prevent such a search omission, there is a technique for performing a search by allowing a certain degree of ambiguity of a search target (hereinafter, referred to as an ambiguous search). JP-A-62-44
In the first prior art of fuzzy search disclosed in Japanese Patent No. 878, when a plurality of candidates are obtained as a result of recognition, candidate characters are embedded in a search target text and search is performed (eg, sentence [ [Character]] [Motogi]
Sentence. . . ). In the second prior art of ambiguous search disclosed in Japanese Patent Application Laid-Open No. 8-7033, if a plurality of candidates are obtained for each character that has been subjected to character recognition, those are left in the index. In this case, in the index in which the recognition result is stored for each character, a plurality of character candidates of the recognition result may be stored for one character to be recognized. These two techniques make the recognition result, that is, the search target text ambiguous. On the other hand, JP-A-6-195
The third prior art of ambiguous search disclosed in Japanese Patent Application Laid-Open Nos. 387, 7-152774 and 8-63487 is to add ambiguity to a search character string. This method creates an incomplete search string in which a part of the search string that is easily recognized by mistake is replaced with a character pattern (misrecognition pattern) that may be misrecognized. In addition, a search is also performed using an incomplete search character string. The types of the misrecognition pattern include a character error, a wrong division, and a wrong combination. For example, "character" is easily recognized as "study", but such a type is a character error. Also, “ka” is easily recognized as “hihi”, but such a type is erroneously divided, while “51” is easily recognized as “pull”, but such a type is erroneously combined.

[0006] By performing an ambiguous search, there is a merit that a search omission can be reduced.
There is a possibility that a search error that a character string different from the search character string matches the search character string is included.

[0007]

However, the first prior art of the fuzzy search described above has a problem that the capacity of the search target text as the recognition result increases and that the search is not performed unless the search result remains in the recognition result. In addition, the misrecognition pattern includes
There was also a problem that it was not possible to cope with misconnection.

The second conventional technique of ambiguous search also has a problem that the capacity of the index as a recognition result increases and that the search is not performed unless the result remains in the recognition result.

In the third prior art of ambiguous search, it is necessary to prepare an erroneously recognized pattern separately from a search character string, and there is a problem that the capacity thereof increases. Also, for example, erroneous combination occurs depending on a combination of consecutive characters, so that many patterns can exist. It is difficult to prepare all possible erroneous recognition patterns in advance.
When an erroneous recognition pattern not prepared in advance occurs, there is a problem that a search is omitted. Increasing the capacity of a search target text, an index, an erroneously recognized pattern, and the like used for the search has caused not only a problem that the storage device requires a large capacity, but also a problem that the search process takes time.

[0010] Further, if a plurality of candidates are registered in the index or misrecognition patterns are enriched in order to reduce search omissions, on the other hand, a character string which is originally unrelated to the search character string is hit in the search. There is a possibility that. In other words, there is also a problem that the search results include a lot of “garbage” (search errors), and the reliability of the search results decreases.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem. A text search apparatus which reduces data to be used for an ambiguous search, reduces search omissions, and reduces the effects of search errors. The purpose is to provide.

[0012]

According to a first aspect of the present invention, there is provided a text search apparatus for generating a substring of a search character string in accordance with a designated matching degree between the search character string and the substring. It has a partial string generating means, and candidate searching means for searching in a search target text for a candidate character string including a character string pattern matching the generated partial string.

The "matching degree" is a matching degree of a substring of a search character string to the search character string, and can be defined, for example, by a ratio of the number of characters between the search character string and the substring. The specified degree of coincidence may be a numerical range specification or a value indicating a threshold. The “substring” is obtained by masking a part of characters of the search character string, and the characters constituting the partial string hold information on the position in the original search character string. For example, the substrings “key ## do” and “key ## do” (# represents a masked character) of the search character string “keyword” are composed of the same set of characters, but the masking position Are different and are treated as different subsequences. Also, as shown in this example,
In some cases, not only are the characters forming a certain sub-sequence continuous with each other, but also a masking position is arranged between the constituent characters. The candidate searching means extracts, from the search target text, a character string in which the same character is arranged at the position of each character constituting the subsequence as a candidate character string. That is, in the extraction of the candidate character string, it does not matter whether the character at the masked position in the search character string matches or does not match.

A text search device according to a second aspect of the present invention comprises:
An error character string registration unit that registers an error character string pattern that may occur in the generation of the search target text, a candidate search unit that searches for a candidate character string based on the search character string, and the search character string in the candidate character string. In different parts, an error character string detecting means for detecting a registration error character string pattern registered in the error character string registration unit, and a priority corresponding to the possibility of matching the search character string with the candidate character string. And a priority assigning means for determining the priority of the candidate character string in accordance with the detection of the registration error character string pattern in the candidate character string. It is.

The search target text generated by character recognition or the like may include an erroneous character string, but the erroneous character string pattern is not random, and a pattern which is likely to occur for the original correct character or character string is not included. Can exist. The error character string registration section mainly stores such error character string patterns that are likely to occur. According to the present invention, an error pattern stored in the error character string registration unit is detected in a part of the candidate character string different from the search character string. For example, if the detected error pattern is for a corresponding part of the search string, it is determined that the error pattern part is likely to be a correct character string before the search target text is generated. Thus, the priority according to the possibility of matching can be set high.

In a preferred aspect of the present invention, there is provided a partial string generating means for generating a partial string of the search character string in accordance with a designated degree of matching between the search character string and the partial string, The candidate character string is a character string including a character string pattern that matches the generated partial string.

A text search device according to a third aspect of the present invention comprises:
In the above invention, the candidate search unit may perform the search by including a unit that regards an ambiguous character constituting a difference between the search character string and the subsequence as an arbitrary character in the search target text. Performing the search by including means for considering any two characters in the search target text; and determining two consecutive ambiguous characters constituting a difference between the search character string and the substring. And / or some means for performing the search with means for assuming an arbitrary character in the search target text.

As a result, it is possible to search for candidate character strings each having a character error, a wrong combination, or a wrong division as an error pattern.

According to a fourth aspect of the present invention, there is provided a text search apparatus comprising:
The priority assigning means determines the priority in accordance with the detection frequency of the registration error character string pattern. According to the present invention, for example, it is determined that an original character string corresponding to a registration error character string pattern having a high detection frequency is likely to cause such an error, and a high priority can be given.

In a preferred aspect of the present invention, the erroneous character string registration unit stores the detection frequency in addition to the registered erroneous character string pattern.

[0021] A text search apparatus according to a fifth aspect of the present invention comprises:
There is provided a candidate character string display means for displaying the candidate character string according to the priority. ADVANTAGE OF THE INVENTION According to this invention, a user can grasp | ascertain the possibility of matching with the search character string in several candidate character strings based on a priority, For example, it is convenient at the time of checking the result of a search process. .

[0022]

Next, embodiments of the present invention will be described with reference to the drawings.

[Embodiment 1] FIG. 1 is a schematic block diagram of a text search apparatus according to an embodiment of the present invention. The present apparatus targets a text whose character has been recognized by the OCR as a search target, and stores an index storage unit 2, an input unit 4,
It includes a search unit 6, a target character position information storage unit 8, a search character string development unit 10, a matching unit 12, and an output unit 14.

The index storage unit 2 stores the search target text obtained by the OCR in the form of an index prior to the search. The index is obtained by associating a character that appears in the search target text as a key with an appearance position of the character.

The input unit 4 receives search conditions such as a search character string and a degree of coincidence from the user.

The search unit 6 obtains a search character string from the input unit 4, searches the index stored in the index storage unit 2 for each character included in the search character string, and determines the appearance position of each character in the search character string. The data is output to the target character position information storage unit 8, and the target character position information storage unit 8 stores this.

The search character string developing unit 10 is a subsequence generating means for generating a substring of the search character string. Extract and generate a replacement string containing the subsequence from the column. The replacement character string is generated by replacing some characters of the search character string with, for example, a symbol “#” and masking the original character. The part of the replacement character string other than the part replaced by “#” is a partial string composed of the original characters of the search character string.

The matching section 12 includes a search character string expanding section 1
Using the replacement character string output from 0, matching with the target character position information stored in the target character position information storage unit 8 is performed. The matching result is output to the output unit 14 and displayed on a screen such as a CRT as a search result on a screen.

Next, the operation of each component will be described using a specific example. FIG. 2 is a schematic diagram showing an image of a search target text. Here, the search target texts used in the example are document A, document B, and document C. The document A has a character string “Persia” from the tenth character from the top. Similarly, in the document B, a character string “Belsha” exists from the fifth character from the top, and in the document C, a character string “Penoresha” exists from the 21st character from the top.

FIG. 3 is a schematic diagram showing an index image generated for these search target texts and stored in the index storage unit 2. The index indicates the type of characters that appear in the search target text (in the figure,
Shown on the left end. ) Is used as a key, and the position in the document where the character type appears is classified for each key. The appearance position of the character is a number Ndoc that distinguishes documents A to C in the figure.
(Document A is “1”, Document B is “2”, Document C is “3”)
And the number of characters Nchar from the beginning of each document (Ndoc,
Nchar).

The user inputs a search character string “Persia” and a matching score of 70% to the input unit 4 as search conditions.
Here, the degree of matching η is defined by the ratio of the number of characters in the substring to the search character string. That is, assuming that the number of characters in the search character string is M and the number of characters remaining in the replacement character string without being replaced is m, the degree of coincidence η = m / M × 100 [%].
The coincidence input to the input unit 4 is a threshold ηth of η,
The present apparatus searches for a character string having a matching degree exceeding ηth in the search target text. If the threshold ηth of the matching degree is low, “garbage” included in the search result increases.
Is generally about 70% or more. On the other hand, if the threshold ηth is higher than necessary, the possibility of occurrence of search omission increases. In consideration of this point, here, ηth is set to 70%.

The input unit 4 notifies the search character string “Persia” to the search unit 6. When the search unit 6 obtains this search character string, the characters "pe", "ru", "shi",
The index storage unit 2 is searched using “「 ”as a key, and the search result is stored in the target character position information storage unit 8. Specifically, in this example, the appearance position (1, 1
0), (3,21), the appearance position (1,1
1), (2,6), appearance position (1,12) for the character "shi",
(2,7), (3,24), the appearance position (1,1
3), (2,8), and (3,25) are stored in the target character position information storage unit 8. FIG. 4 is a schematic diagram showing an image of the target character position information stored in the target character position information storage unit 8.

The search character string developing unit 10 receives the search character string and the threshold value ηth of the degree of coincidence from the input unit 4, and replaces the number of characters corresponding to the degree of coincidence in the search character string with the erroneously recognizable character. Generate a replacement string. Misrecognition-allowed characters (ambiguous characters) are represented here by the symbol "#". The character of the search character string in the portion where the misrecognition allowable character is placed is masked. The masking means that the matching between the replacement character string and the search target text, which will be described later, is not considered.

Here, since the number M of characters of the search character string is M = 4, the number m of characters of the partial string that satisfies the threshold ηth = 70% of the matching degree is 3 or 4. Therefore, the search character string expansion unit 1
0 generates a replacement character string that does not include any misrecognition-permitted character (this is equal to the search character string) and a replacement character string that includes only one misrecognition-permitted character. Specifically, in this example, the replacement character strings are “Persia”, “#Russia”,
Five are generated, namely "sha", "per # a", and "persi #".

The matching section 12 is a candidate searching means for matching the replacement character string with the search target text and searching for a candidate character string having a possibility of matching the search character string. Matching is performed by comparing a relative character position in the replacement character string with an appearance position stored in the target character position information storage unit 8. Hereinafter, α and β are normal characters appearing in the replacement character string.

For example, matching of the two-character part "αβ" appearing in the replacement character string is performed as follows. First, the matching unit 12 searches the target character position information storage unit 8 using α and β as keys. Here, the appearance positions corresponding to α and β are (Ndoc (α), Nchar (α)) and (Ndoc (α)), respectively.
(β), Nchar (β)). The matching unit 12 finds Ndoc (α) = Ndoc (β) with respect to the document number and Nchar (β) = Nchar (α) +1 with respect to the character position, thereby finding two consecutive characters “αβ”. Detect presence.

Next, the character string portions "α # β", "α ## β", "α #### β",
Matching processing for “α ####... Β” and the like is performed as follows. There are three basic matching rules for the misrecognition allowable character "#".

(I) "#" is considered to be the same as any one character, (ii) "#" is considered to be the same as any two characters, and (iii) "##" is considered to be any one character. Are considered identical.

(I) is a rule corresponding to a character error. (Ii) and (iii) are rules corresponding to erroneous division and erroneous combination, respectively.

Rule (i) is that Ndoc (α) = Ndoc (β) and Nc
This is realized as a search for an appearance position where har (β) = Nchar (α) +2. Rule (ii) is that Ndoc (α) = Ndoc (β) and Nch
This is realized as a search for an appearance position where ar (β) = Nchar (α) +3. Rule (iii) is realized by searching for an appearance position where Ndoc (α) = Ndoc (β) and Nchar (β) = Nchar (α) +2. Through these searches, the matching unit 12 detects the presence of a character string pattern such as “α # β”.

The matching unit 12 performs the above-described matching process on each part of the replacement character string, and detects the presence of the replacement character string in the search target text. For example,
For the replacement string "Persia", "pe (1,10)"
"Le (1,11)", "Sh (1,12)", "Cha (1,13)" conform to the above basic matching rules, and the matching unit 1
2 outputs a pair “Persia (1, 10)” of the replacement character string and the appearance position of the first character as a matching result. Those that have been matched once are removed from the target so that they do not match with another replacement character string. In addition, for the replacement character string “#Russia”, any one character preceding “R (2,6)” and “R (2,2,
6), “S (2,7)”, and “G (2,8)” are detected, and the matching unit 12 outputs “#Russia (2,
5) is output. Also, for the replacement character string "pe # sha", based on the basic rules and rule (ii), "pe (3,21)",
Any two characters following this, and “shi (3,24)” and “ya (3,25)” following the arbitrary two characters conform to the above-described matching rules, and the matching unit 12 outputs “Pe # sha (3,21)” is output. The search is also performed for the replacement character strings “Per # a” and “Persian #”, but in this example, there is no character string (candidate character string) that hits them.

The output unit 14 displays a search result on a screen based on the matching result obtained by the matching unit 12. The above-described matching unit 12 outputs the position of the candidate character string as a matching result.
For example, "Document A: (1,10) Document B: (2,
5) Document C: (3,21) "can be displayed. In addition, ranking may be performed according to the number of allowable characters for erroneous recognition, such as perfect match, one-character ambiguity, or two-character ambiguity.

The output unit 14 may access the search target text based on the document number and the character position obtained from the matching unit 12, obtain a candidate character string, and display it. Further, the matching unit 12 itself may be configured to access the index storage unit 2 based on the position information of the candidate character string, and output the candidate character string to the output unit 14 together with the position information. With such a configuration,
The output unit 14 can display the content including the candidate character string, for example, “Document A: Persia (1,10) Document B: Belsha (2,5) Document C: Penoresha (3,21)”.

Incidentally, the present apparatus can be configured by using a general-purpose computer. In particular, the functions of the search unit 6, the search character string expansion unit 10, and the matching unit 12 are performed by a central processing unit (CPU). Can be performed.

According to the present invention described in the present embodiment,
For example, by being incorrectly recognized in character recognition,
Respond to character errors, mis-segmentation, and mis-combinations, without using information on what character or character string is converted to the wrong character or character string to generate the search target text , And a candidate character string that may match the search character string can be completely searched.

[Second Embodiment] FIG. 5 is a schematic block diagram of a text search apparatus according to a second embodiment of the present invention. The same components as those of the above-described embodiment among the components of the present apparatus are denoted by the same reference numerals, and the description will be simplified. This device has a text storage unit 2 in addition to the configuration of the above device.
0, an error character string registration unit 24, and a ranking unit 26 are the main differences.

The text storage unit 20 stores search target texts, and each document is assigned a document number and can be distinguished from each other.

The matching section 22 performs the same processing as in the apparatus of the above embodiment to obtain position information of the candidate character string.
The matching unit 22 of the present apparatus further accesses the text storage unit 20 based on the position information, acquires and outputs the candidate character string. At this time, the position information can also be output.

The erroneous character string registration unit 24 stores an erroneous character string pattern that is a character or a character string that is easily erroneously recognized in character recognition.

Upon obtaining the candidate character string from the matching unit 22, the ranking unit 26 searches the candidate character string for an error character string pattern registered in the error character string registration unit 24. Then, the ranking unit 26 determines a priority according to the possibility of matching between the candidate character string and the search character string according to the result (priority assigning means). The ranking unit 26 outputs the candidate character strings and their priorities to the output unit 14.

Next, the characteristics of the operation of the present apparatus will be described using a specific example. FIG. 6 is a schematic diagram showing an image of the search target text. Here, the search target texts used in the example are document A, document B, and document C. Document A has a character string “scanner” from the tenth character from the top. Similarly, the document B has a character string “scan” from the fifth character from the top, and the document C has two character strings “scan” from the top.
The character string "Skyana" exists from the first character.

FIG. 7 is a schematic diagram showing an image of an index generated for these search target texts and stored in the index storage unit 2.

The user inputs a search character string “scanner” and a degree of coincidence of 70% to the input unit 4 as search conditions.

The input unit 4 notifies the search character string “scanner” to the search unit 6. Upon obtaining the search character string, the search unit 6 searches the index storage unit 2 using each of the characters constituting the search character string as a key, and stores the result in the target character position information storage unit 8, as in the above embodiment.

The search character string developing unit 10 receives the search character string and the threshold value ηth of the degree of coincidence from the input unit 4, and generates a replacement character string corresponding to the search character string.

Here, based on the number M of characters of the search character string M = 4 and the threshold value ηth = 70% of the matching degree, the search character string developing unit 10 determines that the replacement character string containing no erroneously recognizable character and the erroneous recognition allowable character string. Generate a replacement string containing only one character.
Specifically, in this example, as the replacement character strings, “scanner”, “#kana”, “scanner”, “scanner”,
Five “scan #” are generated.

The matching section 22 matches this replacement character string with the search target text, and detects the presence of the replacement character string in the search target text. For example, for the replacement character string “scanner”, “s (1,10)”, “ki (1,11)”, “ya (1,12)”, and “na (1,13)” match. . The matching unit 22 determines, based on the position information,
A candidate character string “scanner” is acquired from the search target text stored in the text storage unit 20, and a set “scanner (1,10)” of the candidate character string and the appearance position of the first character is output to the ranking unit 26. For the replacement character string “Ska #”, based on the basic rules and the rule (i) described in the above embodiment, “S (2,5)”, “G (2,6)”, “ 2,7) "and any one character following it. The matching unit 22 accesses the text storage unit 20 based on the position information, acquires the candidate character string “scan”, and outputs “scan (2, 5)” as a matching result. The basic rules and rules for the replacement character string
Based on (i), "su (3,21)", "ki (3,22)", any one character following it, and "na (3,24)" following this one character match I do. The matching unit 22 accesses the text storage unit 20 based on the position information to acquire the candidate character string “SCAN”, and outputs “SCANANA (3, 21)” as the matching result.

Upon obtaining the matching result from the matching unit 22, the ranking unit 26 ranks the candidate character strings. Here, the ranking is a process of determining a priority according to the possibility that the candidate character string matches the search character string. The number of characters in the part where the candidate character string differs from the search character string (error character string) and the registration of the error character string It is determined based on whether or not it is registered in the unit 24 as an error character string pattern.

FIG. 8 is a schematic diagram showing an example of an error character string pattern registered in the error character string registration unit 24. The figure shows that in character recognition when generating the search target text, the character or character string to the left of “→” is likely to be erroneously recognized as the character or character string to the right. For example, "su" becomes "i", "ya" becomes "ya" or "ゃ",
"Na" indicates that it is likely to be mistakenly recognized as "me", and "ru" may be easily recognized as "nore".

For example, when the candidate character string completely matches the search character string, the ranking unit 26 assigns a point “100” as a numerical value indicating the priority, and when one character does not match, the point “10”. Is given. Then, the ranking unit 26 checks whether an error character string that is a difference between the candidate character string and the search character string is registered as an error character string pattern in the error character string registration unit 24, and if the error character string is registered. In this case, for example, "40" points are added to the points already acquired.

Thus, for example, the candidate character string "scanner"
Is an exact match, so you get the point "100"
Since the candidate character string “Sukiyana” does not match one character and “Error → Ya” is registered in the error character string registration unit 24,
A point "50" is obtained by adding the points "10" and "40". On the other hand, although the candidate character string “scan” does not match one character, since the error character string is not registered in the error character string registration unit 24, only the point “10” is obtained. Then, for example, as a ranking result, the ranking unit 26 sets a set of candidate character strings, their position information and points, for example, “scanner (1, 10, 100)”, “scanner (2, 5, 50)”, “scan (3 , 21,10) "to the output unit 14
Output to

When the output section 14 obtains the ranking result from the ranking section 26, the output section 14 can perform display using the points included therein. For example, points are high,
That is, the candidate character strings can be displayed on the screen in descending order of the possibility of matching with the search character string. In addition, the output unit 14 may display only candidate character strings that have obtained points equal to or more than a certain value, or may group and display those having points within a specified range.

Although the ranking unit 26 assigns a fixed number of points to the error character string pattern registered in the error character string registration unit 24, the number of points may not necessarily be uniform. For example, the error character string registration unit 24
, The degree of error susceptibility represented by the frequency of detection of each error character string pattern is stored, and this is reflected in the ranking, whereby a more detailed ranking can be performed. For example, the error probability is set with an adjustment coefficient of 0 to 1,
As the point, there is a method of multiplying a common point of the error character string pattern by an adjustment coefficient of error probability. In such a method, for example, if the adjustment coefficient of the candidate character string “scan” is 0.8 in the above example, the point is 10 + 40 × 0.8 = 42. Further, it is possible to configure so that the user increases or decreases the detection frequency of the error character string pattern based on the search result.

According to the present invention described in the present embodiment,
Similar to the invention described in the first embodiment, the matching itself in the search processing does not require the error character string pattern registered in the error character string registration unit 24, and can handle character errors, mis-divisions, and mis-combinations. It is possible to search for a candidate character string that may match the search character string without fail. By performing a full search, those that are unlikely to match the search string are also detected as candidate strings,
It cannot be denied that the ratio of “garbage” (search error) included in the matching result increases. The present invention makes it possible to perform a complete search and to display the search results in a more probable order, thereby grasping the importance (priority) of each candidate character string when the user uses the search results. Thus, even if a search error occurs, its influence on actual use can be reduced. In the conventional method of performing a search by expanding a search character string using an error character string pattern, if the dictionary in which the error character string pattern is registered is not sufficient to some extent, the number of search omissions increases and the reliability decreases. On the other hand, in the present invention, even when there is no data in the error character string registration unit 24, the search can be performed without fail, and the accuracy of the ranking can be improved by enriching the data in the error character string registration unit 24. .

[Embodiment 3] The third embodiment of the present invention relates to another example of search processing using another search character string. The configuration of a text search apparatus according to this embodiment is as follows.
This is the same as the device of the second embodiment.

FIG. 9 is a schematic diagram showing an image of a search target text. Here, the search target texts used in the example are document A, document B, and document C. Document A has a character string “Altalia” starting from the tenth character from the beginning. Similarly, in the document B, a character string “Al column a” exists from the fifth character from the beginning, and in the document C, a character string “Al yuria” (“Yu” is a kanji) from the 21st character from the beginning. Exists.

FIG. 10 is a schematic diagram showing an image of an index generated for these search target texts and stored in the index storage unit 2.

The user inputs a search character string “Altalia” and a coincidence degree of 60% to the input unit 4 as search conditions.

The input unit 4 notifies the search character string “Altalia” to the search unit 6. Upon obtaining the search character string, the search unit 6 searches the index storage unit 2 using each of the characters constituting the search character string as a key, and stores the result in the target character position information storage unit 8, as in the above embodiment.

The search character string developing unit 10 receives the search character string and the threshold value ηth of the degree of coincidence from the input unit 4, and generates a replacement character string according to the search character string.

Here, from the number of characters M of the search character string M = 5 and the threshold ηth = 60% of the degree of coincidence, two characters are permitted to be erroneously recognized. Specifically, in this example, the search character string expanding unit 10 includes “Altalia”, “## Thalia”, “#Lutalia”, “# Lutah # a”, “# Lutari #”, "A # Rear", "A # T # A", "A # Tari #", "Al # A #", "Al # Re #", "Alta #"
# ”Is generated and output to the matching unit 22.

The matching section 22 matches this replacement character string with the search target text, and detects the presence of the replacement character string in the search target text. For example, for the replacement string "Altalia", "A (1,10)"
"Le (1,11)", "Ta (1,12)", "Li (1,13)",
"A (1,14)" matches. The matching unit 22 obtains the candidate character string “Altalia” from the search target text stored in the text storage unit 20 based on the position information, and sets “Altalia (1, 10) is output to the ranking unit 26. Also,
Any one character for the replacement string "# le # rear", followed by "le (3,22)", followed by any one character,
"Li (3,24)" and "A (3,25)" match. The matching unit 22 uses the text storage unit 20 based on the position information.
To obtain the candidate character string “Al Yuria (“ Yu ”is a kanji).
21) is output. Also, for the replacement character string "Al ## a", "a (2,5)" and "ru (2,6)" follow rule (iii) described in the first embodiment. One character and "a (2,8)" match. The matching unit 22 accesses the text storage unit 20 based on the position information to obtain the candidate character string “Al string A”, and outputs “Al string A (2, 5)” as a matching result.

Upon obtaining the matching result from the matching unit 22, the ranking unit 26 ranks the candidate character strings. In the present apparatus, the points given by the ranking unit 26 are extended to the case where the number of erroneously recognizable characters is two, and are defined for each case that may occur in that case. For example, it can be determined as follows.

(A) In the case of a perfect match with the search character string: point “10”
0 ”, (b) If one character does not match: point“ 5 ”
0 ”, (b-1) If one mismatched character matches the error character string pattern: add point“ 30 ”, (c) If two characters do not match: point“ 1 ”
0 ", (c-1) If one mismatched character matches the error character string pattern: add point" 30 ". (C-2) If two mismatched characters match the error character string pattern: point" 60 ”.

It is assumed that the error character string patterns registered in the error character string registration unit 24 include “Tari → column” and “T (katakana) → Yu (Kanji)”.

The ranking section 26 determines the point “100” in the case of a perfect match for the candidate character string “Altalia”.
Is given, and 10 + 60 = 70 points are given to “Al column a” because two characters do not match and the error character string “Tari → column” is registered in the error character string registration unit 24. Also,
The candidate character string "Al Yuria"("Yu" is a kanji) does not match one character and the erroneous character string "Ta → Yu" is an erroneous character string registration unit 24.
, 50 + 30 = 80 points are given. Then, for example, as a ranking result, the ranking unit 26 sets a set of candidate character strings, their position information, and points, for example, “Altalia (1,10,100)”, “Al sequence (2,5,70)”, “Al Evening rear (3, 21, 80) "is output to the output unit 14.

When the output unit 14 obtains the ranking result from the ranking unit 26, for example,
Display candidate character strings on the screen. In addition, the output unit 14 may display only candidate character strings that have obtained points equal to or more than a certain value, or may group and display those having points within a specified range.

Instead of the ranking using the error character string registration unit 24, it is possible to easily perform the ranking of perfect match, one-character ambiguity, and two-character ambiguity.

[Brief description of the drawings]

FIG. 1 is a schematic block configuration diagram of a text search device according to a first embodiment of the present invention.

FIG. 2 is a schematic diagram illustrating an image of a search target text according to the first embodiment.

FIG. 3 is a schematic diagram showing an image of an index of a search target text according to the first embodiment.

FIG. 4 is a schematic diagram illustrating an image of target character position information stored in a target character position information storage unit.

FIG. 5 is a schematic block diagram of a text search device according to a second embodiment of the present invention.

FIG. 6 is a schematic diagram illustrating an image of a search target text according to the second embodiment.

FIG. 7 is a schematic diagram illustrating an image of an index of a search target text according to the second embodiment.

FIG. 8 is a schematic diagram illustrating an example of an error character string pattern registered in an error character string registration unit.

FIG. 9 is a schematic diagram illustrating an image of a search target text according to the third embodiment.

FIG. 10 is a schematic diagram illustrating an image of an index generated for a search target text and stored in an index storage unit.

[Explanation of symbols]

2 index storage unit, 4 input unit, 6 search unit, 8
Target character position information storage unit, 10 search character string development unit,
12, 22 matching unit, 14 output unit, 20 text storage unit, 24 error character string registration unit, 26 ranking unit.

Claims (9)

[Claims] 1. A text search apparatus for performing a search process on a search target text based on a search character string, wherein the search character string is determined in accordance with a specified degree of matching between the search character string and a substring thereof. Substring generating means for generating a substring of; and candidate searching means for searching in the search target text for a candidate character string including a character string pattern matching the generated substring. Text search device. 2. A text search apparatus for searching a search target text for a candidate character string having a possibility of matching the search character string, wherein an error character string in which an error character string pattern that may be generated in the generation of the search target text is registered A registration unit, a candidate search unit that searches for the candidate character string based on the search character string, and a registration error registered in the error character string registration unit in a part of the candidate character string different from the search character string. An error character string detecting means for detecting a character string pattern; and a priority assigning means for setting a priority to the candidate character string in accordance with a possibility of matching with the search character string, wherein the priority assigning means Determining the priority of the candidate character string according to detection of the registration error character string pattern in the candidate character string. 3. A substring generating means for generating a substring of the search character string in accordance with a designated matching degree between the search character string and the substring, wherein the candidate character string 3. The text search apparatus according to claim 2, wherein the character string includes a character string pattern that matches the substring. 4. The candidate search means includes means for assuming an ambiguous character constituting a difference between the search character string and the subsequence as an arbitrary character in the search target text to perform the search. The text search device according to claim 1 or 3, wherein: 5. The candidate search unit includes a unit that regards an ambiguous character that constitutes a difference between the search character string and the substring as any two characters in the search target text, and performs the search. The text search device according to claim 1 or 3, wherein: 6. The candidate searching means includes means for regarding two consecutive ambiguous characters constituting a difference between the search character string and the partial string as an arbitrary character in the search target text. The text search device according to claim 1 or 3, wherein the search is performed by performing the search. 7. The text search device according to claim 2, wherein the priority assigning unit determines the priority according to a detection frequency of the registration error character string pattern. 8. The text search device according to claim 7, wherein the error character string registration unit further stores a detection frequency of the registration error character string pattern. 9. The text search device according to claim 2, further comprising candidate character string display means for displaying the candidate character strings according to the priority.