Conversational-type natural language analysis apparatus

• US 5,029,085 A
• Filed: 05/10/1990
• Issued: 07/02/1991
• Est. Priority Date: 05/18/1989
• Status: Expired due to Fees

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: Ito, Hideo
• Primary Examiner(s): Fleming, Michael R.
• Assistant Examiner(s): CHUNG TRANS, XUONG MY
• Application Number:
  US07/521,559
• Time in Patent Office: 418 Days
• Field of Search: 364/419, 364/900, 364/917.92, 364/200, 364/274.8
• US Class Current: 704/9
• CPC Class Codes: G10L 15/193 Formal grammars, e.g. finite state automata, context free grammars or word networks

Abstract

A conversational-type of language analysis apparatus includes a sentence input unit inputting a sentence, a dictionary unit and a grammar storage unit. The apparatus also includes an all-relationship detection unit dividing into the syntactical units the sentence, the syntactical units being nodes, and detecting, as candidate relationships, all linguistic relationships on the basis of the grammatical rules in the grammar storage units and dictionary information in the dictionary unit, a conversational-type analysis unit selecting correct relationships by conversation interaction with a user from the candidate relationships, the conversational-type analysis unit having conversational object selecting unit for selecting object relationships which are objects of the conversational interaction with the user from the candidate relationships, and conversation unit selecting the correct relationships by conversational interaction with the user. The apparatus also has analysis unit selecting correct relationships without the conversational interaction from relationships other than the correct relationships selected by the conversational-type analysis unit.

Claims

1. A conversational-type of language analysis apparatus comprising:

(a) sentence input means for inputting a sentence in a language which is an object of analysis;

(b) dictionary means for having an attribute with respect to a syntactical unit of said language, as dictionary information;

(c) grammar storage means for having grammatical rules which are applied to said language;

(d) all-relationship detection means for dividing into said syntactical units said sentence input by said sentence input means, said syntactical units being nodes, and detecting, as candidate relationships, all linguistic relationships which have a possibility of existing between said nodes on the basis of said grammatical rules in said grammar storage means and said dictionary information in said dictionary means;

(e) conversational-type analysis means for selecting first correct relationships by conversational interaction with a user, from said candidate relationships detected by said all-relationship detection means, said conversational-type analysis means including;

(1) conversational object selection means for selecting object relationships which satisfy a predetermined condition so that the object relationships are offered for the conversational interaction with said user, from said candidate relationships, and

(2) conversation means for selecting said first correct relationships by the conversational interaction with said user from said object relationships; and

(f) analysis means for selecting second correct relationships without said conversational interaction from the candidate relationships other than said first correct relationships selected by said conversational type analysis means from said candidate relationships.

2. A conversational-type of language analysis apparatus as claimed in claim 1 wherein said conversational object selecting means includes:

scoring means for giving a score to each of candidate relationships detected by said all-relationship detection means, said score having a value corresponding to a statistical probability that each of the candidate relationships detected by said all-relationship detection means is established,

grouping means for dividing into groups said candidate relationships detected by said all-relationship detection means said candidate relationships belonging to each of said groups being exclusive each other in each group, and

means for selecting object groups, each of said object groups in which the difference between a highest score and a next highest score, among scores which are given to said candidate relationships belonging to each of said object groups, is less than or equal to a predetermined value, from said candidate relationships belonging to each of said object groups one of said first correct relationships being selected by said conversation means.

3. A conversational-type of language analysis apparatus as claimed in claim 2 wherein said candidate relationships detected by said all-relationship detection means includes a grammatical rule which has a possibility of being established between said nodes, said scoring means has calculation means for calculating a statistical probability that said grammatical rule is established between said nodes.

4. A conversational-type of language analysis apparatus as claimed in claim 2 wherein each of said candidate relationships detected by said all-relationship detecting means includes linguistic attributes of said nodes and a grammatical rule which has a possibility of being between said nodes, said scoring means has second calculation means for calculating a score in accordance with a predetermined formula using a statistical probability that said grammatical rule is established between said nodes and a statistical probability that each of said nodes has a corresponding linguistic attribute.

5. A conversational-type language analysis apparatus as claimed in claim 2 wherein said conversational means selects said first correct relationships starting from one of said object groups which has the smallest difference between the highest score and the next highest score.

6. A conversational-type language analysis apparatus as claimed in claim 2 wherein said conversational object selecting means has elimination means for eliminating relationships which are exclusive to one of said first correct relationships selected by said conversation means from said candidate relationships which are object of said conversational interaction with said user, and said scoring means again giving a score to each of said candidate relationships from which said relationships are eliminated.

7. A conversational-type of language analysis apparatus as claimed in claim 1 wherein said conversation means has determination means for determining the number of times of said conversational interaction with said user.

8. A conversational-type language analysis apparatus as claimed in claim 1 wherein said analysis means includes:

second scoring means for giving a second score to each of said candidate relationships other than said first correct relationships selected by said conversational-type analysis means, said second score having a value corresponding to a statistical probability that each of said candidate relationships other than said first correct relationships is established,

second grouping means for dividing into second groups said candidate relationships other than said first correct relationships selected by said conversational-type analysis means, said candidate relationships belonging to each of said second groups being exclusive to each other in each second group, and

selecting means for selecting one of said second correct relationships which has the largest second score among said candidate relationships belonging to each of said second groups.

9. A conversational-type of language analysis apparatus as claimed in claim 8 wherein said selecting means selects one of said second correct relationships starting from one of said second groups which has the largest difference between highest second score and the next highest second score.

10. A conversational-type of language analysis apparatus as claimed in claim 8 wherein said analysis means includes:

second elimination means for eliminating relationships which are exclusive to one of said second correct relationships selected by said selecting means from said candidate relationships, said second scoring means again giving a second score to each of said relationships.

11. A conversational-type language analysis apparatus as claimed in claim 1 wherein said analysis means includes:

third scoring means for giving a third score to each of said candidate relationships other than said first correct relationships selected by said conversational-type analysis means, said third score having a value corresponding to statistical probability that each of said candidate relationships other than said first correct relationships is established,

third grouping means for dividing into third groups said candidate relationships other than said first correct relationships selected by said conversational-type analysis means, said candidate relationships belonging to each of said third groups being inclusive in each third group,

total calculation means for calculating a total score of second scores given to said candidate relationships belonging to each of said third groups, and

second selecting means for selecting one of said third groups which has largest total score, said candidate relationships belonging to said one of said third groups being said second correct relationships.