JP2010122739A - Content distribution server, content distribution method, and communication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To select and distribute appropriate contents that are easy to be received by a user by precisely setting the priorities of contents based on a tendency for information access of the user (the randomness of the information access). <P>SOLUTION: The content distribution server calculates the current random level based on the Web access pattern of the current session of a user with the matching result of a unique access pattern (ST16), and selects distribution candidates based on the current information acceptance scores obtained from the matching result and a preliminarily set threshold (ST17, ST18), and calculates the optimal random level of each distribution candidate from the random level in the distribution of each piece of distribution information, and determines the distribution priority order of the distribution candidates based on the optimal random level of each distribution candidate and the current random level (ST20), and distributes the distribution information as the distribution candidates in a sequence based on the distribution priority order (ST21). <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a content distribution server that distributes content to a mobile terminal, a content distribution method that is executed by the content distribution server, and a communication system that includes the content distribution server and the mobile terminal.

  The user can access various contents from the portable terminal using his spare time such as free time. Since the content that can be accessed from the portable terminal is very large, a system that selects and distributes content that is useful to the user from the huge amount of content is desired. In such a system, it is conceivable to prioritize contents in order to select contents to be distributed.

On the other hand, the content accessed by the user often differs depending on the user's access purpose at that time. The access purposes here can be roughly divided into the following two, for example.
(1) For an access purpose of checking predetermined information, the user periodically checks, for example, a weather forecast site, a train transfer guidance site, a coupon issue site, and the like.
(2) For the purpose of searching for useful information other than the information that is always checked, the user searches for interesting blogs and articles other than the blogs that are always browsed.
JP 2002-108923 A JP 2002-334104 A

  However, prior art prioritization of content rarely takes into account the user's access purpose, so it is difficult to improve the accuracy of prioritization, and as a result, it is difficult to select and distribute content useful for the user. It was.

  SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and has an object to accurately prioritize content based on a user's information access tendency (information access randomness) and to select and distribute appropriate content that is easily accepted by the user. And

  In order to achieve the above object, the content distribution server according to the present invention provides a unique access pattern database storing a unique access pattern for each user, and an information acceptance score when the unique access pattern appears for each unique access pattern. Stored information acceptance score database, distribution information database storing distribution information to be distributed, and distribution information acceptance history storing pattern conformance rate history and information acceptance rate history for each unique access pattern at the time of distribution For each piece of distribution information accepted by the user, a distribution randomness history database storing randomness at the time of distribution, a frequent access pattern extraction unit for extracting a plurality of frequent access patterns from the user's web access pattern, Extracted frequent access patterns A unique access pattern clustering unit for clustering frequent access patterns having a mutual similarity of a predetermined level or more, and storing the obtained access patterns as unique access patterns in the unique access pattern database; An access pattern matching unit that matches the session web access pattern with the unique access pattern stored in the unique access pattern database, and based on the matching result, the current session web access pattern for the unique access pattern A pattern matching rate calculation unit for calculating a pattern matching rate representing a matching level, the calculated pattern matching rate, and information acceptance for each unique access pattern stored in the information acceptance score database An information acceptance score calculation unit that calculates a current information acceptance score based on the core and stores the current information acceptance score in the information acceptance score database; and a web of the current session based on the matching result Based on a randomness calculation unit that calculates a current randomness representing a degree of randomness of an access pattern and the current information acceptance score, a distribution candidate is selected from the distribution information stored in the distribution information database. An information selection unit, and an optimal randomness calculation unit that calculates an optimal randomness for each of the selected distribution candidates from a randomness at the time of distribution for each distribution information with reference to the distribution randomness history database. Based on the optimal randomness for each distribution candidate and the current randomness calculated by the randomness calculator A distribution priority determining unit that determines a distribution priority for the distribution candidate, an information distribution unit that distributes the distribution information determined as the distribution candidate in an order based on the determined distribution priority, and a distributed distribution A feedback receiving unit that receives feedback on a user's information acceptance time for information from a distribution-destination mobile terminal, and stores distribution randomness in the distribution randomness history database for each distribution information accepted by the user An information acceptance rate is calculated based on the distribution randomness history storage unit, the received information acceptance time, the delivery time of the delivery information, and a predetermined reference time interval, and the information acceptance at the time of the current information delivery calculated The rate is stored in the distribution information acceptance history database together with the calculated pattern matching rate, and the distribution information acceptance rate is recorded. Pattern-specific information acceptance for each unique access pattern based on the pattern conformance rate and information acceptance rate at the time of information delivery in the past stored in the database, and the pattern conformance rate and information acceptance rate at the time of the current information delivery An information acceptance score updating unit that calculates a score and updates the information acceptance score stored in the information acceptance score database with the calculated information acceptance score for each pattern.

  The above content distribution server realizes accurate prioritization of distribution candidates based on the user's information access tendency (randomness of information access), thereby selecting appropriate content that is easily accepted by the user. Can be delivered.

  In the content distribution server, the randomness calculation unit may determine that the length of the character string of the portion that does not match the unique access pattern in the web access pattern of the current session is the character of the web access pattern of the current session. A ratio that occupies the total length of the column is calculated as a randomness for each unique access pattern, and an average value of the calculated randomness for each unique access pattern is calculated as a current randomness; It is desirable to do.

  In addition, the optimal random degree calculation unit obtains a random degree at the time of distribution of another user who has accepted the distribution candidate for each of the selected distribution candidates from the distribution random degree history database. It is desirable that the average value of the randomness is calculated as the optimal randomness for the distribution candidate.

  Further, the distribution priority order determination unit is configured to determine the distribution priority order for the distribution candidates in the order in which the calculated optimal randomness for each distribution candidate is closer to the calculated current randomness. Is desirable.

  By the way, the content distribution server according to the present invention described above can be regarded as an invention according to the following content distribution method and an invention according to a communication system, and has the same operations and effects.

  A content distribution method according to the present invention includes a unique access pattern database storing a unique access pattern for each user, an information acceptance score database storing an information acceptance score at the time of appearance of the unique access pattern for each unique access pattern, and distribution Distribution information database storing distribution information to be distributed, distribution information acceptance history database storing pattern conformance history and information acceptance history for each unique access pattern at the time of distribution, and distribution accepted by the user A content distribution method that is executed by a content distribution server including a distribution randomness history database that stores distribution randomness for each piece of information, and a plurality of frequent access patterns from a user's Web access history The step of extracting and extracting Clustering frequent access patterns whose similarity is equal to or higher than a predetermined level among the frequent access patterns, storing the obtained access patterns in the unique access pattern database as unique access patterns, and the current session of the user Matching the web access pattern of the current session with the unique access pattern stored in the unique access pattern database, and representing the matching level of the web access pattern of the current session with the unique access pattern based on the matching result A current information acceptance score based on the step of calculating a pattern adaptation rate, the calculated pattern adaptation rate, and the information acceptance score for each unique access pattern stored in the information acceptance score database Calculating and storing the current information acceptance score in the information acceptance score database, and calculating a current randomness representing a degree of randomness of the web access pattern of the current session based on the matching result A step of selecting a delivery candidate from the delivery information stored in the delivery information database based on the current information acceptance score; a delivery for each delivery information with reference to the delivery randomness history database; A step of calculating an optimum random degree for each of the selected delivery candidates, a calculated optimum random degree for each delivery candidate, and a current random degree calculated by the random degree calculating unit. And determining a distribution priority order for the distribution candidates, and the determined distribution priority. Distributing the distribution information set as the distribution candidates in an order based on the priorities, receiving a feedback of the user's information acceptance time for the distributed distribution information from the distribution-destination portable terminal, and accepting by the user For each piece of distribution information, information on the basis of the step of storing the distribution randomness in the distribution randomness history database, the received information acceptance time, the distribution information distribution time, and a predetermined reference time interval The acceptance rate is calculated, and the calculated information acceptance rate at the time of information delivery is stored in the delivery information acceptance history database together with the calculated pattern matching rate, and the past stored in the delivery information acceptance history database Pattern conformance rate and information acceptance rate at the time of information delivery of the predetermined times, and the pattern at the time of this information delivery Calculating a pattern-specific information acceptance score for each unique access pattern based on the combination rate and the information acceptance rate, and updating the information acceptance score stored in the information acceptance score database with the calculated pattern-specific information acceptance score And.

  A communication system according to the present invention is a communication system including a mobile terminal and a content distribution server that distributes information to the mobile terminal, and the mobile terminal is a current session of the mobile terminal. An access history transmitting unit that transmits a Web access pattern to the content distribution server, a distribution information receiving unit that receives distribution information from the content distribution server, a distribution information display unit that displays received distribution information, and the distribution information A feedback transmission unit that transmits the information acceptance time to the content distribution server, and the content distribution server includes an access history database that stores a Web access pattern for each user and a unique access pattern that stores a unique access pattern for each user. Access pattern database and appearance of unique access patterns An information acceptance score database storing information acceptance scores for each unique access pattern, a distribution information database storing distribution information to be distributed, and a history and information of pattern conformance rates for each unique access pattern at the time of distribution Distribution information acceptance history database storing acceptance rate history, distribution randomness history database storing distribution randomness for each distribution information accepted by the user, and target user stored in the access history database Access pattern obtained by clustering frequent access patterns whose similarity is equal to or higher than a predetermined level among the frequent access patterns extracted from the frequent access pattern extraction unit that extracts multiple frequent access patterns from the web access patterns The unique access pattern A unique access pattern clustering unit for storing in the unique access pattern database as an instance, a web access pattern of a current session in the portable terminal received from the portable terminal, and an access history receiving unit for storing in the access history database; An access pattern matching unit that matches the web access pattern of the current session with the unique access pattern stored in the unique access pattern database, and the web access of the current session to the unique access pattern based on the matching result A pattern matching rate calculating unit that calculates a pattern matching rate that represents the degree of pattern matching, the calculated pattern matching rate, and each unique access pattern stored in the information acceptance score database. An information acceptance score calculation unit that calculates a current information acceptance score based on the information acceptance score, stores the current information acceptance score in the information acceptance score database, and the current session based on the matching result A randomness calculation unit for calculating a current randomness representing a degree of randomness of the web access pattern, and a distribution candidate from distribution information stored in the distribution information database based on the current information acceptance score An information selection unit to select, an optimal randomness calculation unit that refers to the distribution randomness history database, calculates an optimal randomness for each of the selected distribution candidates, from a randomness at the time of distribution for each distribution information; The optimal randomness for each distribution candidate calculated, and the current randomness calculated by the randomness calculator A distribution priority determining unit that determines a distribution priority for the distribution candidates, an information distribution unit that distributes the distribution information set as the distribution candidates in an order based on the determined distribution priority, and a distribution completed A feedback receiving unit that receives feedback of the user's information acceptance time for the distribution information from the distribution-destination mobile terminal, and the distribution randomness history database for each distribution information accepted by the user A distribution-time randomness history storage unit to store, an information acceptance rate is calculated based on the received information acceptance time, the delivery time of the delivery information, and a predetermined reference time interval. An information acceptance rate is stored in the delivery information acceptance history database together with the calculated pattern matching rate, and the delivery A pattern for each unique access pattern based on the pattern conformance rate and information acceptance rate at the time of information delivery in the past stored in the report acceptance history database, and the pattern conformance rate and information acceptance rate at the time of the current information delivery An information acceptance score updating unit that calculates another information acceptance score, and updates the information acceptance score stored in the information acceptance score database with the calculated information acceptance score for each pattern.

  According to the present invention, in the content distribution server, distribution information (content) can be prioritized with high accuracy based on the user's information access tendency (information access randomness), thereby being accepted by the user. It is possible to select and distribute appropriate content that is easily handled.

  Hereinafter, embodiments according to the present invention will be described with reference to the drawings.

(Configuration of communication system)
First, the configuration of the communication system according to the present embodiment will be described. As shown in FIG. 1, the communication system 1 according to this embodiment includes a mobile terminal 30 and a content distribution server 10, and the mobile terminal 30 is connected via a network N (for example, a mobile communication network or a wireless LAN communication network). Data can be transmitted to and received from the content distribution server 10.

  The mobile terminal 30 is a terminal having a wireless data communication function, and corresponds to, for example, a mobile phone, a PHS, or a PDA equipped with a wireless communication card.

  The content distribution server 10 is a server that controls the timing set for each content and performs content distribution. Although details will be described later, the content distribution server 10 extracts a user-specific access pattern from the user's Web access history. In addition, the compatibility rate between the current access log and the unique access pattern is calculated, the current information acceptance rate of the user is calculated, and the distribution candidate is selected. Then, the selected distribution candidates are prioritized based on the current information access randomness, and are distributed from distribution candidates with higher priorities. Furthermore, the feedback about the timing when the user accepted the information is received from the portable terminal, and the information acceptance rate for each unique access pattern is updated.

  The content distribution server 10 includes an access history DB 201, a unique access pattern DB 202, an access pattern-specific information acceptance score DB 203, a distribution information DB 204, a distribution information acceptance history DB 205, and a content-specific distribution randomness history DB 206 shown in FIG. A database group 20 is provided.

(Configuration of mobile device)
As illustrated in FIG. 2, the mobile terminal 30 includes an access history transmission unit 301 that transmits a web access history accessed from the terminal to the content distribution server 10, and a distribution information reception unit 302 that receives distribution information from the content distribution server 10. A distribution information display unit 303 that displays the received distribution information, and a feedback transmission unit 304 that transmits the time when the user viewed the distributed information to the content distribution server 10.

(Content distribution server configuration)
As shown in FIG. 3, the content distribution server 10 includes an offline control unit 100 </ b> A that controls execution of offline processing described later, an online control unit 100 </ b> B that controls execution of online processing described later, and a database group 20.

  The offline control unit 100A includes a frequent access pattern extraction unit 101 and a unique access pattern clustering unit 102. Among them, the frequent access pattern extraction unit 101 extracts the access sequence for each session from the access history DB 201 storing the URL accessed by the user (access destination URL), and compares the access sequences of each session, A plurality of frequent Web access patterns (hereinafter referred to as “frequent access patterns”) are extracted for each user. The unique access pattern clustering unit 102 calculates the similarity between the extracted frequent access patterns, clusters the access patterns whose mutual similarity is a predetermined level or more, and uses the obtained access pattern as a unique access pattern. Store in the DB 202.

  The online control unit 100 </ b> B includes the following function groups 103 to 114 for receiving a web access history from the mobile terminal 30, calculating an information acceptance score, and learning a distribution information acceptance rate.

  The access history receiving unit 103 receives the user's Web access history from the access history transmitting unit 301 of the mobile terminal 30 and stores it in the access history DB 201.

  The access pattern matching unit 104 compares the web access history (web access pattern of the user's current session) received at that time by the access history receiving unit 103 and the user's unique access pattern stored in the unique access pattern DB 202. Perform matching.

  From the matching result of the access pattern matching unit 104, the pattern matching rate calculation unit 105 determines that the character string length of the common part between the Web access pattern and the unique access pattern of the user's current session is the character string of the unique access pattern. The proportion of the total length is calculated for each unique access pattern as a pattern matching rate.

  The information acceptance score calculation unit 106 uses the adaptation rate for each unique access pattern calculated by the pattern adaptation rate calculation unit 105 and the information acceptance rate at the time of appearance of each unique access pattern stored in the information acceptance score DB 203 for each access pattern. Based on the method described later, the current information acceptance score is calculated.

  Based on the matching result of the access pattern matching unit 104, the randomness calculation unit 107 determines that the length of the character string of the part that does not match the unique access pattern in the Web access pattern of the user's current session is the Web access pattern of the current session Is calculated for each unique access pattern as a degree of randomness, and the added average value of the calculated degrees of randomness is used as the degree of randomness of information access for the current user.

  The recommendation information selection unit 108 compares the current information reception score calculated by the information reception score calculation unit 106 with the threshold value for each piece of distribution information stored in the distribution information DB 204, and the current information reception score is equal to or greater than the threshold value. Select distribution information as a distribution candidate.

  The optimal randomness calculation unit 109 refers to the content-specific distribution randomness history DB 206 and calculates the optimal randomness for each distribution candidate from the randomness of the distribution information viewed by other users.

  The distribution priority order determination unit 110 distributes the distribution candidates selected by the recommendation information selection unit 108 based on the optimal randomness calculated by the optimal randomness calculation unit 109 and the randomness calculated by the randomness calculation unit 107. Determine priority.

  The recommendation information distribution unit 111 distributes the distribution candidates to the distribution information reception unit 302 of the mobile terminal 30 in the order based on the distribution priority determined by the distribution priority determination unit 110.

  The feedback receiving unit 112 provides feedback of the user's information acceptance time for the distributed distribution information (specifically, the identification information of the distribution information browsed by the user and the browsing time information) as a feedback transmission unit of the mobile terminal 30 Received from 304.

  The distribution-by-content distribution randomness history storage unit 113 stores the distribution-by-content distribution randomness history DB 206 for each distribution information viewed by the user.

  The information acceptance score update unit 114 for each access pattern calculates the current information acceptance rate based on the feedback of the information acceptance time received by the feedback receiver 112, the delivery time of the delivery information, and a predetermined reference time interval. The information acceptance rate at the time of this information delivery is stored in the delivery information acceptance history DB 205 together with the calculated pattern matching rate for each unique access pattern. Further, the information acceptance score update unit 114 by access pattern stores the pattern conformance rate and information acceptance rate in the past predetermined information distribution stored in the distribution information acceptance history DB 205, and the pattern conformance rate in the current information distribution. Based on the information acceptance rate, an information acceptance score for each pattern is calculated for each unique access pattern, and the information acceptance score stored in the information acceptance score for each access pattern DB 203 is updated with the calculated information acceptance score for each pattern.

  Hereinafter, an outline of each DB included in the database group 20 will be described.

  The access history DB 201 illustrated in FIG. 4 is a DB that stores a URL accessed by the user (access destination URL). As shown in FIG. 4, the same session ID (S001) is assigned to URLs a, b, k, j accessed in the same session.

  The unique access pattern DB 202 illustrated in FIG. 5 is a DB that stores the unique access pattern for each user extracted by the unique access pattern clustering unit 102.

  The information acceptance score DB 203 for each access pattern shown in FIG. 6 is a DB that stores an information acceptance score for each access pattern (information acceptance score for each pattern).

  The distribution information DB 204 illustrated in FIG. 7 is a DB that stores IDs of distribution information (contents) distributed to users and information acceptance score thresholds that are set in advance for each piece of distribution information.

  The distribution information acceptance history DB 205 shown in FIG. 8 is a history of pattern conformance rates with respect to each unique access pattern when information is distributed to the user, and an information acceptance rate history calculated based on an elapsed time until information browsing. Is a DB that stores.

  The distribution-by-content distribution randomness history DB 206 shown in FIG. 9 is a DB that stores the distribution randomness for each distribution information (content) viewed by the user.

  By the way, from the viewpoint of hardware, for example, the content distribution server 10 includes a CPU 10A that executes an operating system, an application program, and the like, a main storage unit 10B that includes a ROM and a RAM, and a nonvolatile memory as illustrated in FIG. An auxiliary storage unit 10C composed of a volatile memory, a communication control unit 10D that performs data communication, an output unit 10E that displays information, prints out information, and the like, and performs character / number input and execution instructions And an operation unit 10F composed of keys. Note that each function described in FIG. 3 reads predetermined software on the CPU 10A and the main storage unit 10B shown in FIG. 20, executes a predetermined application, and controls the communication control unit 10D under the control of the CPU 10A. It is realized by operating and reading and writing data in the main storage unit 10B and the auxiliary storage unit 10C. Further, as a hardware configuration, the above-described portable terminal 30 basically has the same configuration.

(Processing flow related to timing control)
The processing flow according to the present embodiment includes (a) offline processing for extracting a unique access pattern for each user, and (b) a current information acceptance score based on a matching result between the current access log and the unique access pattern, and Randomness is calculated, distribution candidates obtained based on the current information acceptance score are distributed according to the distribution priority obtained based on the randomness, and the information acceptance score for each pattern is updated based on feedback of information acceptance time Including online processing. Hereinafter, it demonstrates in order.

((A) Offline processing flow)
FIG. 10 shows a processing flow of offline processing. The frequent access pattern extraction unit 101 refers to the access history DB 201 and extracts an access sequence for each session as shown in FIG. 11 (step ST01). Further, the frequent access pattern extraction unit 101 compares the access sequences of the sessions and extracts the frequent access patterns (step ST02). For example, as shown in FIG. 12, “bkl”, “bkk”, and “ghi” are extracted as frequent access patterns of length “3 characters”.

  Next, the unique access pattern clustering unit 102 compares the extracted frequent access patterns with each other, and calculates the length of the character string of the common part as the similarity (step ST03). In the example of FIG. 12, for the frequent access patterns “b k l” and “b k k”, the length “2” of the character string “b k” of the common part is calculated as the similarity. Then, the unique access pattern clustering unit 102 clusters frequently accessed patterns having a certain degree of similarity or more and stores them in the unique access pattern DB 202 as unique access patterns (step ST04). As the clustering method here, there are a method of ANDing frequent access patterns and a method of ORing. In FIG. 12, OR is performed for frequent access patterns “bkl” and “bkk”. An example in which the obtained “bkkl” is extracted as a unique access pattern is shown.

((B) Online processing flow)
FIG. 13 shows a processing flow of online processing. When the access history receiving unit 103 receives the web access history from the portable terminal 30, an affirmative determination is made in step ST11, the process proceeds to step ST12, and the access history receiving unit 103 stores the received web access history in the access history DB 201. Next, the access pattern matching unit 104 matches the unique access patterns P001, P002,... Of the corresponding user stored in the unique access pattern DB 202 with the Web access pattern of the current session (step ST13). Next, the pattern matching rate calculation unit 105 calculates the matching rate for each unique access pattern from the matching result (step ST14). Here, as shown in FIG. 14, the length of the character string (character string consisting of the characters circled in FIG. 14) of the common part of the Web access pattern and the unique access pattern of the current session is the unique access pattern. Is calculated for each unique access pattern, and the calculated ratio is used as the matching rate for each unique access pattern.

  Next, the information acceptance score calculation unit 106 calculates the current information acceptance score of the user (step ST15). Here, for example, as shown in FIG. 15, the pattern matching rate calculated by the pattern matching rate calculating unit 105 and the information acceptance score for each unique access pattern stored in the access acceptance information DB for each access pattern DB 203 (information acceptance for each pattern) Score)), the addition average of the multiplication results is calculated, and the obtained addition average value “0.4” is set as the current information acceptance score.

  Further, the randomness calculation unit 107 calculates the randomness of the user's information access at the present time (step ST16). Here, for example, as shown in FIG. 16, the length of the character string of the part that does not match the unique access pattern in the Web access pattern of the user's current session is the length of the entire character string of the Web access pattern of the current session. The ratio of the calculated randomness is calculated for each unique access pattern as the randomness, and the added average value of the calculated randomness is used as the randomness of information access for the current user.

  After calculating the current information acceptance score and randomness of the target user as described above, the recommendation information selection unit 108 is determined for each of the calculated current information acceptance score and the distribution information stored in the distribution information DB 204. And the distribution information (content) whose current information acceptance score is equal to or greater than the threshold among the stored distribution information is selected as a distribution candidate (step ST18). That is, distribution information (content) whose current information acceptance score is lower than the threshold is not distributed (that is, not selected as a distribution candidate: step ST19). For example, when the current information acceptance score is 0.4, distribution information C002 shown in FIG. 17 is not selected as a distribution candidate, and distribution information C001 and C003 are selected as distribution candidates.

  In step ST20, the optimal randomness calculation unit 109 refers to the distribution-specific distribution randomness history DB 206 for each selected delivery candidate, and calculates an average value of randomness of other users who have browsed the delivery candidate. Calculated as the optimal randomness for the distribution candidate. For example, as shown in FIG. 17, 0.7, which is an average value of randomness (0.5, 0.9, 0.7) at the time of distribution of users U001, U002, and U003 who browsed the distribution candidate C001, is optimal for the distribution candidate C001. Calculated as randomness. Then, the distribution priority order determination unit 110 determines the distribution priority order of the distribution candidates in the order in which the optimal randomness is close to the current user information access randomness calculated in step ST16. For example, as shown in FIG. 17, when the current user's information access randomness is “0.5”, among the distribution candidates C001 and C003, the distribution candidate C001 with the optimal randomness “0.7” has the first priority. The priority of the distribution candidate C003 having the optimal randomness “0.9” is set to the second place.

  And the recommendation information delivery part 111 delivers delivery information (content) to the portable terminal 30 in the order based on the determined delivery priority (step ST21). In the portable terminal 30, the distribution information receiving unit 302 receives the distribution information (content), and the distribution information display unit 303 displays the received distribution information (content). At this time, for example, the distribution information (content) may be distributed to the portable terminal 30 by normal data distribution, or may be distributed to the portable terminal 30 by e-mail as shown in the lower right part of FIG. Moreover, as a display mode of the distribution information (content) by the distribution information display unit 303, it may be displayed on the e-mail reception screen as shown in the lower right part of FIG. 18, or a telop as shown in the lower left part of FIG. You may display.

  After the distribution information (content) is distributed, the feedback receiving unit 112 receives feedback of the user's information acceptance time for the distributed distribution information (specifically, the distribution viewed by the user) from the feedback transmission unit 304 of the mobile terminal 30. Information identification information and browsing time information). However, if feedback cannot be received from the feedback transmission unit 304 even after a certain time has elapsed, it is determined that the distributed distribution information (content) has not been browsed by the user.

  After receiving the feedback, the content-specific distribution randomness history storage unit 113 stores the distribution randomness in the content-specific distribution randomness history DB 206 for each distribution information viewed by the user.

  Then, the information acceptance score update unit 114 by access pattern is based on the ratio of the elapsed time from the delivery time of the delivered delivery information (content) to the viewing time with respect to a predetermined fixed reference time interval. As described later, the information acceptance rate is calculated, and the calculated information acceptance rate at the time of information delivery is stored in the delivery information acceptance history DB 205 together with the calculated pattern matching rate.

  Further, the information acceptance score update unit 114 by access pattern stores the pattern conformance rate and information acceptance rate in the past predetermined information distribution stored in the distribution information acceptance history DB 205, and the pattern conformance rate in the current information distribution. Based on the information acceptance rate, an information acceptance score for each access pattern is calculated for each unique access pattern based on the information acceptance rate, and the information acceptance stored in the access acceptance information DB for each access pattern DB 203 with the calculated information acceptance score for each pattern. The score is updated (step ST22).

As shown in FIG. 19, the information distribution time is T1, the information browsing time (time when the user clicks on the information) is T2, and the elapsed time t ₁₂ (= T2−T1) from information distribution to information browsing is 2. When the fixed time interval t is 10 minutes, the information acceptance rate is calculated by the following formula (1), for example.
Information acceptance rate = 1− (t ₁₂ /t)=1−(2/10)=0.8 (1)

As for the information acceptance score for each pattern, for example, the sum of the products of the pattern conformance rate and the information acceptance rate at the time of each information distribution is expressed as the following equation (2). The result obtained by dividing by the sum is used as an information acceptance score for each pattern.
Information acceptance score by pattern = (0.6 x 0.7 + 0.2 x 0.1 + 0.8 x 0.8) / (0.6 + 0.2 + 0.8) (2)

  According to the present embodiment described above, in the content distribution server, distribution information (content) can be prioritized with high accuracy based on the user's information access tendency (randomness of information access). It is possible to select and distribute appropriate content that is easily accepted by the user.

1 is a configuration diagram of a communication system according to an embodiment of the invention. It is a functional block block diagram of a portable terminal. It is a functional block block diagram of a content delivery server. It is a figure which shows the example of data memorize | stored in access history DB. It is a figure which shows the example of data memorize | stored in specific access pattern DB. It is a figure which shows the example of data memorize | stored in information acceptance score DB classified by access pattern. It is a figure which shows the example of data memorize | stored in delivery information DB. It is a figure which shows the example of data memorize | stored in distribution information acceptance log | history DB. It is a figure which shows the example of data memorize | stored in random degree log | history DB at the time of delivery according to contents. It is a processing flow figure of offline processing. It is a figure which shows the example of an access sequence for every session. It is a figure which shows the frequent access pattern extracted from the access sequence of FIG. 11, and an example of a clustering result. It is a processing flow figure of online processing. It is a figure which shows the example of calculation of a pattern precision. It is a figure which shows the example of calculation of the present information acceptance score. It is a figure which shows the example of calculation of randomness. It is a figure which shows the example of selection of delivery information. It is a figure which shows the delivery information management screen in a content delivery server, and the information delivery image to a portable terminal. It is a figure which shows the example of calculation of the information acceptance score according to access pattern. It is a hardware block diagram of a content delivery server.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Communication system, 10 ... Content delivery server, 10A ... CPU, 10B ... Main memory part, 10C ... Auxiliary memory part, 10D ... Communication control part, 10E ... Output part, 10F ... Operation part, 20 ... Database group, 30 ... Mobile terminal, 100A ... offline control unit, 100B ... online control unit, 101 ... frequent access pattern extraction unit, 102 ... unique access pattern clustering unit, 103 ... access history reception unit, 104 ... access pattern matching unit, 105 ... pattern matching rate Calculation unit 106 ... Information acceptance score calculation unit 107 ... Randomness calculation unit 108 ... Recommendation information selection unit 109 ... Optimal randomness calculation unit 110 110 Distribution priority order determination unit 111 111 Recommendation information distribution unit 112 Feedback receiving unit, 113 ... Randomness during distribution by content Storage unit 114 ... Access pattern-specific information acceptance score update unit 201 ... Access history DB, 202 ... Unique access pattern DB, 203 ... Access pattern-specific information acceptance score DB, 204 ... Distribution information DB, 205 ... Distribution information reception history DB , 206 ... Randomness history DB at the time of distribution, 301 ... Access history transmission unit, 302 ... Distribution information reception unit, 303 ... Distribution information display unit, 304 ... Feedback transmission unit.

Claims (6)

A unique access pattern database storing unique access patterns for each user;
An information acceptance score database storing an information acceptance score at the time of appearance of the unique access pattern for each unique access pattern;
A distribution information database storing distribution information to be distributed;
Distribution information acceptance history database storing pattern conformance rate history and information acceptance rate history for each unique access pattern at the time of delivery,
For each piece of distribution information accepted by the user, a distribution randomness history database that stores the distribution randomness,
A frequent access pattern extraction unit that extracts a plurality of frequent access patterns from a user's Web access pattern;
Among the extracted frequent access patterns, a unique access pattern clustering unit that clusters frequent access patterns whose mutual similarity is equal to or higher than a predetermined level, and stores the obtained access pattern in the unique access pattern database as a unique access pattern;
An access pattern matching unit that matches a web access pattern of the current session of the user with a unique access pattern stored in the unique access pattern database;
Based on the matching result, a pattern matching rate calculation unit that calculates a pattern matching rate that represents the matching degree of the Web access pattern of the current session with respect to the unique access pattern;
Based on the calculated pattern matching rate and the information acceptance score for each unique access pattern stored in the information acceptance score database, a current information acceptance score is calculated, and the current information acceptance score is calculated as the information acceptance score. An information acceptance score calculation unit to be stored in the score database;
Based on the matching result, a randomness calculation unit that calculates a current randomness representing a degree of randomness of the web access pattern of the current session;
An information selection unit that selects a delivery candidate from the delivery information stored in the delivery information database based on the current information acceptance score;
An optimal randomness calculation unit that refers to the distribution randomness history database and calculates an optimal randomness for each of the selected distribution candidates from a randomness during distribution for each distribution information;
A distribution priority determining unit that determines a distribution priority for the distribution candidates based on the calculated optimal randomness for each distribution candidate and the current randomness calculated by the randomness calculating unit;
An information distribution unit that distributes the distribution information set as the distribution candidates in an order based on the determined distribution priority;
A feedback receiving unit that receives feedback of a user's information acceptance time for distributed distribution information from a distribution-destination mobile terminal;
For each piece of distribution information accepted by the user, a distribution randomness history storage unit that stores a distribution randomness in the distribution randomness history database;
An information acceptance rate is calculated based on the received information acceptance time, the delivery time of the delivery information, and a predetermined reference time interval, and the calculated information acceptance rate at the time of information delivery is calculated based on the calculated pattern adaptation. The distribution information acceptance history database together with the rate, and the pattern conformance rate and information acceptance rate at the time of the past predetermined information distribution stored in the distribution information acceptance history database, and the pattern conformance at the time of the current information distribution Information acceptance score for each unique access pattern is calculated based on the rate and the information acceptance rate, and the information acceptance score stored in the information acceptance score database is updated with the calculated information acceptance score for each pattern. A score update unit;
A content distribution server comprising: The randomness calculation unit
The proportion of the length of the character string that does not match the unique access pattern in the web access pattern of the current session accounts for the total length of the character string of the web access pattern of the current session. Calculate as the randomness for each pattern, and calculate the average of the randomness for each unique access pattern calculated as the current randomness,
The content distribution server according to claim 1. The optimal randomness calculation unit includes:
From the distribution randomness history database, for each of the selected distribution candidates, obtain the randomness at the time of distribution of other users who have accepted the distribution candidates, and calculate the addition average value of the randomness at the distribution, Calculate as the optimal randomness for the delivery candidate,
The content distribution server according to claim 1, wherein the content distribution server is a content distribution server. The delivery priority determining unit
Determining the distribution priority for the distribution candidates in the order in which the calculated optimal randomness for each distribution candidate is close to the calculated current randomness;
The content distribution server according to claim 1, wherein the content distribution server is a content distribution server. A unique access pattern database storing a unique access pattern for each user, an information acceptance score database storing an information acceptance score when the unique access pattern appears for each unique access pattern, and distribution information storing distribution information to be distributed Randomness at the time of distribution for the database, distribution information acceptance history database storing the history of pattern conformance rate and information acceptance rate for each unique access pattern at the time of distribution, and distribution information accepted by the user A content distribution method executed by a content distribution server including a distribution randomness history database storing
Extracting multiple frequent access patterns from the user's web access history;
Clustering frequent access patterns having similarities of a predetermined level or more among the extracted frequent access patterns, and storing the obtained access patterns in the unique access pattern database as unique access patterns;
Matching a web access pattern of the user's current session with a unique access pattern stored in the unique access pattern database;
Based on the matching result, calculating a pattern matching rate representing a matching degree of the web access pattern of the current session with respect to the unique access pattern;
Based on the calculated pattern matching rate and the information acceptance score for each unique access pattern stored in the information acceptance score database, a current information acceptance score is calculated, and the current information acceptance score is calculated as the information acceptance score. Storing in the score database;
Calculating a current degree of randomness representing a degree of randomness of the web access pattern of the current session based on the matching result;
Selecting a delivery candidate from the delivery information stored in the delivery information database based on the current information acceptance score;
Referring to the distribution randomness history database and calculating the optimal randomness for each of the selected distribution candidates from the randomness during distribution for each distribution information;
Determining a distribution priority for the distribution candidates based on the calculated optimal randomness for each distribution candidate and the current randomness calculated by the randomness calculation unit;
Delivering the delivery information as the delivery candidates in an order based on the decided delivery priority;
Receiving feedback of the user's information acceptance time about the distributed distribution information from the mobile terminal of the distribution destination;
Storing distribution randomness in the distribution randomness history database for each distribution information accepted by the user;
An information acceptance rate is calculated based on the received information acceptance time, the delivery time of the delivery information, and a predetermined reference time interval, and the calculated information acceptance rate at the time of information delivery is calculated based on the calculated pattern adaptation. The distribution information acceptance history database together with the rate, and the pattern conformance rate and information acceptance rate at the time of the past predetermined information distribution stored in the distribution information acceptance history database, and the pattern conformance at the time of the current information distribution Calculating an information acceptance score for each pattern for each unique access pattern based on the rate and the information acceptance rate, and updating the information acceptance score stored in the information acceptance score database with the calculated information acceptance score for each pattern; ,
A content distribution method comprising: A communication system including a mobile terminal and a content distribution server that distributes information to the mobile terminal,
The portable terminal is
An access history transmission unit for transmitting the web access pattern of the current session in the mobile terminal to the content distribution server;
A distribution information receiving unit for receiving distribution information from the content distribution server;
A distribution information display unit for displaying the received distribution information;
A feedback transmission unit for transmitting the information acceptance time of the distribution information to the content distribution server;
With
The content distribution server is
An access history database that stores web access patterns for each user,
A unique access pattern database storing unique access patterns for each user;
An information acceptance score database storing an information acceptance score at the time of appearance of the unique access pattern for each unique access pattern;
A distribution information database storing distribution information to be distributed;
Distribution information acceptance history database storing pattern conformance rate history and information acceptance rate history for each unique access pattern at the time of delivery,
For each piece of distribution information accepted by the user, a distribution randomness history database that stores the distribution randomness,
A frequent access pattern extraction unit that extracts a plurality of frequent access patterns from the target user's Web access patterns stored in the access history database;
Among the extracted frequent access patterns, a unique access pattern clustering unit that clusters frequent access patterns whose mutual similarity is equal to or higher than a predetermined level, and stores the obtained access pattern in the unique access pattern database as a unique access pattern;
An access history receiving unit that receives the web access pattern of the current session in the mobile terminal from the mobile terminal and stores the access history database;
An access pattern matching unit that matches the web access pattern of the current session with the unique access pattern stored in the unique access pattern database;
Based on the matching result, a pattern matching rate calculation unit that calculates a pattern matching rate that represents the matching degree of the Web access pattern of the current session with respect to the unique access pattern;
Based on the calculated pattern matching rate and the information acceptance score for each unique access pattern stored in the information acceptance score database, a current information acceptance score is calculated, and the current information acceptance score is calculated as the information acceptance score. An information acceptance score calculation unit to be stored in the score database;
Based on the matching result, a randomness calculation unit that calculates a current randomness representing a degree of randomness of the web access pattern of the current session;
An information selection unit that selects a delivery candidate from the delivery information stored in the delivery information database based on the current information acceptance score;
An optimal randomness calculation unit that refers to the distribution randomness history database and calculates an optimal randomness for each of the selected distribution candidates from a randomness during distribution for each distribution information;
A distribution priority determining unit that determines a distribution priority for the distribution candidates based on the calculated optimal randomness for each distribution candidate and the current randomness calculated by the randomness calculating unit;
An information distribution unit that distributes the distribution information set as the distribution candidates in an order based on the determined distribution priority;
A feedback receiving unit that receives feedback of a user's information acceptance time for distributed distribution information from a distribution-destination mobile terminal;
For each piece of distribution information accepted by the user, a distribution randomness history storage unit that stores a distribution randomness in the distribution randomness history database;
An information acceptance rate is calculated based on the received information acceptance time, the delivery time of the delivery information, and a predetermined reference time interval, and the calculated information acceptance rate at the time of information delivery is calculated based on the calculated pattern adaptation. The distribution information acceptance history database together with the rate, and the pattern conformance rate and information acceptance rate at the time of the past predetermined information distribution stored in the distribution information acceptance history database, and the pattern conformance at the time of the current information distribution Information acceptance score for each unique access pattern is calculated based on the rate and the information acceptance rate, and the information acceptance score stored in the information acceptance score database is updated with the calculated information acceptance score for each pattern. A score update unit;
Comprising
A communication system characterized by the above.

Images