Keyword extracting device

Claim 2

A computer implemented keyword extracting device comprising: a text data input part for inputting a text, a pattern storage part storing at least a pattern to generate keyword candidates, said pattern is represented by character strings in regular expression or its equivalent, an extracting part for extracting character strings from the text through pattern matching using said pattern stored in said pattern storage part, a keyword candidate generating part generating keyword candidates, the keyword candidates including at least portions of the character strings being extracted at the extracting part, and an output part outputting the keyword candidates as keywords, wherein the pattern to generate keyword candidates is an unnecessary pattern representing unnecessary head or end character strings, the keyword candidate generating part includes an unnecessary character string removing part, and said unnecessary character string removing part extracts keyword candidates including character strings matched with the unnecessary pattern from the keyword candidates so that (i) if the extracted keyword candidates are same as character strings matched with the unnecessary pattern, removes the extracted keyword candidates, and (ii) if the extracted keyword candidates are not same as character strings matched with the unnecessary pattern, takes the keyword candidates, in which said matched character string is removed, as a keyword candidates.

Claim 3

A computer implemented keyword extracting device of claim 1 , wherein the pattern storage part stores a keyword component pattern represented by character strings configured to be keyword components in regular expression or its equivalent, and the extracting part extracts character strings matched with the keyword component pattern through pattern matching while not extracting a character string overlapping with another character string already extracted.

Claim 4

A computer implemented keyword extracting device of claim 2 further comprising an extra keyword component extracting part; wherein the pattern to generate keyword candidates is an extra keyword component pattern, which represents keyword components hardly matched with the keyword component pattern, the extra keyword component extracting part extracts character strings matched with the extra keyword component pattern through pattern matching and replaces the extracted character strings in the text with special character strings, the extracting part extracts character strings matched with the keyword component pattern, from the text being replaced at the extra keyword component extracting part, and the keyword candidate generating part takes the character strings being extracted at the extra keyword component extracting part as keyword candidates as well as the character strings extracted at the extracting part.

Claim 5

A computer implemented keyword extracting device of claim 2 : wherein the keyword component pattern including patterns corresponding to type or field of text or kind of pattern the extracting part extracts character strings based on said patterns.

Claim 6

A computer implemented keyword extracting device of claim 4 : wherein the patterns includes patterns of a same kind, and the extracting part carries out the pattern matching applying said patterns of the same kind at a same time to extract a longest character string among character strings matched at a same headmost position in the text.

Claim 7

A computer implemented keyword extracting device of claim 5 further comprising a pattern integrating part for integrating the patterns to generate a new pattern of each kind, through a process including at least either of selection, concatenation, and repetition, while sorting components of the pattern in descending order of string lengths to be matched.

Claim 8

A computer implemented keyword extracting device of claim 1 further comprising a part-of-speech analyzing part for dividing the text into words and analyzing a part-of-speech of each divided word, wherein the extracting part extracts character strings corresponding to a prescribed sequence of part-of-speeches from the text, based on a result of the part-of-speech analysis at the part-of-speech analyzing part.

Claim 9

A computer implemented keyword extracting device of claim 1 , wherein the extracting part extracts the character strings with their positions in the text, and the keyword candidate generating part generates a keyword candidate by coupling the extracted character strings which are recognized as continuous with each other from their string length and positions in the text.

Claim 10

A computer implemented keyword extracting device of claim 1 : wherein the pattern to generate keyword candidates is an unnecessary word pattern representing unnecessary words, and the keyword candidate generating part includes an unnecessary word removing part, said unnecessary word removing part extracts character strings matched with the unnecessary word pattern from the text through pattern matching and, if same character strings thereof are included in the keyword candidates, removes the character strings from the keyword candidates.

Claim 11

A computer implemented keyword extracting device of claim 9 : wherein the pattern to generate keyword candidates further includes a necessary word pattern representing necessary words, and the unnecessary word removing part does not remove the character strings matched with the necessary word pattern.

Claim 12

A computer implemented keyword extracting device of claim 1 : wherein the pattern to generate keyword candidates further includes a necessary pattern representing necessary head or end character strings, and if the keyword candidate including character strings to be removed is a character strings matching with the necessary pattern, the unnecessary character string removing part does not perform said removal concerning with the keyword candidate.

Claim 13

A computer implemented keyword extracting device of claim 1 : wherein the pattern to generate keyword candidates includes patterns corresponding to type or field of text kind of pattern, and the keyword candidate generating part generates the keyword candidates based on the patterns.

Claim 14

A computer implemented keyword extracting device of claim 12 : wherein the patterns includes patterns of a same kind, and the keyword candidate generating part carries out the pattern matching, for the text or the keyword candidates, applying said patterns of the same kind at a same time to extract a longest character string among character strings matched at a same headmost position to generate keyword candidates based an the extracted character string.

Claim 15

A computer implemented keyword extracting device of claim 12 : wherein the patterns includes unnecessary word patterns or necessary word patterns of a same kind, and the keyword candidate generating part includes an unnecessary word removing part, said unnecessary word removing part carries out pattern matching using the unnecessary word patterns or the necessary word patterns at a same time for the text, and if any of the patterns is matched, the other patterns are not used and, if same character strings thereof are included in the keyword candidate, removes the character strings from the keyword candidate.

Claim 16

A computer implemented keyword extracting device of claim 12 further comprising a pattern integrating part for integrating the patterns to generate a new pattern of each kind, through a process including at least either of selection, concatenation, and repetition, while sorting components of the pattern in descending order of string lengths to be matched.

Claim 17

A computer implemented keyword extracting device of claim 15 : wherein the patterns include unnecessary word patterns or necessary word patterns of a same kind, and the pattern integrating part generates a new pattern having the patterns of the same kind as a selection choice with regard to the unnecessary word patterns and necessary word patterns.

Claim 18

A computer implemented keyword extracting device comprising: a text data input part for inputting a text. a pattern storage part storing at least a pattern to generate keyword candidates, said pattern is represented by character strings in regular expression or its equivalent, an extracting part for extracting character strings from the text through pattern matching using said pattern stored in said pattern storage part, a keyword candidate generating part generating keyword candidates, the keyword candidates including at least portions of the character strings being extracted at the extracting part, and a part-of-speech analyzing part for dividing the text into words and analyzing a part-of-speech of each divided word, wherein the extracting part extracts character strings corresponding to a prescribed sequence of part-of-speeches from the text, based on a result of the part-of-speech analysis at the part-of-speech analyzing part, the pattern to generate keyword candidates is an extra keyword component pattern, which represents keyword components hardly generated as character strings corresponding to the prescribed sequence of part-of-speeches, the extra keyword component extracting part extracts character strings matched with the extra keyword component pattern through pattern matching and replaces the extracted character strings in the text with special character strings, the extracting part extracts character strings corresponding to the prescribed sequence of part-of-speeches from the text being replaced at the extra keyword component extracting part, based on a result of the part-of-speech analysis at the part-of-speech analyzing part, the keyword candidate generating part takes the character strings being extracted at the extra keyword component extracting part as the keyword candidates as well as the character strings extracted at the extracting part strings extracts at the extracting part, and an outputting part outputting the keyword candidates as keywords.

Claim 19

A computer implemented keyword extracting device comprising: a text data input part for inputting a text. a pattern storage part storing at least a pattern to generate keyword candidates, said pattern is represented by character strings in regular expression or its equivalent, an extracting part for extracting character strings from the text through pattern matching using said pattern stored in said pattern storage part, a keyword candidate generating part generating keyword candidates, the keyword candidates including at least portions of the character strings being extracted at the extracting part, and a part-of-speech analyzing part for dividing the text into words and analyzing a part-of-speech of each divided word, wherein the extracting part extracts character strings corresponding to a prescribed sequence of part-of-speeches from the text, based on a result of the part-of-speech analysis at the part-of-speech analyzing part, the pattern to generate keyword candidates is an extra keyword component pattern, which represents keyword components hardly generated as character strings corresponding to the prescribed sequence of part-of-speeches, the extra keyword component extracting part extracts character strings matched with the extra keyword component pattern through pattern matching and replaces the extracted character strings in the text with special character strings, the extracting part extracts character strings corresponding to the prescribed sequence of part-of-speeches from the text being replaced at the extra keyword component extracting part, based on a result of the part-of-speech analysis at the part-of-speech analyzing part, the keyword candidate generating part takes the character strings being extracted at the extra keyword component extracting part as the keyword candidates as well as the character strings extracted at the extracting part strings extracts at the extracting part, and an outputting part outputting the keyword candidates as keywords.