JP2015035026A - Information processing device, information processing method, and information processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processing device, an information processing method, and an information processing program that can improve convenience to a user when reporting living-up.SOLUTION: An information processing device includes: a keyword extraction part which extracts a keyword from a document posted within a predetermined time; a document extraction part which extracts a representative document as a representative from one group of posted documents; and a word order change part which rearranges keywords extracted by the keyword extraction part in the order of words appearing in the representative document extracted by the document extraction part.

Description

  The present invention relates to an information processing apparatus, an information processing method, and an information processing program.

  There is one that calculates the degree of excitement of the TV program being broadcast from the number of messages posted to Twitter (registered trademark) and notifies the user of the degree of excitement (see Non-Patent Document 1). Here, the excitement is a state in which the number of posted messages at the current time exceeds a predetermined number and the number of posted messages at the current time is larger than the number of posted messages at the same time. In order to enjoy recorded TV programs efficiently, important scenes are automatically detected from fluctuations in the number of posted messages related to TV programs, and the main characters and event contents in each scene are estimated from the contents of posted messages. However, some labels are given as scene labels (see Non-Patent Document 2).

“Teach me remote control”, [online], Sharp Corporation, [searched July 1, 2013], Internet <URL: https://play.google.com/store/apps/details? id = jp.co.sharp.clpf.android.cloudremocon> Masami Nakazawa, Keiichiro Hoashi, Tomohiro Ono, “Twitter TV Program Important Scene Detection and Labeling Technique”, Proceedings of the 3rd Forum on Data Engineering and Information Management, F5-6, February 2011.

  However, even if the technology described in Non-Patent Document 1 is used or the degree of climax is notified to the user, there is a drawback that the reason for the climax is not known. Moreover, even if the technique described in Non-Patent Document 2 is used, there is a drawback that the contents of events that are not related to the main person cannot be estimated. As described above, there is a problem that the convenience of the user when notifying the excitement is not sufficient.

  The present invention has been made in view of the above points, and it is an object of the present invention to provide an information processing apparatus, an information processing method, and an information processing program that can improve user convenience when notifying a rise. To do.

  (1) The present invention has been made to solve the above-described problems, and a first aspect of the present invention is an information processing apparatus that extracts keywords from sentences posted within a predetermined time. An extraction unit; a sentence extraction unit that extracts a representative sentence from the group of the posted sentences; and a word that appears in the representative sentence extracted by the sentence extraction unit by the keyword extracted by the keyword extraction unit An information processing apparatus comprising: a word order changing unit for rearranging in the order of

  (2) Moreover, the second aspect of the present invention is an information processing method, comprising: a keyword extraction process for extracting a keyword from a sentence posted within a predetermined time; and a representative from a group of the posted sentences. A sentence extraction process for extracting a representative sentence, and a word order changing process for rearranging the keywords extracted in the keyword extraction process in the order of words appearing in the representative sentence extracted in the sentence extraction process. Is an information processing method characterized by

  (3) Further, a third aspect of the present invention is an information processing program, the keyword extracting step for extracting a keyword from a sentence posted within a predetermined time to a computer of the information processing apparatus, and the posted A sentence extraction step for extracting a representative sentence from the group of sentences, and a word order for rearranging the keywords extracted in the keyword extraction step into the order of words appearing in the representative sentence extracted in the sentence extraction step An information processing program for executing the change step.

  According to the information processing apparatus, the information processing method, and the information processing program of the present invention, it is possible to improve the convenience of the user when notifying the excitement.

It is the schematic which shows an example of a structure of the information processing system which concerns on the 1st Embodiment of this invention. It is a schematic block diagram which shows an example of a structure of the information processing apparatus which concerns on the 1st Embodiment of this invention. It is the schematic which shows an example of a structure of the detection part and extraction part which concern on the 1st Embodiment of this invention. It is a schematic block diagram which shows an example of a structure of the analysis part which concerns on the 1st Embodiment of this invention. It is explanatory drawing explaining an example of a process of the keyword extraction part which concerns on the 1st Embodiment of this invention, a weight calculation part, and a weighting part. It is explanatory drawing explaining an example of a process of the text extraction part which concerns on the 1st Embodiment of this invention, a word order change part, and an output part. It is a flowchart which shows an example of the process of the analysis part which concerns on the 1st Embodiment of this invention. It is a schematic block diagram which shows an example of a structure of the analysis part which concerns on the 2nd Embodiment of this invention. It is explanatory drawing explaining an example of the process of the keyword extraction part which concerns on the 2nd Embodiment of this invention, a weight calculation part, and a weighting part. It is explanatory drawing explaining an example of a process of the text extraction part which concerns on the 2nd Embodiment of this invention, a word order change part, and an output part. It is a schematic block diagram which shows an example of a structure of the analysis part which concerns on the 3rd Embodiment of this invention. It is explanatory drawing explaining an example of a process of the text extraction part which concerns on the 3rd Embodiment of this invention, a word order change part, and an output part. It is a schematic block diagram which shows an example of a structure of the analysis part which concerns on the 4th Embodiment of this invention.

(First embodiment)
Hereinafter, a first embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a schematic diagram showing an example of the configuration of the information processing system S1 according to the first embodiment of the present invention.
The information processing system S1 includes a posting terminal device 1, a posting service server device 2, an information processing device 3, and an electronic device 4. Each of posting terminal device 1, posting service server device 2, information processing device 3, and electronic device 4 performs wired or wireless communication via network N1.

  The posting terminal device 1 is, for example, a PC (Personal Computer) and receives a message input from the user. The posting service server device 2 is, for example, a social media server device. Here, social media is a community in which information transmission between individuals and individuals, individuals and organizations, and organizations and organizations is meaningful by itself via web services, and is widely spread and influenced in the real world. For example, blogs, miniblogs, electronic bulletin boards, social network services, posting services, etc.

  The user posts a message to the posting service server device 2 using the posting terminal device 1. In the following description, a message posted to the posting service server device 2 is referred to as a posting message.

The information processing device 3 acquires a posted message from the posting service server device 2 via the network N1. Then, the information processing apparatus 3 extracts a plurality of keywords from a plurality of posted messages within a predetermined time, and extracts a representative message as a representative from the plurality of posted messages. The information processing device 3 rearranges the extracted keywords in the order of words appearing in the representative message, and transmits them to the electronic device 4 via the network N1. The display unit 41 of the electronic device 4 displays the keywords rearranged in the order of representative messages.
The electronic device 4 is, for example, a PC, a multi-function mobile phone (smart phone) with a PDA (personal digital assistant) function in addition to a telephone or mail function, a tablet PC, or the like.

FIG. 2 is a schematic block diagram showing an example of the configuration of the information processing apparatus 3 according to the first embodiment of the present invention.
The information processing apparatus 3 includes an acquisition unit 31, a storage unit 32, a detection unit 33, and an extraction unit 34.
The acquisition unit 31 acquires a plurality of posted message information from the posting service server device 2 using an API (Application Program Interface). The posted message information includes the message body of the posted message and the time when the message was posted to the posting service server device 2. The acquisition unit 31 stores the acquired posted message information in the storage unit 32, and the storage unit 32 stores the posted message information.

The detection unit 33 detects the number of posted messages within a predetermined time, and operates the extraction unit 34 based on the detection result. In addition, the detection unit 33 transmits the detection result and keyword information input from the extraction unit 34 described later to the electronic device 4 via the network N1. Detailed processing of the detection unit 33 will be described later.
The extraction unit 34 extracts a plurality of keywords from the posted message. The extraction unit 34 rearranges the extracted keywords to generate keyword information. The extraction unit 34 outputs the generated keyword information to the detection unit 33. Detailed processing of the extraction unit 34 will be described later.

FIG. 3 is a schematic diagram illustrating an example of the configuration of the detection unit 33 and the extraction unit 34 according to the first embodiment of the present invention.
The detection unit 33 includes a detection unit 331 and a notification unit 332.
The detection unit 331 detects the number of posted messages within a predetermined time, and detects the swell when the number of posted messages detected within the predetermined time is equal to or greater than the predetermined number, and operates the extraction unit 34. In addition, the detection unit 331 outputs excitement information indicating the detected excitement to the notification unit 332.
The notification unit 332 notifies the electronic device 4 of the climax information input from the detection unit 331 and the keyword information input from the analysis unit 342 via the network N1.

The extraction unit 34 includes a reading unit 341 and an analysis unit 342.
When the detecting unit 331 detects the excitement, the reading unit 341 reads posted message information within a predetermined time from the storage unit 32 and outputs the posted message information to the analyzing unit 342.
The analysis unit 342 extracts a plurality of keywords from the posted message text included in the posted message information. The analysis unit 342 extracts keywords using, for example, known means. The analysis unit 342 extracts a representative message based on the extracted keyword. Then, the analysis unit 342 generates keyword information by rearranging the extracted keywords in the order of words appearing in the representative message text. The analysis unit 342 outputs the generated keyword information to the notification unit 332. Detailed processing of the analysis unit 342 will be described later.

FIG. 4 is a schematic block diagram illustrating an example of the configuration of the analysis unit 342 according to the first embodiment of the present invention.
The analysis unit 342 includes a keyword extraction unit 3421, a weight calculation unit 3422, a weighting unit 3423, a sentence extraction unit 3424, a word order change unit 3425, and an output unit 3426.
The keyword extracting unit 3421 extracts a keyword from each posted message body of a plurality of posted messages included in the posted message information input from the reading unit 341. The keyword extraction unit 3421 outputs the extracted keyword to the weight calculation unit 3422.

  The weight calculation unit 3422 counts the number of appearances of the keyword extracted by the keyword extraction unit 3421 for the posted message. Further, the weight calculation unit 3422 counts the number of appearances of the keyword counted for each posted message for the posted messages within a predetermined time. Then, the weight calculation unit 3422 calculates weights for the tabulated keywords. The weight calculation unit 3422 outputs the keyword extracted by the keyword extraction unit 3421 and the weight calculated by the weight calculation unit 3422 for the keyword to the weighting unit 3423. Further, the weight calculation unit 3422 outputs the counted number of appearances of the keyword to the text extraction unit 3424.

  The weighting unit 3423 assigns a weight to the keyword based on the keyword input from the weight calculation unit 3422 and the weight for the keyword. The weighting unit 3423 outputs the weighted keyword to the sentence extracting unit 3424 and the word order changing unit 3425.

  The text extraction unit 3424 determines the number of posted messages from the posted message texts of a plurality of posted messages included in the posted message information input from the reading unit 341, based on the keywords to which the weights input from the weighting unit 3423 are assigned. Is calculated. The posted message score is calculated from the weight assigned to the keyword and the number of appearances of the keyword counted for each posted message. Here, the posted message score is the sum of the element scores, and the element score is a score calculated by multiplying the number of appearances of each keyword by the weight of each keyword. The text extraction unit 3424 extracts a posted message having the highest posted message score from a plurality of posted messages as a representative message based on the calculated posted message score. The sentence extracting unit 3424 outputs the extracted representative message to the word order changing unit 3425.

The word order changing unit 3425 causes the word order of the weighted keywords to appear in the representative message based on the weighted keyword input from the weighting unit 3423 and the representative message input from the text extracting unit 3424. The keyword information is generated by rearranging the words in the order. The word order changing unit 3425 outputs the generated keyword information to the output unit 3426.
The output unit 3426 outputs the keyword information input from the word order change unit 3425 to the notification unit 332.

FIG. 5 is an explanatory diagram illustrating an example of processing of the keyword extraction unit 3421, the weight calculation unit 3422, and the weighting unit 3423 according to the first embodiment of the present invention.
In the processing content example Ex1, the keyword extraction unit 3421 extracts keywords from a plurality of posted message texts included in the posted message information. For example, the keyword extraction unit 3421 determines that the posting message “[ABC Opening Game Japan vs. Thailand] Game End Japan 5-3 Thailand Japan National Team is suffering, but it ’s a great reversal victory!” From “ABC Opening Game”, “Japan vs. Thailand”, “ The keywords “End of game”, “Japan”, “5-3”, “Thailand”, “Japan national team”, “Hard”, “Superb”, and “Reversal victory” are extracted.

  The weight calculation unit 3422 counts the number of appearances of the keyword extracted by the keyword extraction unit 3421 for the posted message. For example, the weight calculation unit 3422 determines the number of appearances of each extracted keyword as “ABC opening game [1]”, “Japan vs. Thailand [1]”, “Game end [1]”, “Japan [1]”, “5-3 [1]”, “Thai [1]”, “Japan National Team [1]”, “Bad [1]”, “Superb [1]”, “Reversal victory [1]” are counted. Here, the number in [] given to each keyword represents the appearance frequency of the keyword.

  The weight calculation unit 3422 uses posted messages within a predetermined time (for example, 1 minute) as extracted frames, and totals keywords for the posted messages in the extracted frames. In the processing content example Ex2, for example, the weight calculation unit 3422 sets the keywords in the extracted frame to “win [31]”, “Japan national team [22]”, “Japan [14]”, “match end [13]”. , “Thailand [7]”, “Victory [6]”, “First Battle [5]”, “5-3 [5]”, “Bad [4]”, “Japan Victory [4]”, And sum up. Here, the number in [] given to each keyword represents the number of posted messages in the extracted frame in which the keyword appears.

  The weight calculation unit 3422 uses a posted message within a predetermined time (for example, 10 minutes) as a reference frame, extracts the same keyword as each keyword extracted in the extracted frame from the posted message in the reference frame, and tabulates it. In the processing content example Ex3, for example, the weight calculation unit 3422 determines each keyword in the reference frame as “win [5]”, “Japan national team [46]”, “Japan [4]”, “match end [0]”. ”,“ Thailand [3] ”,“ Victory [0] ”,“ First Battle [0] ”,“ 5-3 [1] ”,“ Bad [0] ”,“ Japan Victory [0] ”, etc. , And extract and aggregate. Here, the number in [] given to each keyword represents the number of posted messages in the reference frame in which the keyword appears.

  Based on the number of posted messages in the extracted frame and the number of posted messages in the reference frame, the weight calculating unit 3422 calculates the weight for each keyword using Equation (1).

Here, Weight (w) represents a weight for each keyword. _{Also, r 0 (w), r} 10 (w) , the formula (2), is calculated by the equation (3).

Here, r ₀ (w) and r ₁₀ (w) represent the appearance rate of each keyword in the extracted frame and the appearance rate of each keyword in the reference frame, respectively, and x ₀ (w) and x ₁₀ (w) are This represents the number of posted messages in which each keyword appears in the extracted frame and the number of posted messages in which each keyword appears in the reference frame. N ₀ and N ₁₀ represent the total number of posted messages in the extracted frame and the total number of posted messages in the reference frame, respectively.

The weighting unit 3423 gives the weight for each keyword calculated by the weight calculation unit 3422 according to the expression (1) to the keyword as in the processing content example Ex4.
For example, the weighting unit 3423 may include “game end 3.10”, “win 1.69”, “win 0.71”, “first match 0.51”, “Japan 0.45”, “sad 0.34”, Weights are given to keywords such as “Japan win 0.34”, “5-3 0.26”, “Thailand 0.14”, “Japan national team 0.13”.

  Thereby, a low weight is assigned to a keyword that frequently appears in a posted message in the reference frame, and a high weight is assigned to a keyword that does not appear (not appear) in the posted message in the reference frame. In this example, the weight of keywords such as “End of game”, “Victory”, “First game” is high, and the weight for “Japan national team” is low.

FIG. 6 is an explanatory diagram illustrating an example of processing of the text extraction unit 3424, the word order change unit 3425, and the output unit 3426 according to the first embodiment of the present invention.
The processing content example Ex5 is an example of a keyword given a weight by the weighting unit 3423. The text extracting unit 3424 calculates the element score of the posted message of the extracted frame based on the keyword given the weight by the weighting unit 3423. Specifically, the sentence extraction unit 3424 calculates the element score by multiplying the number of appearances of each keyword in the extracted frame counted by the weighting unit 3423 by the weight of each keyword. The text extraction unit 3424 calculates the total sum of the calculated element scores as a posted message score for each posted message.

In the processing content example Ex6, the text extraction unit 3424 rearranges each posted message in the extracted frame in descending order of the number of posted messages, and extracts a predetermined number, for example, the top five posted messages. Then, the text extraction unit 3424 extracts a posted message having the highest calculated posted message score from the top five posted messages as a representative message.
For example, the text extraction unit 3424, like the processing example Ex7, posted the message “ABC game end qualifying round Japan 5-3 Thailand * 7 times, Japan took 3 points at a time and won the opening game” Extracted as a representative message.

The word order changing unit 3425 generates keyword information by rearranging the word order of the keywords weighted by the weighting unit 3423 into the order of words appearing in the representative message. For example, the processing content example Ex8 is an example of keyword information generated by the word order changing unit 3425 and output from the output unit 3426 to the notification unit 332.
The word order change unit 3425 changes the word order of each keyword to which the weight of the processing content example Ex5 is given, as in the processing content example Ex8, with the keywords “game end”, “Japan”, “5-3”, “ Rearrange with “Victory” to generate keyword information.

FIG. 7 is a flowchart showing an example of processing of the analysis unit 342 according to the first embodiment of the present invention.
In step ST101, the keyword extraction unit 3421 extracts keywords from each posted message in the extraction frame.
In step ST102, the weight calculation unit 3422 calculates the weight for each keyword from the keyword appearance rate of the posted message in the extracted frame and the keyword appearance rate of the posted message in the reference frame.

In step ST103, the weighting unit 3423 assigns a weight for each keyword calculated by the weight calculation unit 3422 to each keyword extracted by the keyword extraction unit 3421.
In step ST104, the text extraction unit 3424 calculates the element score and the posted message score from the weight for each keyword and the number of appearances of the keyword, and extracts the representative message based on the calculation result.

  In step ST105, the word order changing unit 3425 rearranges the word order of the keywords extracted by the keyword extracting unit 3421 into the order of words appearing in the representative message, and generates keyword information.

  As described above, according to the present embodiment, the information processing apparatus 3 includes the keyword extraction unit 3421 that extracts a keyword from a sentence (post message) posted within a predetermined time, and a group of posted sentences (post message). A text extraction unit 3424 that extracts representative text (representative message) from (post message in the extraction frame) and a keyword extracted by the keyword extraction unit 3421 appear in the representative text (representative message) extracted by the text extraction unit 3424. And a word order changing unit 3425 for rearranging the words in order.

  Accordingly, the information processing device 3 can extract keywords from a plurality of posted sentences and rearrange the extracted keywords in the order of words appearing in the representative sentences. You can follow the topic of the message. Moreover, since the information processing apparatus 3 can notify the user of the reason for the excitement by following the topicality, the convenience when notifying the user of the excitement can be improved.

(Second Embodiment)
Hereinafter, a second embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 8 is a schematic block diagram showing an example of the configuration of the analysis unit 342A according to the second embodiment of the present invention.
The analysis unit 342A includes a keyword extraction unit 3421, a weight calculation unit 3422, a weighting unit 3423, a sentence extraction unit 3424A, a word order change unit 3425A, and an output unit 3426A. When the analysis unit 342 according to the first embodiment and the analysis unit 342A according to the second embodiment are compared, the sentence extraction unit 3424A, the word order change unit 3425A, and the output unit 3426A are different. Since the other configuration is the same as that of the first embodiment, the same reference numerals are given and description thereof is omitted.

  The sentence extraction unit 3424A obtains the number of posted messages from the posted message texts of a plurality of posted messages included in the posted message information input from the reading unit 341 based on the weighted keyword input from the weighting unit 3423. Is calculated. The posted message score is calculated from the weight assigned to the keyword and the number of appearances of the keyword counted for each posted message. First, the text extraction unit 3424A extracts a posted message having the highest posted message score from a plurality of posted messages as a first representative message based on the calculated posted message score. The sentence extracting unit 3424A outputs the extracted representative message to the word order changing unit 3425A. Next, the text extracting unit 3424A extracts a posted message having the second highest posted message score from the plurality of posted messages as a second representative message based on the calculated posted message score. The sentence extracting unit 3424 outputs the extracted representative message to the word order changing unit 3425A.

The word order changing unit 3425A determines the first word order of the weighted keywords based on the weighted keyword input from the weighting unit 3423 and the first representative message input from the text extracting unit 3424A. The keyword information is generated by rearranging in the order of words appearing in the representative message. Next, the word order changing unit 3425A determines the word order of the weighted keywords based on the weighted keyword input from the weighting unit 3423 and the second representative message input from the text extracting unit 3424A. The keyword information is generated by rearranging in the order of words appearing in the second representative message. Word order changing unit 3425A outputs the generated keyword information to output unit 3426A.
The output unit 3426A outputs the keyword information input from the word order change unit 3425A to the notification unit 332.

FIG. 9 is an explanatory diagram illustrating an example of processing of the keyword extraction unit 3421, the weight calculation unit 3422, and the weighting unit 3423 according to the second embodiment of the present invention.
In the processing content example Ex9, the keyword extraction unit 3421 extracts keywords from a plurality of posted message texts included in the posted message information. For example, the keyword extraction unit 3421 reads from the posted message “Earthquake early warning sounds → Wakes up → Starts shaking → A variety of checks are being performed at the broadcasting station A” to “Earthquake Early Warning”, “Sounds”, “Wakes up”, “ The keywords “shaking”, “beginning”, “A broadcasting station”, and “under various checks” are extracted.

  The weight calculation unit 3422 counts the number of appearances of the keyword extracted by the keyword extraction unit 3421 for the posted message. For example, the weight calculation unit 3422 displays the number of appearances of each extracted keyword as “Earthquake Early Warning [1]”, “Sound [1]”, “Wake [1]”, “Swing [1]”, “Starts”. [1] "," A broadcasting station [1] ", and" various checking [1] "are counted. Here, the number in [] given to each keyword represents the appearance frequency of the keyword.

  The weight calculation unit 3422 uses posted messages within a predetermined time (for example, 1 minute) as extracted frames, and totals keywords for the posted messages in the extracted frames. In the processing content example Ex10, for example, the weight calculation unit 3422 sets the keywords in the extracted frame as “tsunami concern [40]”, “earthquake [18]”, “see [16]”, “attach [15]”, “□□ Oshima [14]”, “High seismic intensity 5 [12]”, “Earthquake early warning [8]”, and so on. Here, the number in [] given to each keyword represents the number of posted messages in the extracted frame in which the keyword appears.

  The weight calculation unit 3422 uses a posted message within a predetermined time (for example, 10 minutes) as a reference frame, extracts the same keyword as each keyword extracted in the extracted frame from the posted message in the reference frame, and tabulates it. In the processing content example Ex11, for example, the weight calculation unit 3422 displays the keywords in the reference frame as “Tsunami Concern [21]”, “Earthquake [34]”, “View [36]”, and “Add [61]”. , “□□ Oshima [59]”, “Seismic intensity 5 strong [39]”, “Earthquake Early Warning [22]”, and so on. Here, the number in [] given to each keyword represents the number of posted messages in the reference frame in which the keyword appears.

Based on the number of posted messages in the extracted frame and the number of posted messages in the reference frame, the weight calculating unit 3422 calculates the weight for each keyword using the above equation (1).
The weighting unit 3423 weights the keyword for each keyword calculated by the weight calculation unit 3422 according to the expression (1) as in the processing content example Ex12.
For example, the weighting unit 3423 includes “tsunami worry 0.35”, “earthquake 0.05”, “view 0.03”, “seismic intensity 5 strong 0.02”, “attach 0.02”, “□□ Oshima 0.02 ”,“ Earthquake Early Warning 0.01 ”,...

FIG. 10 is an explanatory diagram illustrating an example of processing of the text extraction unit 3424A, the word order change unit 3425A, and the output unit 3426A according to the second embodiment of the present invention.
The processing content example Ex13 is an example of a keyword given a weight by the weighting unit 3423. The text extraction unit 3424A calculates the number of elements of the post message in the extracted frame based on the keyword given the weight by the weighting unit 3423. Specifically, the sentence extracting unit 3424A calculates the number of element points by multiplying the number of appearances of each keyword in the extracted frame counted by the weighting unit 3423 by the weight of each keyword. The text extraction unit 3424A calculates the total sum of the calculated element scores as a posted message score for each posted message.

In the processing content example Ex14, the text extraction unit 3424A sorts the posted messages in the extracted frame in descending order of the posted message score, and extracts a predetermined number, for example, the top five posted messages. Then, the text extraction unit 3424A extracts the posted message having the highest calculated posted message score from the top five posted messages as the first representative message.
For example, the sentence extraction unit 3424A, as shown in the processing example Ex15, the posted message “around 5:33, a very strong earthquake with the epicenter of YYＸ in XX prefecture. No worries about tsunami "is extracted as the first representative message.

In addition, the sentence extraction unit 3424A extracts a posted message having the second highest calculated posted message score from the top five posted messages as the second representative message.
For example, the sentence extraction unit 3424A, like the processing content example Ex16, posted the message “Are you okay? Now on TV, you are not worried about the tsunami. Extracted as a representative message.

The word order changing unit 3425A rearranges the word order of the keywords weighted by the weighting unit 3423 into the order of words appearing in the first representative message, and generates keyword information on the first line. Next, the word order changing unit 3425A rearranges the word order of the keywords weighted by the weighting unit 3423 into the order of the words appearing in the second representative message, and generates keyword information on the second line. For example, the processing content example Ex17 is an example of keyword information generated by the word order changing unit 3425A and output from the output unit 3426A to the notification unit 332.
The word order changing unit 3425A, as shown in the processing content example Ex17 based on the first representative message, indicates the word order of each keyword to which a weight is given as in the processing content example Ex13, as shown in the processing content example Ex17. ”,“ Seismic intensity 5 strong ”,“ Tsunami worry ”, line feed, and on the second line, as shown in the processing example Ex17 based on the second representative message, the keyword“ Tsunami worry ”, The keyword information is generated by rearranging “Seismic intensity 5+”.

  As described above, according to the present embodiment, the information processing apparatus 3A includes the keyword extracting unit 3421 that extracts a keyword from a sentence (post message) posted within a predetermined time, and a group of posted sentences (post message). A sentence extracting unit 3424A that extracts representative sentences (first representative message and second representative message) from (post message in the extraction frame) and a keyword extracted by the keyword extracting unit 3421 are extracted by the sentence extracting unit 3424A. And a word order changing unit 3425A for rearranging the words in the order of words appearing in the representative sentences (first representative message, second representative message).

  Accordingly, the information processing apparatus 3A can extract keywords from a plurality of posted sentences and rearrange the extracted keywords in the order of words appearing in the representative sentences. You can follow the topic of the message. Moreover, since the information processing apparatus 3A can notify the user of the reason for the excitement by following the topicality, the convenience when notifying the user of the excitement can be improved.

(Third embodiment)
Hereinafter, the third embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 11 is a schematic block diagram illustrating an example of the configuration of the analysis unit 342B according to the third embodiment of the present invention.
The analysis unit 342B includes a keyword extraction unit 3421, a weight calculation unit 3422B, a weighting unit 3423B, a sentence extraction unit 3424B, a word order change unit 3425B, and an output unit 3426B. When the analysis unit 342 according to the first embodiment and the analysis unit 342B according to the third embodiment are compared, the weight calculation unit 3422B, the weighting unit 3423B, the sentence extraction unit 3424B, the word order changing unit 3425B, and the output unit 3426B are different. . Since the other configuration is the same as that of the first embodiment, the same reference numerals are given and description thereof is omitted.

The weight calculation unit 3422B counts the number of appearances of the keyword extracted by the keyword extraction unit 3421 for the posted message. Further, the weight calculation unit 3422B totals the number of appearances of the keyword counted for each posted message for the posted messages within a predetermined time. Then, the weight calculation unit 3422B calculates weights for the tabulated keywords. The weight calculation unit 3422B outputs the keyword extracted by the keyword extraction unit 3421 and the weight calculated by the weight calculation unit 3422B for the keyword to the weighting unit 3423B.
Then, when the keyword information is input from the word order changing unit 3425B, the weight calculating unit 3422B excludes the weights for the keywords included in the keyword information. The weight calculation unit 3422B outputs, to the weighting unit 3423B, a keyword obtained by excluding the keyword included in the keyword information input from the word order change unit 3425B from the keyword input from the keyword extraction unit 3421 and the weight for the keyword.

  The weighting unit 3423B gives a weight to the keyword based on the keyword input from the weight calculation unit 3422B and the weight for the keyword. The weighting unit 3423B outputs the weighted keyword to the text extracting unit 3424B and the word order changing unit 3425B.

  The sentence extraction unit 3424B obtains the number of posted messages from the posted message texts of a plurality of posted messages included in the posted message information input from the reading unit 341, based on the keyword to which the weight input from the weighting unit 3423B is assigned. Is calculated. The posted message score is calculated from the weight assigned to the keyword and the number of appearances of the keyword counted for each posted message. First, the text extraction unit 3424B extracts a posted message having the highest posted message score from a plurality of posted messages as a first representative message based on the calculated posted message score. The sentence extracting unit 3424B outputs the extracted first representative message to the word order changing unit 3425B. In addition, the text extraction unit 3424B calculates again the posted message score based on the keyword to which the weight input from the weighting unit 3423B is given. Then, based on the recalculated post message score, the text extraction unit 3424B extracts, as a second representative message, a post message having the highest post message score other than the first representative message from the plurality of post messages. To do. The sentence extracting unit 3424B outputs the extracted second representative message to the word order changing unit 3425B.

The word order changing unit 3425B determines the first word order of the weighted keywords based on the weighted keyword input from the weighting unit 3423B and the first representative message input from the text extracting unit 3424B. The keyword information is generated by rearranging in the order of words appearing in the representative message. Next, the word order changing unit 3425B determines the word order of the weighted keywords based on the weighted keyword input from the weighting unit 3423B and the second representative message input from the text extracting unit 3424B. The keyword information is generated by rearranging in the order of words appearing in the second representative message. Word order changing unit 3425B outputs the generated keyword information to weight calculation unit 3422B and output unit 3426B.
The output unit 3426B outputs the keyword information input from the word order change unit 3425B to the notification unit 332.

FIG. 12 is an explanatory diagram illustrating an example of processing of the text extraction unit 3424B, the word order change unit 3425B, and the output unit 3426B according to the third embodiment of the present invention.
The processing content example Ex18 is an example of a keyword given a weight by the weighting unit 3423B.
The text extraction unit 3424B calculates the element score of the post message of the extracted frame based on the keyword given the weight by the weighting unit 3423B. Specifically, the sentence extraction unit 3424B calculates the element score by multiplying the number of appearances of each keyword in the extracted frame counted by the weighting unit 3423B by the weight of each keyword. The text extraction unit 3424B calculates the total sum of the calculated element scores as a posted message score for each posted message.

In the processing content example Ex19, the text extraction unit 3424B sorts the posted messages in the extracted frame in descending order of the posted message score, and extracts, for example, the top five posted messages. Then, the text extraction unit 3424B extracts the posted message having the highest calculated posted message score from the top five posted messages as the first representative message.
For example, the sentence extraction unit 3424B, like the processing example Ex20, posted the message “around 5:33, a very strong earthquake with an epicenter of YYＹ in XX prefecture. No worries about tsunami "is extracted as the first representative message.

The word order changing unit 3425B generates keyword information by rearranging the word order of the keywords weighted by the weighting unit 3423B into the order of the words appearing in the first representative message. For example, the processing content example Ex21 is an example of keyword information generated by the word order changing unit 3425B and output from the output unit 3426B to the notification unit 332.
The word order changing unit 3425B changes the word order of each keyword to which a weight such as the processing content example Ex18 is assigned, using the keywords “earthquake”, “□□ Oshima”, “5 seismic intensity 5”, “ The keyword information is generated by rearranging based on the tsunami concern and the first representative message.

  Next, the weight calculation unit 3422B excludes the weight for the keyword included in the keyword information generated by the word order change unit 3425B. For example, as in the processing content example Ex22, the weighting unit 3423B performs “viewing 0.03”, “attaching 0.02”, “emergency earthquake early warning 0.01”, “tsunami is not concerned 0.01” for the keyword 0.01 The weight is given as follows.

The sentence extraction unit 3424B extracts, as the second representative message, the posted message having the highest posted message score other than the first representative message among the top five posted messages.
For example, the sentence extraction unit 3424B extracts the posted message “Earthquake early warning! This morning, I am not worried about the tsunami, saying at the A broadcast station” as the second representative message, as in the processing example Ex23. .

Next, the word order changing unit 3425B generates keyword information by rearranging the word order of the keywords weighted by the weighting unit 3423B into the order of words appearing in the second representative message. For example, the processing content example Ex24 is an example of keyword information generated by the word order changing unit 3425B and output from the output unit 3426B to the notification unit 332.
The word order changing unit 3425B sets the keywords “Earthquake Early Warning” and “Don't worry about tsunami” as in the case of the processing content example Ex24. The keyword information is generated by rearranging based on the representative message.

  In this way, the keywords extracted by the keyword extraction unit 3421 are rearranged in the order of words appearing in the first representative message and the second representative message extracted by the text extraction unit 3424B, and the first representative message is sorted. By repeating the calculation of the weight for the keyword so as not to select a keyword that overlaps with the second representative message, the rearranged keywords such as the processing content example Ex25 can be rearranged without duplication. .

  As described above, according to the present embodiment, the information processing apparatus 3B includes the keyword extraction unit 3421 that extracts a keyword from a sentence (post message) posted within a predetermined time, and a group of posted sentences (post message). A sentence extracting unit 3424B that extracts representative sentences (first representative message and second representative message) from (post message in the extraction frame) and a keyword extracted by the keyword extracting unit 3421 are extracted by the sentence extracting unit 3424B. And a word order changing unit 3425B for rearranging words in the order of words appearing in the representative sentences (first representative message, second representative message).

  Thereby, the information processing apparatus 3B can extract keywords for a plurality of posted sentences and rearrange the extracted keywords in the order of words appearing in the representative sentences. It is possible to extract changes in topics. Further, the information processing device 3B can notify the user of the reason for the excitement by extracting the transition of the topic, so that it is possible to improve convenience when notifying the user of the excitement. it can.

(Fourth embodiment)
Hereinafter, a fourth embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 13 is a schematic block diagram illustrating an example of the configuration of the analysis unit 342C according to the fourth embodiment of the present invention.
The analysis unit 342C includes a keyword extraction unit 3421, a weight calculation unit 3422, a weighting unit 3423, a sentence extraction unit 3424, a word order change unit 3425, an output unit 3426, and a filter unit 3427C. . When the analysis unit 342 according to the first embodiment and the analysis unit 342C according to the fourth embodiment are compared, a filter unit 3427C is added. Since the other configuration is the same as that of the first embodiment, the same reference numerals are given and description thereof is omitted.

  The filter unit 3427C performs a filtering process on the posted message information input from the reading unit 341. Specifically, the filter unit 3427C performs filtering processing on the posted message information input from the reading unit 341 based on identification information such as, for example, a broadcasting station, a channel, a program, and the like. Each post message information is extracted. The filter unit 3427C outputs the extracted posted message information to the keyword extraction unit 3421 and the text extraction unit 3424.

  As described above, according to the present embodiment, the information processing apparatus 3C includes the keyword extraction unit 3421 that extracts a keyword from a sentence (post message) posted within a predetermined time, and a group of posted sentences (post message). A text extraction unit 3424C that extracts representative text (representative message) from (post message in the extracted frame) and a keyword extracted by the keyword extraction unit 3421 appear in the representative text (representative message) extracted by the text extraction unit 3424C. And a word order changing unit 3425 for rearranging the words in order.

  Thereby, the information processing apparatus 3C can extract keywords from a plurality of posted sentences and rearrange the extracted keywords in the order of words appearing in the representative sentences. It is possible to extract changes in the topic of posted messages for each channel and program. In addition, the information processing apparatus 3C can notify the user of the reason for the excitement by extracting the transition of the topic for each broadcast station, channel, and program. Convenience can be improved.

  In each of the embodiments described above, an arbitrary posted message may be extracted by including the filter unit 3427C as in the fourth embodiment and performing filtering processing on the posted message information.

  In each of the above-described embodiments, the weight for each keyword is calculated from the ratio between the keyword appearance rate in the extracted frame and the keyword appearance rate in the reference frame. However, the keyword appearance rate in the extracted frame and the keyword appearance rate in the reference frame are calculated. The weight may be calculated from the difference from the appearance rate.

In each of the above-described embodiments, it has been described that the extracted keywords are rearranged in the order of words appearing in the representative message. However, the representative message may be notified as it is.
In each of the above-described embodiments, one or two representative messages are extracted and the keywords are rearranged. However, three or more representative messages may be extracted and the keywords may be rearranged.

In each embodiment described above, an example has been described in which the text extraction unit calculates the element score by multiplying the number of appearances of each keyword by the weight of each keyword. However, the element score is not limited to this, for example, The weight given to the keyword and the number of appearances of the keyword may be used as the element score as they are.
In each of the embodiments described above, the sentence extraction unit has shown an example of extracting a representative message based on the number of posted messages, but the representative message is extracted based on the number of postings of posted messages that are repeatedly posted within a predetermined time. The representative message may be extracted based on the number of characters in the posted message body, or the representative message may be extracted by combining a plurality of these extraction means.

  In addition, you may make it implement | achieve a part or all of information processing apparatus 3, 3A, 3B, 3C in each embodiment mentioned above with a computer. In that case, the program for realizing the control function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by a computer system and executed. Here, the “computer system” is a computer system built in the information processing apparatuses 3, 3 A, 3 B, and 3 C, and includes an OS and hardware such as peripheral devices.

  The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” is a medium that dynamically holds a program for a short time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line, In such a case, a volatile memory inside a computer system that serves as a server or a client may be included that holds a program for a certain period of time.

  The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

  Moreover, you may implement | achieve part or all of the information processing apparatus 3, 3A, 3B, 3C in embodiment mentioned above as integrated circuits, such as LSI (Large Scale Integration). Each functional block of the information processing apparatuses 3, 3A, 3B, and 3C may be individually made into a processor, or a part or all of them may be integrated into a processor. Further, the method of circuit integration is not limited to LSI, and may be realized by a dedicated circuit or a general-purpose processor. Further, in the case where an integrated circuit technology that replaces LSI appears due to progress in semiconductor technology, an integrated circuit based on the technology may be used.

  As described above, the embodiments of the present invention have been described in detail with reference to the drawings. However, the specific configuration is not limited to that described above, and various design changes and the like can be made without departing from the scope of the present invention. It is possible to

  (Additional remark 1) The keyword extraction part which extracts a keyword from the text posted within the predetermined time, the sentence extraction part which extracts the representative sentence from the group of the posted sentences, and the keyword extraction part extract An information processing apparatus comprising: a word order changing unit that rearranges the keywords in the order of words appearing in the representative sentence extracted by the sentence extracting unit.

  (Appendix 2) The information processing apparatus according to (Appendix 1), further comprising a weighting unit that assigns a weight to each of the keywords extracted by the keyword extraction unit.

  (Additional remark 3) The said text extraction part extracts the said representative sentence based on the said keyword which the said weighting part gave the weight, The information processing apparatus as described in (Additional remark 2) characterized by the above-mentioned.

  (Additional remark 4) The detection part which detects the number of the said posted sentences is further provided, and when the said detection part detects the number of the said sentence more than a predetermined number within the said predetermined time, the said keyword extraction part is the said The information processing apparatus according to (Appendix 3), wherein a keyword is extracted from the sentence posted within a predetermined time.

  (Supplementary Note 5) Keyword extraction process for extracting a keyword from sentences posted within a predetermined time, sentence extraction process for extracting representative sentences from a group of the posted sentences, and extraction by the keyword extraction process And a word order changing step of rearranging the keyword in the order of words appearing in the representative sentence extracted by the sentence extraction process.

  (Additional remark 6) The keyword extraction step which extracts a keyword from the text posted in the predetermined time in the computer of the information processing apparatus, and the text extraction step which extracts a representative text from the group of the posted text An information processing program for executing a word order changing step of rearranging the keywords extracted in the keyword extraction step in the order of words appearing in the representative sentence extracted in the sentence extraction step.

S1 ... Information processing system, 1 ... Posting terminal device, 2 ... Posting service server device, 3, 3A, 3B, 3C ... Information processing device, 4 ... Electronic equipment, N1 ... Network 31 ... acquisition unit 32 ... storage unit 33 ... detection unit 34 ... extraction unit 331 ... detection unit 332 ... notification unit 341 ... reading unit , 342, 342A, 342B, 342C ... analysis unit, 3421 ... keyword extraction unit, 3422, 3422B ... weight calculation unit, 3423, 3423B ... weighting unit, 3424, 3424A, 3424B ... sentence Extraction unit, 3425, 3425A, 3425B ... word order change unit, 3426, 3426A, 3426B ... output unit, 3427C ... filter unit

Claims (6)

A keyword extraction unit that extracts keywords from sentences posted within a predetermined time;
A sentence extraction unit that extracts representative sentences from the group of the posted sentences;
A word order changing unit for rearranging the keywords extracted by the keyword extracting unit in the order of words appearing in the representative sentence extracted by the sentence extracting unit;
An information processing apparatus comprising: The information processing apparatus according to claim 1, further comprising a weighting unit that assigns a weight to each of the keywords extracted by the keyword extraction unit. The information processing apparatus according to claim 2, wherein the sentence extraction unit extracts the representative sentence based on the keyword given a weight by the weighting unit. A detection unit for detecting the number of posted sentences;
The keyword extraction unit extracts a keyword from the sentence posted within the predetermined time when the detection unit detects the number of the sentences equal to or greater than a predetermined number within the predetermined time. The information processing apparatus according to any one of claims 1 to 3. A keyword extraction process to extract keywords from sentences posted within a predetermined time;
A sentence extraction process for extracting representative sentences from the group of the posted sentences;
A word order changing process for rearranging the keywords extracted in the keyword extraction process in the order of words appearing in the representative sentence extracted in the sentence extraction process;
An information processing method characterized by comprising: In the computer of the information processing device,
A keyword extraction step for extracting keywords from sentences posted within a predetermined time;
A sentence extraction step of extracting representative sentences from the group of the posted sentences;
A word order changing step of rearranging the keywords extracted in the keyword extraction step in the order of words appearing in the representative sentence extracted in the sentence extraction step;
Information processing program to execute.

Images