Correlate multiple notes in a database using bi-directional link
Disclosed are a method and device for calculating a correlation between notes using a database constructed on a basis of artificial intelligence, and supporting a service for the notes on a basis of the calculated correlation. A method by which a note providing device that interworks with a user terminal provides notes, includes: constructing a keyword DB by extracting a keyword from a note generated through the user terminal and reflecting a weight calculated through machine learning using the extracted keyword; and calculating a correlation score for each of a plurality of target notes correlated with a reference note using the keyword DB. Therefore, the method and device for providing the notes using the artificial intelligence-based correlation calculation can more accurately recommend the correlated notes by reflecting the interaction of the user.
1. A method performed by at least one processor comprising:
extracting, using the at least one processor, a keyword from a document;
generating, using the at least one processor, a first note from the extracted keyword in a database;
correlating, using the at least one processor, the first note with at least one second note in the database using a bi-directional link between the first note and the at least one second note.
2. The method of claim 1 , wherein the extracting the keyword is performed at a user terminal.
3. The method of claim 2 , wherein:
the generating the first note and is performed on at least one computer system separate from the user terminal;
the correlating the first note with the at least one second note is performed on the at least one computer system; and
the at least one computer system is network connected to the user terminal.
4. The method of claim 3 , wherein the at least one computer system is at least one server.
5. The method of claim 1 , wherein:
the correlating the first note comprises reflecting a weight calculated through machine learning using the extracted keyword; and
the method comprises calculating a correlation score for each of the first note and the plurality of second notes correlated with a reference note using the database.
6. The method of claim 5 , wherein the calculating the correlation score comprises:
calculating, using the at least one processor, a number of times that the extracted keyword is included in the reference note; and
associating, using the at least one processor, a weight to the extracted keyword based on the calculated number of times that the extracted keyword is included in the reference note.
7. The method of claim 6 , wherein the calculating the correlation score comprises applying the weight associated with the extracted keyword to a significance of the extracted keyword in the reference note.
8. The method of claim 5 , comprising providing, using the at least one processor, the plurality of second notes correlated with the reference note according to a ranking of the correlation score.
9. The method of claim 8 , comprising receiving feedback, using the at least one processor, through a user terminal on whether or not to accept each of the plurality of second notes according to the ranking of the correlation score.
10. The method of claim 9 , comprising updating, using the at least one processor, the weight associated with the extracted keyword according to acceptability of each of the plurality of second notes.
11. A system comprising:
at least one processor;
an extraction unit configured to extract a keyword from a document using the at least one processor;
a generation unit configured to generate a first note from the extracted keyword in a database using the at least one processor;
a correlation unit configured to correlate the first note with at least one second note in the database using a bi-directional link between the first note and the at least one second note using the at least one processor.
12. The system of claim 11 , comprising a user terminal, wherein the extraction unit is comprised in the user terminal.
13. The system of claim 12 , comprising at least one computer system, wherein:
the at least one computer system is separate from the user terminal;
the at least one computer system is network connected to the user terminal;
the generation unit is comprised in the at least one computer system; and
the correlation unit is comprised in the at least one computer system.
14. The system of claim 13 , wherein the at least one computer system comprises at least one server.
15. The system of claim 11 , wherein:
the correlation unit is configured to correlate the first note by reflecting a weight calculated through machine learning using the extracted keyword using the at least one processor; and
the correlation unit is configured to calculate a correlation score for each of the first note and the plurality of second notes correlated with a reference note using the database using the at least one processor.
16. The system of claim 15 , wherein the correlation score is calculated by:
calculating a number of times that the extracted keyword is included in the reference note using the at least one processor; and
associating a weight to the extracted keyword based on the calculated number of times that the extracted keyword is included in the reference note using the at least one processor.
17. The system of claim 16 , wherein the correlation score is calculated by applying the weight associated with the extracted keyword to a significance of the extracted keyword in the reference note using the at least one processor.
18. The system of claim 15 , wherein the plurality of second notes correlated with the reference note is provided according to a ranking of the correlation score using the at least one processor.
19. The system of claim 18 , wherein the system is configured to receive feedback through a user terminal on whether or not to accept each of the plurality of second notes according to the ranking of the correlation score using the at least one processor.
20. The system of claim 19 , the system is configured to update the weight associated with the extracted keyword according to acceptability of each of the plurality of second notes using the at least one processor.