JPS6128134A - Symbol string collecting device and its control system - Google Patents

Abstract

PURPOSE:To extract flexibly an optional string among non-structure symbol strings in a short time by inputting a symbol string sequentially and transmitting externally the location of the symbol string similar to a registered collation symbol string. CONSTITUTION:The titled device is provided with a means 110 storing a symbol string in a bit pattern related to a symbol, plural counters 120 using a read signal 12 to control the change in the content, a register 130 as a temporary storage means and a means 140 adding the content of the counter 120 of the lowest stage and the content of the register 130, and the counter 120 includes a logical operation means 123 and a means 124 storing the result of operation. When a transfer signal 122 is applied at the leading of one shot pulse of a clock 121, the content of the counter 120 is transferred from the upper to the lower position, the collation processing to all symbol strings is executed and the content of the counter 120 of the final stage represents the similarity between the input symbol string and the collated symbol string. The similarity and the content of the register 130 are added (140) and the result is transmitted externally as the similarity output 125.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention relates to a component of an information processing system, and more specifically CC is a symbol string matching device that extracts a specific symbol string from a long symbol string. and its matching method.

(Prior art and its problems) - The F symbol string matching device can be used to extract % symbol sequences in a batan recognition system, - extract keywords from the original text file of sentences created in a professional, and support language translation. It is used to construct a valve structure take base using abbreviated explanations of communication texts, figures, images, texts, etc., and is indispensable for the form of these information processing systems that are becoming intelligent.

Conventional symbol string matching requires sequential processing using general-purpose computer software, requires an enormous amount of processing time, and is limited to small-scale applications. In addition, the target of matching was limited to structured symbol strings separated by words. - as a column.

If you want to find out where the button is located from Kn symbol strings in a text consisting of Kn symbol strings, m (n, -
m + 1) times of verification processing is required. Searching for a sentence with n-=10" character strings from a text with m=109 character strings stored on a magnetic disk or original disk requires 10 matching processes. Therefore, the text,
Since it is impractical to search using supplementary source information such as images, figures, and sounds, searches have been limited to searching by adding keywords to the source information in advance or searching for takes structured in a tabular format. .

Another disadvantage is that the processing time is too long for flexible symbol string matching that allows variations in the constituent elements of symbol strings.

Furthermore, the problems of the conventional symbol string matching device and its matching method will be explained in detail.

FIG. 6 shows the text that is the subject of symbol string matching. This text shows the beginning of the report as a single column. A large number of such texts are stored in the file memory of the word processor. When searching for what you need from such texts, you are required to be able to search for it directly using words that indicate the content of your request.

For example, the text in Figure 1 is memory, bu
In order to know whether the paper contains a symbol string such as ble, K
In that text memory, memories
It is necessary to check whether there is a matching part in the symbol string such as Jpbubble. Comparing such symbol strings with text would take a very long time if conventional computers and software were used.

A typical A4 size English text is approximately 3,000 characters long, including spaces between words. On the other hand, the length of the symbol string used for comparison and matching is de
Vice also has 6 characters. Matching between 6-character and 3000-character strings generally requires character comparisons a number of times on the order of the product. Even if the character comparison time in the microprocessor is 1 μsec, it takes 18 m5 ec to check each symbol string.

In reality, the number of characters in the text to be skipped is ten, and the number of characters in the symbol string to be compared may exceed 100. The number of symbol strings that can be matched is not just one, but tens. In that case, the verification time spans several hundreds of hours. Therefore, such matching is practically impossible, and in reality, key words are extracted in advance by acquisition, and matching is limited to the extracted keywords.

Furthermore, it is difficult to match symbol strings in which some symbols are missing or incorrect, or symbol strings to which extra symbols are added. For example, take the string 'MEMORY'' and enter the string MEMXR,Y in which the middle character '0' is replaced with another character 11 'X'.
MRMXOI (,Y
If " is input, the collation or character string '"M, BΔ40R
It is desirable for a symbol string matching VC to be able to output information similar to "Y".However, in conventional symbol string matching devices and matching methods, errors in some symbols can be detected by superimposing them on the matching symbol string and comparing them. Although it is possible to check for similarities, it was not possible to deal with missing some symbols or adding extra characters.

(Objective of the Invention) The object of the present invention is to easily solve the drawbacks of the conventional symbol string matching device and its method, and to shorten any symbol string from one of unstructured symbol strings such as text, images, figures, etc. In time,
An object of the present invention is to provide a symbol string matching device and its matching method that can perform flexible extraction.

An object of the present invention is to provide a low-cost symbol string matching device that can match even a symbol string that is truncated, missing or erroneous in some symbols, or has extra symbols added. (Structure of the Invention) Therefore, the present invention According to , the following symbol string matching device 4 and its open side j method are obtained. That is, a symbol string storage means that takes a symbol code as an address input and stores a collation symbol string, a first calculation means connected to each of the output lines, and a first temporary storage means that interconnects the plurality of calculation means. a second temporary storage means connected to the output of a specific first temporary storage means; and a second calculation means connected to the second temporary storage means.

A symbol string storage means that uses the first code code as an address input and stores a collation symbol string, and an Ef41 connected to an output line during this period.
ε that mutually connects the calculation means of and the calculation means of the number of screws.
temporary storage means and a specific first temporary storage means 〒:+21'
, a second temporary storage means connected to the output of the means, and a second temporary storage means connected to the output of the means;
a second calculation means connected to the temporary storage means V; a register that stores all allowable similarities given from the outside; and a comparator that compares the contents of this register with the output of the second calculation means and outputs a comparison result. Symbol string matching device including

and a symbol string collation device 4 which includes 4#+i encoders connected to the registers connected to the comparators, and a symbol string storage means for storing the collated symbol strings in bit patterns associated with the symbols. He (he) milestone and (he +
1) N-th first connected to the th output line respectively;
Contents of the temporary storage means'(C(N-1)th content of the first temporary storage means V?CNth 1 self-sequence storage means' output line controlled by the Nth first calculation means jIfQ
Assume that the content of the specific first temporary storage means at different times is stored in the second performance:! :'? ] A symbol string matching system (I\, OIIi: I & II system type) that finds the similarity between the contents of the symbol storage means and the sequentially input symbol string by measuring the arithmetic processing by the means. > Le, (embodiment)] -11 sets (explanations are given to the history of Akira, using drawings.

FIG. 1 is an explanatory diagram of an embodiment of the present invention. This symbol string matching device I'7 encodes a long symbol string such as the one shown in FIG. 7
, a symbol string storage means 110 for storing the symbol string in a bit pattern associated with the symbol, and a symbol string storage means 110 for storing the symbol string in a bit pattern associated with the symbol. A register 1 as a second temporary storage means is connected to the output of a multi-stage multi-bit counter 120 whose contents are changed by the signal 112 and the lowest stage counter 120.
30 consists of an adder 140 which is a second arithmetic means for adding the content N6 of the counter 120 at the lowest stage and the content of the register 130. Counter 120 Fi
It includes a logic operation means 123 and a temporary storage means 124 for storing the results of the operation. Here, the logical operation means 123 and the temporary storage means 124 are used as the counter 120.
Each symbol of the collation symbol string is stored in each bit of the symbol string storage means 110 in a bit pattern associated with the symbol. In the example shown in the figure, 'MEMO' is stored in the symbol string storage means 110 of Fi6 Binoto.
A collation symbol string consisting of RY"06 symbols is stored. That is, each of the first to sixth bits of the symbol string storage means 110 is a symbol 1M",...E, RM II.
, pigeon 0〃, ``''R〃, ``■〃 is stored only in the addresses specified by ``Y'', and ``0'' is stored in the other addresses. The address of the symbol string storage means 110 corresponds to the type of symbol, and the first to sixth bits of the read signal 112 are the symbols 1-M〃, It I,; /7. Only when IXM /lkiO〃,kiR〃,paY〃 are input, ■'' is displayed.

The counter 120 in each stage operates in synchronization with the clock signal 121, and the output of each stage is connected to the data input of the next stage. Zero is supplied to the data input of the first stage counter 120, and the output N6 of the last stage counter 120 is applied to the register 130.

When the transfer signal 122 is applied at the rising edge of the clock signal 12, the contents of the counters 120 at each stage are transferred one stage at a time from the upper stage to the lower stage in the figure. Further, when the transfer signal 122 is not applied and a read signal of 1'' is supplied from the symbol string storage means 110, the content of the counter 120 increases by 1 (hereinafter referred to as increment) and becomes 0''. The counter 120 operated in this manner retains its contents in the case of a read signal of
It can be used in commercially available integrated circuits such as 5N74L8161N' manufactured by Texas Instruments.

FIG. 2 shows a timing diagram of clock signal 121 and transfer signal 122. The symbol string to be matched is 1 in period T@
symbols are input from the symbol string input terminal 111, during which two pulse signals before and after are applied as a clock signal 121. Transfer signal 122 is clock signal 12
1 is applied as a pulse signal including the rising edge of the pulse signal.

Therefore, each time one symbol is input, the transfer operation and increment of the counter 120 in each stage are performed alternately (provided that the read signal 112 is 1"). counter 120
Import the contents of.

FIG. 3 is an explanatory diagram of the operation of the symbol string matching device shown in FIG. 1.

As shown in FIG.
The same figure (al
I will explain. In addition, the same figure (bl, (cl
, (d).

In the figure, the first row shows the symbol string input from the symbol string input terminal 111, the second row shows the time ']'o-'l'7 temporarily determined for each symbol input, and the third row and below show the clock. The contents of the counter 120 after the subsequent pulse signal of the signal 121 is applied are shown as Nl, N2 . N3°N4.
N5. It is represented by N6 and the diameter value is shown. The last row shows the value to the output of adder 140.

First, before the verification operation, the contents N1 to N6 of the counters 120 in each stage are set to zero. Although this is not shown,
This can be easily done by applying a pulse signal to the clear terminal of the counter 120.

Figure 3 (with reference to al, the symbol string 'M' which is the same as the collation symbol string)
The operation when EMORY" is input from the symbol string input terminal 111 will be explained. First, the first symbol "'M"
Since the transfer signal 122 is applied at the rising edge of the pulse signal of the clock signal 121 when the counter 120 is input, the contents of the counters 120 in each stage are transferred all at once from the upper stage to the lower stage. At this point, the contents N1 to N6 of the counters 120 in all stages are zero. In addition, only the first and third bits of the reading signal 1]2 of the symbol string storage means 110 become 1", L7?:, and the subsequent clock signal of the clock signal ]21 is applied. 1 to the contents of the counters 120 in the first and third rows.
and N3 are incremented from 0 to 1. Now, the matching process for the first symbol "XM" of the symbol string to be matched is performed.

Similarly, when the next symbol "'E" is input, only the content N2 of the counter in the second stage is incremented, and becomes 2 from J. In this way, all symbol strings are collated. The contents of the final stage counter 120 at each time (NO in this example) are determined by the input 'J' before that time.
! 'l- indicates the f-constant number of the matching symbol with each symbol in the symbol string. In Fig. 3(a), at time 111T6 ~
The value 6 indicates that it matches the symbol string input before that time, that is, from Tl to T6.

Therefore, if the number of symbols in the verification symbol string is M (in the example shown in the figure, M = 6), the value of ~6 will be M if the same symbol string is input, and only one symbol is different. becomes M-1. That is, the contents of the counter 120 at the final stage indicate the degree of similarity between the input symbol string and the verification symbol string. Adder 1
20 adds the contents of the final stage counter 120 and the contents of the register 130 that stores the contents one time before, and transmits the result to the outside as a similarity output 125.

FIG. 3(b) shows the counter 12 in each stage when a symbol example in which some symbols '0' are incorrectly replaced with other symbols 'X' is input.
Indicates the content of 0. Three symbols from the beginning ゛M〃, ’E“,
The counter 120 at each stage for the input of 'M'' operates in the same manner as in FIG. 3(a). However, when the next fourth symbol
When ``X'' is input, the content N4 of the fourth stage counter 120 is not incremented because the read signal 112 of the fourth bit of the symbol string storage means 110 is O''.
Keep the same value 3 all. Therefore, the symbol string consisting of 6 symbols"
IMEMXR, Y" is input, 'f?,' and the value of ~6 after is 5, which is 1 smaller than in the case of Fig. 3(a). In other words, the symbol differs from the collation symbol string by only one symbol. It can be recognized that a column has been entered.

Figure 3 (C1 shows the inner boundaries of the counters in each stage when a symbol string with a part of the sequence "0" missing in the collation symbol string "M E Mll, Y 2 for input of 3 symbols from i'Ti: 2B 3 (XI
It operates in the same way as l.

However, the subsequent symbol string "R, γX"' is the collation symbol and 1
The time is 'J'4. T5. N4 at T6. ~5. ~6 of 11? It does not change to i and remains 3. ] F, the time when the symbol "R//" is input '■' 4 ~5 is incremented from 0 to 1. When the next symbol "Y" is input, ~6 is incremented,
It goes from 1 to 2. As a result, the final stage counter 120
The maximum number of contents of N6 is 3, and if you are only monitoring N6, the input symbol string will be
It is not possible to recognize that this is a symbol string without the 1lIII symbol. However, the output of adder]40 is N&:1:5, so the symbol string input from time T1 at 'I'6-+ contains five symbols that match the collation symbol string. Recognize that. By monitoring the output of the adder 140, even for symbol strings in which some symbols are missing, verification is possible. In addition, the output ~ of the device J 40 is transmitted to the outside as a similar inverse output 125 .

Figure 3 (d+ is the 7th column 'Mi' containing the extra symbol ''X')
The counter of each stage when the ray J do.

However, since the symbol string ir. I'i contains the extra sushi symbol + S
It has one symbol with the combination, so the time '1゛4,
1゛5,'J. ' N4, ~5, j in each of 6
It is not incremented by 1 to i1i of ~0 and remains at 3. However, when one shouts out, 'fr': 21j 'i' 5, 1
゛6, 9゛7 each i/-2fu・ke4, ~4, ~5, ~
The 1st line b of 6 is an increment type, and goes from 0 to 3. Once upon a time, N6λ・'. fK Ill'. L
,/Koketeguma It cannot be recognized that the input symbol string is a symbol example in which one extra symbol is added to the string. However, since the value of the output of the adder 140 becomes 6 at time 117, it can be recognized that the input symbol strings have a strong similarity.

In general, if the symbol string length of the verification symbol string is fn, the content N of the last stage counter]20 at time Ti is the time T
It shows the number of symbols in which the symbol string input from i −n+1 to time Ti matches the verification symbol string. The output of the adder 140 at time Ti, that is, the similarity output 125, indicates the sum of the contents of the final stage counter 120 at time Ti and one time before. That is, similarity output 25 indicates the sum of the number of symbols in which the input symbol string and the symbol string shifted by one symbol match the collation symbol string. Therefore, symbol strings with some symbols missing and symbol strings with extra symbols added can also be compared by monitoring the similarity output 125.

As mentioned above, the symbol string matching device shown in Figure 1 can extract not only symbol strings that completely match the verification symbol string, but also symbol strings in which any number of symbols are incorrect, making it flexible. It can be said to be a deep symbol string matching device.

Generally, when matching symbol strings, it is easier to output the number of symbols that differ from the matching symbol string than the number of symbols that match the matching symbol string. When the number of matching symbols is counted as shown in FIG. 1, the counter 120 calculates the log2(M-f
-1) bit number is required. However, if you count the number of different symbols, the acceptable percentage of discrepancies is 5%.
In this case, the counter 120 only needs to be able to count about 5 units of M, so the number of bits can be reduced.

A symbol string matching device that makes this possible can be realized by slightly modifying the symbol string matching device shown in FIG.

For example, if the read signal 112 of the symbol string storage means 110 is "
0", the color/re 120 is incremented or ticked (the value is decreased by 1). Alternatively, the contents of the symbol string storage means 110 are inverted, and when the reading number 112 is 1", The counter 120 may be increased or decreased. In this case, if the initial value of the counter 120 is to be incremented, Ku, zero, or ticked if it is to be incremented, it is necessary to prohibit the increment operation or decrement operation. . This is counter 1
A Nants gate circuit that receives all 20 bits as input detects that all bits are 1, or a NOR gate circuit detects that all bits are 0, and the detected output is used as a read signal 21 of the symbol string storage means/means 110.
This can be easily achieved by masking 2. However, if the counter 120 is a counter with a built-in circuit for detecting the maximum value or minimum value, the Nant gate circuit or the Nant gate circuit is not necessary.

It is also possible to replace the ifr counter 120 with a shift register that can be input in parallel. In this case, parallel transfer to the next stage shift register is opened at transfer time number 122, and left shift or right shift of the contents is controlled by read signal 112. Data that differs only in the most significant bit (6) or the least significant bit is supplied to the data input of the first stage shift register.

This kind of shift register is sold out by Texas Instrument Company74. LS 299 integrated circuits are available. If a shift register is used instead of the counter 120, when the number of symbols different from the collation symbol string is counted, the contents of the shift register will be automatically set to zero if the number of symbols different from the number of bits of the shift register is input. It swells, making it easier to control.

Furthermore, it is also possible to easily replace the counter with other logic operation means rather than just the 20-digit shift register.

Further, in the above description, a symbol string collation device that can tolerate the omission or addition of one symbol has been described, but it can be expanded to allow the omission or addition of a plurality of symbols. this is,
This is possible by using a plurality of stages of registers connected in series as the register 130 and by having the adder 140 add the contents of each register.

FIG. 4 is also an explanatory diagram of an embodiment according to the present invention.

This symbol string matching device extracts only the symbol strings showing high similarity # to the matching symbol string from the symbol strings to be matched, and the symbol string matching device 400 shown in FIG.
10 and a level setting register 420 are added.

Before starting the matching operation, the allowable similarity of the symbol string to be extracted is set in the level setting register 420. this is,
This is done by supplying a level signal 421 indicating the allowable similarity and a register signal 422 to the level setting register 420.

Comparator 410 (inputs the U degree output 125 and the contents of the level setting register 420 to inputs A and B, respectively, and compares them. Then, the AZB output indicating whether the former is thicker or equal is compared with the output signal 41) For the number M of symbols in the collation symbol string stored in the symbol string symbol storage means 110,
When the symbol string input terminal 111 is set to 1l-1), if the symbol string input from the symbol string input terminal 111 and the collation symbol string match νΔ with (M-1) or more symbols, the comparator 4]0
generates a verification output 411. Therefore, by monitoring the IAφ combined output 411, it is possible to extract only the verification symbol string and the symbol string indicating the 1h degree.

FIG. 5 is also an explanatory diagram of an embodiment according to the present invention.

This symbol string matching device collates multiple matching symbol strings in parallel and outputs a matching code indicating the matched matching symbol string. Symbol string matching unit 51 including a plurality of units 500
0 and each matching output 411 f output from each symbol string matching unit 500 and an encoder 520 inputting it,
Register set signal 42 to each symbol string matching unit 500
2 and a decoder 530 that supplies 2.

Symbol code input from symbol string input terminal 111, riO
, a transfer signal 121, a level signal 421 indicating an acceptable degree of similarity, and a level signal 421 indicating the permissible similarity are commonly supplied to each symbol string matching unit) 500.

To set the allowable similarity to the symbol string matching unit 500, the basic register set signal 532 is supplied to the decoder 530 together with the unit selection value +j53x, and the level signal 4
21 In supplying the entire symbol string matching unit 510,! : It is done. At this time, the decoder 530 applies the register set signal 422 only to the symbol string matching unit 500 for which the allowable similarity is to be set.

When a symbol code is sequentially inputted to form a symbol string 71' which becomes a verification null name from the symbol string input Aj5 child 111, each symbol string verification unit 50 (1j4j combination is performed in parallel with each verification symbol string registered in l). 17, reduce the 'Jul Ijdu dust with a high degree of similarity above the allowable similarity 11f (memorize the combined Be No. 1 island;) 1 Shi symbol string matching unino) 5 (1 (+ to I
An I-beta output 411 is generated.

En-1-E 52 (1) - Take this collation output 411 and combine it with the input power/6 classification code 5; 21 and 7. 4-Shows 0〃 No. 1 Young Iba No. 522 No. 2 occurs. Which one!]
l, the encoder has a 1" error indicating inconsistency (+'j No. 522 attack occurs 1").

Therefore, by monitoring the illusion code 522 and the classification code 521, it is possible to classify the same line as input. In addition, the encoder 520 and the tecoter '530
As a result, the number of input/output terminals can be reduced, making it easy to convert a one-inch column collation device into an LSI.

(Effects of the Invention) As described above, the symbol string matching section ■? of the present invention? Since it allows matching based on similarity, it also allows matching of symbol strings in which some of the symbols are incorrect. English words, especially nouns, often change into singular or plural forms, resulting in different final texts. For example, '
"rnemories" becomes plural form Vmemories
“In this case, the basic” nIemo r 3'
” is a matching symbol string, and the allowable similarity is set to exclude literary errors.
can also be extracted. It is also possible to match symbol strings with some symbols missing or symbol strings with extra symbols added. In general, it is not possible to determine in advance where in the symbol string the missing part of a symbol or the addition of an extra symbol will occur or how the symbol will be missing or added. Therefore, all possible string variations must be accounted for, and conventional string matching must be programmed to accommodate them. However, in the symbol string matching method and symbol string matching device of the present invention, this problem can be easily dealt with by setting an allowable degree of similarity. In other words, the symbol string matching device ζC and its method of the present invention enable flexible symbol string matching! ! In addition, human υ3 power Qif1: Reduces the number of children, easily enables J, SI, and lowers prices.

If the current 256 kilobit H and AM half-251 technology is used, a total of 128 collation symbol strings consisting of eight 8-bit code symbols can be stored in one chip, and they can be collated in parallel.

] The fact that it is possible to classify a symbol string consisting of 256 Zen symbols into 128 classes is possible by classifying up to 128 keywords ( It emphasizes that the extraction of the symbol string) can be completed all at once. In the past, it was difficult to simultaneously detect a large number of keywords, so the impact of the above chip is significant.

This symbol string identification device can be used to recognize OCR, devices, voice recognition devices, etc. with a bang! g! It is also useful for classifying feature sequences in systems that perform Converting this symbol string identification device into a one-chip LSI can also serve as a dictionary necessary for language translation.

If you connect a normal I(, AM) to this chip and store the translation of a word in correspondence with the classification code of each symbol string, the translation of 128 words per chinofu can be requested immediately upon completion of inputting the symbol string. The normal R and AM connected to the symbol string identification chip can store various types of information in correspondence with the classification code of each symbol, and depending on the F'L, various types of information can be stored. A symbol string information processing structure is achieved.For example, if the part-of-speech code of a word, the number of occurrences of a symbol string, and processing instructions for symbol string sentences are stored in correspondence with the classification code of the symbol string, knowledge information collection and It will be easier to organize.

The processing speed of this symbol string extraction device is symbol storage means 2]0
．． 4] Cycle time Tc of semiconductor RAM used for 0
corresponds approximately to the processing time for one symbol. If 1゛C is 100 m5, matching of 10 symbol strings against the text of 10 symbol strings can be performed in 10 seconds.

Since verification using current software takes about 10 hours, the symbol string verification device of the present invention significantly shortens the verification time.

In summary, according to the present invention, the drawbacks of long processing time and lack of flexibility due to symbol string processing 11 due to the combination of a conventional microcomputer and software can be easily solved. Considering that the symbol string identification device of the present invention can be easily assembled into a single chip, such an LSI can be used for extracting keywords from original text files, electronic dictionary for language translation, and slam recognition. It becomes an indispensable functional element in the classification of system feature series.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of an embodiment of the symbol string matching device according to the present invention, FIG. 2 is a timing diagram of Kuro 2248 and transfer signals, FIG. 3 is an explanatory diagram of the operation of the symbol string matching device, FIG. 45
The figure is an explanatory diagram of another embodiment of the present invention, and FIG. 6 is a symbol 1.
It is an explanatory diagram of collation. 1] ()...Symbol string storage means, 120...
・Counter, 123...First calculation means, ]2
4.-Temporary storage means, 130...Register, 14
0... Adder, 41o... Song comparator, '4
20...Level setting register, 500...
... Symbol string matching unit, 5]0... Symbol string matching section, 520... Encoder, 530...
...Tekota. Figure 2 MEMRYX ME7. lX0RYTo
T'ThT3T4T51 lower 2
Ding OT+ Ding 2 Ding 3 Ding 4 Ding 5 Ding 6- "i Figure 4 Figure 5 Figure 6

Claims (7)

[Claims] (1) Symbol string storage means that takes the symbol code as address input and stores the collation symbol string, and a first
a first computing means that interconnects the plurality of computing means;
A symbol string characterized by comprising a temporary storage means, a second temporary storage means connected to the output of a specific first temporary storage means, and a second calculation means connected to the second temporary storage means. Verification device. (2) The symbol string matching device according to claim 1, wherein the first calculation means and the first temporary storage means are each a counter or a shift register. (3) Symbol string storage means that takes the symbol code as address input and stores the collation symbol string, and a first
a first computing means that interconnects the plurality of computing means;
a temporary storage means, a second temporary storage means connected to the output of the specific first temporary storage means, a second calculation means connected to the second temporary storage means, and a permissible similarity given from the outside. register, the contents of this register, and the second
and a comparator that compares the output of the arithmetic means with the output of the arithmetic means and outputs the result of the comparison. (4) The symbol string matching device according to claim 3, wherein the first calculation means and the first temporary storage means are each a counter or a shift register. (5) Symbol string storage means that takes the symbol code as an address input and stores a collation symbol string, and a first
a first computing means that interconnects the plurality of computing means;
a temporary storage means, a second temporary storage means connected to the output of the specific first temporary storage means, a second calculation means connected to the second temporary storage means, and a permissible similarity given from the outside. register, the contents of this register, and the second
1. A symbol string matching device comprising: a comparator that compares the output of a calculation means with the output of a calculation means and outputs a matching result; an encoder connected to the comparator; and a decoder connected to the register. (6) The symbol string matching device according to claim 5, wherein the first calculation means and the first temporary storage means are each a counter or a shift register. (7) The Nth (positive integer) and (N+1) symbol string storage means that stores the collation string in the bit pattern associated with the symbol.
The contents of the Nth first temporary storage means between the Nth first logical calculation means and the (N+1)th first calculation means respectively connected to the output line of the (N-1)th first temporary storage means. It is assumed that the arithmetic processing of the Nth first arithmetic output stage controlled by the output line of the Nth symbol string storage means is performed on the data, and that the contents of the specific first temporary storage means at mutually different times are
A control method for a symbol string matching device, characterized in that a degree of similarity between the contents of the symbol string storage means and the sequentially input symbol strings is determined by performing arithmetic processing by the arithmetic means.