JP4379365B2 - Content providing apparatus, content providing system, and program - Google Patents

Description

  The present invention relates to a content providing apparatus that provides content to a user of a client device, a content providing system, and a program usable in these.

  2. Description of the Related Art In recent years, content providing systems have been proposed in which content that is presumed to be of interest to a user is retrieved and output from a user interface or other display unit based on user input to the user interface.

As a specific example, in a system that recognizes a voice taken in from a microphone and outputs a content retrieved (read) based on a keyword extracted (recognized) from the voice to a display unit (display device). Yes (see Patent Document 1).
JP 2004-164678 A

  However, in the above-described system, the detected contents are simply output sequentially, so that it takes time until the detection. Even content that has a high possibility of fading will be displayed on the display unit upon detection thereof.

  In this way, in the above-described technology, even if the content is estimated to be not suitable for the user because the user's interest is diminished, the content is displayed with the detection. However, even if it is suitable for the user's interest, such content cannot be displayed, and as a result, there is a possibility that the content estimated to be highly interested by the user cannot be provided at an appropriate timing.

  The present invention has been made to solve such a problem, and an object of the present invention is to provide a technique for providing content estimated to be of high user interest at an appropriate timing.

In order to solve the above-described problem, the content providing apparatus according to the first configuration is configured to obtain information from a client device having a user interface, at any time, one or both of input information to the user interface and output information from the user interface. According to the information acquired by the information acquisition means, the interest estimation means for estimating the interest information indicating the interest of the user of the user interface based on the information when the information acquisition means acquires the information. Content acquisition means for acquiring the content, and the content estimation by the content acquisition means based on the information used in the detection of the storage unit associated with each of the plurality of interest information. Storage unit corresponding to interest information estimated by means Content storage means for storing, and content output means for providing the content to the user of the client device by causing the output unit to output the content stored in the storage unit by the content storage unit, The content output unit selects a storage unit corresponding to the latest interest information estimated by the interest estimation unit from the plurality of storage units when outputting the content to the output unit, and the selected storage unit The content stored in is output to the output unit.

  In the content providing apparatus configured as described above, the content corresponding to the input information or output information (hereinafter referred to as “input / output information”) acquired from the client device is converted into the interest information estimated based on the input / output information. Of the contents stored in the corresponding storage unit, the content stored in the storage unit associated with the latest interest information can be output to the output unit.

  As described above, when the content is actually output, the content stored in the storage unit associated with the latest interest information is targeted, and thus the content is output after obtaining the input / output information. In the meantime, even if the interest information, that is, the user's interest has changed, it is possible to obtain information according to the current interest simply by changing the storage unit that reads the content following such change. , It is possible to provide content that is estimated to be of high user interest at the appropriate time.

  In this configuration, the content output unit is a unit that outputs the content to the output unit. For example, if the output unit has a function of displaying information, the content output unit displays the content on the output unit. What is necessary is just to comprise. Further, if the output unit has a function of outputting sound, it may be configured to output sound indicating content. Further, if the output unit has a function of printing an image on a recording medium, the image indicating the content may be printed on the recording medium.

  In addition, the specific configuration for the content acquisition unit to acquire the content is not particularly limited. For example, if the content providing device itself has various contents, input / output from these contents A configuration can be considered in which information corresponding to information (including information related thereto, and so on) is read and acquired.

In addition, when configured to be communicable with a content search unit having a function of searching for content based on information, as in the second configuration , the content acquisition unit is acquired by the information acquisition unit. A content search unit having a function of searching for content based on the received information and the interest information estimated by the interest estimation unit, and requests the content search based on the acquired information, and in response to the request What is necessary is just to comprise so that the searched content may be acquired.

If comprised in this way, the processing burden concerning acquisition of content can be reduced.
Further, the content search realized by such a content search unit may be configured to be performed by the content providing device itself. In this case, for example, the content search is performed based on information acquired by the information acquisition unit. Content search means for searching may be provided, and the content acquisition means may be configured to acquire the content detected by causing the content search means to search for content when information is acquired by the information acquisition means. In this case, the content providing device itself can search and acquire the content based on the input / output information.

  In the content providing apparatus, the content output by the output unit is the content stored in the storage unit corresponding to the latest interest information. Here, a plurality of content items are stored in the corresponding storage unit. In such a case, it may be configured to preferentially output, for example, recently stored content, recently acquired content, or randomly selected content. Note that “preferentially output” means, for example, that only the content to be prioritized is output at the output unit, or that the content to be prioritized can be specified in a manner that can be specified in the output unit together with other content. And so on.

  Further, the content search unit may search for content based on the interest information estimated by the interest estimation means. By comprising in this way, the content close | similar to a user's interest can be searched from a search step, and it can provide to a user.

Further, an evaluation value evaluated based on some evaluation criterion may be assigned to each content stored in the same storage unit, and a content with a high evaluation value may be preferentially output. In this case, for example, a degree of similarity such as how much content is similar to the latest interest information may be adopted as an evaluation value. For this purpose, the third configuration may be adopted. Conceivable.

The content providing apparatus according to the third configuration evaluates, for each content acquired by the content acquisition unit, how much the content is similar to the latest interest information estimated by the interest estimation unit The content storage means stores the content acquired by the content acquisition means in the storage unit together with the evaluation result by the similarity evaluation means for the content, and the content output means After selecting the storage unit corresponding to the latest interest information estimated by the interest estimation unit, among the contents stored in the storage unit, the content having a higher degree of similarity in the evaluation result by the similarity evaluation unit is given priority. It is configured to output to the output unit.

  If comprised in this way, when several content is memorize | stored in the memory | storage part corresponding to the newest interest information, content with a high degree similar to the newest interest information will be output preferentially among these contents. Is done. For this reason, the latest interest information is stored in the associated storage unit, and the content that is estimated to be highly interested by the user is highly similar to the latest interest information. Content presumed to be even higher can be preferentially provided.

In this configuration, as a specific configuration for causing the content output means to preferentially output content having a high degree of similarity in the evaluation result by the similarity evaluation means, for example, the fourth configuration is used. Can think.

In the content providing apparatus according to the fourth configuration, each time the content is stored in the storage unit by the content storage unit in the case where the storage unit is configured to sequentially store the content by the queue method, Order reordering means for reordering each of the contents stored in the storage unit so that the content that is more similar in the evaluation result by the similarity evaluation means is read more quickly. The content output unit outputs a content stored in the storage unit according to the order rearranged by the order rearrangement unit after selecting the storage unit corresponding to the latest interest information estimated by the interest estimation unit. It is comprised so that it may output to a part.

  With this configuration, since the content is stored in the storage unit according to the queue method, the content is normally read from the storage unit in the stored order. Since the content having a high degree of similarity is rearranged so as to be read first, the content having a high degree of similarity to the latest interest information can be read by simply reading according to this arrangement order. Therefore, when outputting the content, it is only necessary to read the content according to the queue method from the storage unit corresponding to the latest interest information, and each time the content is similar to the latest interest information from the corresponding storage unit. There is no need to identify high content.

  By the way, as described above, in the configuration in which the degree of similarity such as how much content is similar to the latest interest information is used as the evaluation value, the evaluation is performed only at the time when the content is acquired. In this case, if the latest interest information changes, the evaluation result will not be highly similar to the latest interest information. There is a possibility that content that is not content to be provided is preferentially provided.

  In order to prevent such a situation, for example, the evaluation is performed not only at the time when the content is acquired but also periodically thereafter, and the evaluation result stored together with the content is updated. Can be considered.

  Further, as another configuration, after the content is acquired, the evaluation result stored in the storage unit together with the content is updated as needed to the degree of similarity to the latest interest information according to the elapsed time thereafter. Such a configuration is also conceivable.

As a configuration for this, for example, as in the fifth configuration , the content storage unit stores the content acquired by the content acquisition unit together with timing information indicating a specific time after the content is acquired. The evaluation result is stored in the storage unit together with the content according to the specific time indicated by the time information stored in the storage unit together with the content by the content storage unit. It is conceivable to have an evaluation result changing means for changing

If comprised in this way, the evaluation result memorize | stored with a content can be changed according to the specific time after the content was acquired.
The “specific time” in this configuration is a specific time after the content is acquired. For example, the time or date and time when the content is acquired, or the time or date and time when the content is stored in the storage unit Etc. can be adopted.

In this configuration, after the content is acquired, it is expected that the degree of similarity described above will decrease with the passage of time thereafter. Therefore, the evaluation result stored together with the content is The configuration may be such that the degree of similarity with the latest interest information is lowered according to the elapsed time from the specific time, and the configuration for this is, for example, the configuration of the sixth configuration do it.

In the content providing apparatus according to the sixth configuration, the degree of similarity decreases as the elapsed time from the time indicated by the time information stored in the storage unit together with the content to the current time increases. As described above, the content storage means changes the evaluation result stored in the storage unit together with the content.

  With this configuration, as the elapsed time from a specific time becomes longer, the degree of similarity between the evaluation result stored with this content and the latest interest information for the content that is expected to be less interested by the user Therefore, it is possible to prevent content with less interest of the user from being preferentially provided over time.

  In addition, as described above, after the content is acquired, it is expected that the degree of similarity will decrease with the passage of time thereafter. However, the content with the passage of time has increased to some extent. There is also a possibility that the user's interest itself may be completely lost rather than the above-mentioned similarity degree being lowered. Therefore, it can be said that it is desirable from the viewpoint of preferentially providing content presumed to be highly interested by the user to be able to delete from the storage unit content whose elapsed time has been increased to some extent after acquisition.

Therefore, in order to realize this, for example, as shown in the seventh configuration , among the contents stored in the storage unit by the content storage unit, the time indicated by the time information stored together with the content It is preferable to include a content deleting unit that deletes content whose value is more than a predetermined threshold value from the storage unit.

  According to this configuration, the degree of similarity decreases according to the elapsed time from a specific time, and a time (predetermined threshold) that is estimated to have completely lost the user's interest has elapsed. Since the content can be deleted from the storage unit, it is possible to prevent the content presumed that the user's interest has been lost from being preferentially provided. .

The content providing system according to the eighth configuration includes the content providing apparatus according to any one of the first to seventh configurations .
According to the content providing system configured as described above, the same operations and effects as those obtained by the content providing apparatus according to any one of the first to seventh configurations can be obtained.

  In addition to the above-described content providing apparatus, this content providing system is a system including one or more of the above-described client device, an information collecting apparatus including content search means, and the like as components.

The program according to the ninth configuration is a program for causing a computer system to execute various processing procedures for causing the content providing apparatus according to any of the first to seventh configurations to function as all means. is there.

According to the computer system controlled by such a program, a part of the content providing apparatus according to any one of the first to seventh configurations can be configured.
The above-described program is composed of an ordered sequence of instructions suitable for processing by a computer. For example, the program is recorded via a recording medium such as an FD, a CD-ROM, a memory card, or a communication line network such as the Internet. Thus, it is provided to a content providing apparatus or a user who uses them. In addition, as a form of providing these programs to the user, it is also possible to provide a form pre-installed in a hard disk or memory.

Embodiments of the present invention will be described below with reference to the drawings.
1. First Embodiment (1) Overall Configuration of System As shown in FIG. 1, this system 1 is connected to a WAN (Wide Area Network) such as the Internet by a web server 10 having various contents, an information search server 20, and the like. A personal computer (hereinafter referred to as “PC”) 30, a digital telephone device (hereinafter referred to as “telephone device”) 50, and information collection are connected to a LAN (local area network) connected to the WAN via a router R or the like. This is a system in which an analysis device 70 and the like are connected.

  The information retrieval server 20 includes a control unit 21 composed of a CPU, ROM, RAM, etc. (not shown), a storage unit 23 composed of a hard disk device, etc., a communication unit 25 that controls data communication with a WAN or LAN, and the like. ing.

  Among these, the storage unit 23 stores a database including address information (URL information) groups of various contents that can be provided by the web server 10, various search programs, and the like. This search program is activated when a query (inquiry) is received from the outside, and the process of searching for content based on this query and returning the search result to the sender of the query is controlled. This is executed by the unit 21.

  The personal computer 30 controls data communication with a control unit 31 composed of a CPU, ROM, RAM, etc. (not shown), a storage unit 33 composed of a hard disk device, a monitoring unit 35 described later, and a WAN or LAN. The communication unit 37 includes a user interface 40 configured by an input unit 41 that receives user input operations and a display unit 43 that displays various types of information.

  Among these, the storage unit 33 stores various application software (hereinafter simply referred to as “text data creation based on text input by the input unit 41”, display of text indicated by the text data by the display unit 43, etc.). "Software") is installed.

  In addition, the monitoring unit 35 monitors predetermined software among the software executed by the control unit 31 as a data collection target, and the text input from the input unit 41 or the display unit 43 based on the software. It is configured to execute a process of collecting text data indicating the displayed text from the software (a personal computer monitoring process described later).

  The telephone device 50 controls data communication with a control unit 51 including a CPU, ROM, RAM, and the like (not shown), a storage unit 53 including a hard disk device, a monitoring unit 55 described later, and a WAN or LAN. A communication unit 57, a sound input unit 61 for inputting sound and converting it to sound data, and a sound output unit 63 for outputting sound represented by the sound data. An IP (Internet Protocol) telephone function that communicates with an external telephone device via the network can be realized.

  Among these, the monitoring unit 55 monitors voice data exchanged via the user interface 60, and is configured to execute processing for collecting the voice data (telephone monitoring processing described later). Yes.

  The information collection and analysis device 70 includes a control unit 71 composed of a CPU, ROM, RAM, etc. (not shown), a storage unit 73 composed of a hard disk device, etc., and a communication unit that controls data communication with the WAN or LAN 75, a display unit 77 for displaying various information, a sound output unit 79 for outputting sound represented by audio data, and the like.

Among these, the storage unit 73 stores various programs executed by the control unit 71 and various data tables used when the programs are executed.
In addition, as shown in FIG. 2, the display unit 77 is an area for displaying text and images, and a text / image area for scrolling and displaying new information, and a moving image for displaying the reproduction result of the moving image file. This is for displaying a content providing screen including a reproduction area and a term display area for displaying a term search result.
(2) Personal computer monitoring process by personal computer 30 Here, the processing procedure of the personal computer monitoring process executed by the monitoring unit 35 of the personal computer 30 will be described with reference to FIG. This personal computer monitoring process is a process that is repeatedly executed after the personal computer 30 is activated.

  When this personal computer monitoring process is started, first, it is checked whether or not there is software that holds text data that has not been collected at the present time among predetermined software to be collected (s110).

  If it is determined in s110 that the corresponding software does not exist, the process waits until any of the data collection target software holds uncollected text data (s110: NO). On the other hand, if it is determined that there is software that holds uncollected text data (s110: YES), the timer is reset and counting is restarted (s120).

  Next, the document mode of the software determined to hold uncollected text data in s110 is checked (s130). This “document mode” can be determined by receiving a notification from the software when the software is checked in s110, and the software determines the text data and the display unit according to the input operation to the input unit 41. When the display contents on 43 are sequentially updated, it is “document creation mode”, and when the text indicated by the text data designated by the input unit 41 is displayed on the display unit 43. It is notified that it is in “document browsing mode”. Therefore, in s130, the document mode is determined according to the notification content from the software.

  If it is determined in s130 that the mode is “document creation mode” (s130: YES), device usage information indicating “document creation” is generated (s140), while the mode is not the document creation mode (that is, in the document browsing mode). If it is determined that there is (S130: NO), device usage information indicating “document browsing” is generated (s145).

  After generating the device usage information in this way, writing of the text data as a transmission target file is started while collecting uncollected text data from the software (s150).

  Thereafter, the writing of the text data to the transmission target file is performed until the count value of the timer restarted in s120 passes a predetermined unit time T1 (1 minute in the present embodiment) (s160: NO). Source generation time information indicating the current time when the unit time T1 has elapsed (s160: YES) is generated (s170).

Then, a transmission target file in which text data for unit time T1 is written, source generation time information generated in s170, device usage information generated in s140 or s145, and input indicating that the transmission source is the personal computer 30 After associating the original information and transmitting these as source information to the information collecting and analyzing apparatus 70 (s180), the process returns to s110 and waits until the next uncollected data is generated.
(3) Telephone Monitoring Processing by Telephone Device 50 Next, the telephone monitoring processing procedure executed by the monitoring unit 55 of the telephone device 50 will be described with reference to FIG. This telephone monitoring process is a process that is repeatedly executed after the telephone device 50 is activated.

  When the telephone monitoring process is activated, the process waits until the telephone device 50 is connected to another telephone device (communication line is closed) (s210: NO), and the connection to the other telephone device is established. If it is made (s210: YES), it is checked whether or not the connection is caused by the telephone device 50 calling (s220).

  If it is determined in s220 that the telephone device 50 itself has caused a call (s220: YES), device usage information indicating “calling” is generated (s230), while other phone calls are made. If it is determined that the cause is that the device originates a call (the telephone device 50 has received a call), device usage information indicating “incoming call” is generated (s235).

  Thus, after the device usage information is generated, the audio data input / output by the sound input unit 61 and the sound output unit 63 is acquired until the unit time T2 elapses thereafter, and the audio indicated by the audio data is obtained. The text data is converted into text data by a known voice recognition technique and written into a transmission target file (s240).

  Next, after generating source generation time information indicating the current time (s250), a transmission target file generated in s240, source generation time information generated in s250, device usage information generated in s230 or s235, and Input source information indicating that the transmission source is a telephone device is associated and transmitted as source information to the information collection and analysis device 70 (s260).

Then, after transmitting the data in s260, it is checked whether the connection has been disconnected (communication line has been released). If it is determined that the connection has not been disconnected (s270: NO), the process returns to s240. Then, while repeating the process of converting the audio data for the next time T2 into text data, if it is determined that the connection has been disconnected (s270: YES), the process returns to s210 and waits until the next connection is made ( s210: NO).
(4) Processing by Information Collection Analysis Device 70 Various processes executed by the information collection analysis device 70 will be described below.
(4-1) Source Information Registration Process First, the processing procedure of the source information registration process repeatedly executed by the control unit 71 after the information collection / analysis apparatus 70 is started will be described with reference to FIG.

  When this source information registration process is started, first, either source information transmitted from the personal computer 30 at s180 in FIG. 3 or source information transmitted from the telephone device 50 at s260 in FIG. (S310: NO).

  When the source information is received in s310 (s310: YES), a predetermined file name is assigned to the transmission target file in the source information and stored in the storage unit 73 (s320).

Then, a record including the file name (or path name) of the transmission target file stored in s320 and the source generation time information, device usage information, and input source information among the source information received in s310 is sourced. After registering in the management table (s330), the process returns to s310 and waits until the next source information is received. In the present embodiment, as shown in FIG. 6, data for registering a record in which source generation time information, device usage information, and input source information are associated with the file name of a transmission target file as a source management table. Since the table is stored in the storage unit 73, in s330, each information is registered in the corresponding record in this table.
(4-2) Keyword Extraction Processing Subsequently, the processing procedure of the keyword extraction processing that is repeatedly executed in parallel with other processing by the control unit 71 after the information collection analysis device 70 is started will be described with reference to FIG. .

  When this keyword extraction processing is started, first, it is checked whether or not an unprocessed record that is not a processing target in the subsequent processing is registered in the above-described source management table (s410). If the unprocessed record is not registered, the process waits until the unprocessed record is registered in the above-described source information registration process (s410: NO).

  On the other hand, when it is determined in s410 that an unprocessed record is registered (s410: YES), the record is set as a record to be processed (hereinafter referred to as “target record” in this section). The text data of “file name” in this target record is read, this text data is analyzed by a well-known morphological analysis, and the noun included in the text indicated by the text data is extracted. Among them, nouns other than the specific noun are extracted as keywords (s420). Here, general nouns that are not suitable for use as keywords when searching for content are defined as specific nouns, and nouns obtained by excluding such specific nouns from nouns extracted by morphological analysis are extracted as keywords.

  Next, an interest vector is calculated by starting an interest vector calculation process described later based on the device usage information in the target record and the keyword group including one or more keywords extracted in s420 (s430). This interest vector calculation process is a process for calculating a vector obtained by quantifying the interest information indicating the user's interest as an interest vector, and this “interest vector” corresponds to each type of interest information determined in advance. This is a 1-by-n vector (V = v1, v2,..., Vn) composed of the numerical elements (v1 to vn) attached.

  Next, a new record is added to the search information management table (s450). In the present embodiment, as a search information management table, a data table for registering source generation time information, interest vectors, search keys, the number of search keys, and the like is stored in the storage unit 73 as shown in FIG. Yes. In s450, a new record necessary for registering the information in the subsequent processing is added to the search information management table.

Then, the time indicated by “source generation time information” in the target record of the source management table is registered in “source generation time information” in the new record registered in the search information management table in s450 (s460), and the new The interest vector calculated in s430 is registered in the “interest vector” in the new record (s470), and each keyword extracted in s420 is registered in the “search key” in the new record (s480). After registering the number of keywords extracted in s420 in the “number of search keys” in a simple record (s490), the process returns to s410, and the processes after s410 are repeatedly performed on the unprocessed records.
(4-3) Vector Calculation Processing Next, the processing procedure of the vector calculation processing which is s430 in FIG. 7 will be described based on FIG.

  When the interest vector calculation process is started, first, the interest vector V is initialized by setting “0” for each element (v1 to vn) for the interest vector V (s510). .

  Next, from the keyword group passed from the keyword extraction process at the time of starting the vector calculation process, one of the keywords not processed in the subsequent processes is selected as a process target (s520). In the present embodiment, even if a plurality of identical keywords are included in the keyword group passed from the keyword extraction process, the subsequent processes are performed after each is selected separately as a processing target. It will be.

  Next, the keyword selected in s520 (hereinafter referred to as “target keyword”) is searched from the keywords registered in the interest estimation table (s530). In the present embodiment, as shown in FIG. 10, as the “interest estimation table”, for each of a plurality of predetermined types of interest information (v1 to vn), keywords belonging to that type of interest information, and A data table in which definition information used for weighting the interest vector in the subsequent processing is associated is stored in the storage unit 73. Therefore, in s530, it is checked whether or not the same keyword as the target keyword is registered in the interest estimation table.

  If the target keyword is detected from the interest estimation table in s530 (s540: YES), the type of interest information corresponding to the keyword thus detected is specified as the type of interest information in the target keyword (s550). After adding “1” to the value of the element (any one of v1 to vn) in the interest vector corresponding to the type of information (s560), the process proceeds to the next process (s570).

  On the other hand, if the target keyword is not detected from the interest estimation table in s530 (s540: NO), the process proceeds to the next process (s570) without performing the above-described s550 and s560.

  Next, it is checked whether or not there is a keyword that has not been processed so far in the keyword group passed from the keyword extraction process (s570), and if there is a keyword that has not been processed, (S570: YES), the process returns to s520, and thereafter s520 to s570 are repeated for each keyword in the keyword group passed from the keyword extraction process.

  Thereafter, in s570, if there is no keyword that is not a processing target, that is, after the above-described processing (s520 to s560) has been repeatedly performed on all keywords (s570: NO), the keyword is registered in the interest estimation table. After weighting the current interest vector based on the definition information corresponding to the target keyword among the defined definition information (s580), the weighted interest vector is returned as a return value to s440 in FIG. Transition.

Here, the definition information in the interest estimation table will be described in detail. The interest estimation table includes device definition information (v1, v2 in FIG. 10) configured by device usage information and weighting coefficients according to the type of interest information. , V4), or time definition information (see definition information corresponding to v3 in FIG. 10) composed of time information indicating a predetermined time zone and a weighting coefficient is registered as definition information. Yes. In s580, if the definition information corresponding to the same keyword as the target keyword is device definition information, a value indicated by the device definition information and a value indicated by the device usage information passed from the keyword extraction process are obtained. Weight only if they match. On the other hand, if the definition information corresponding to the same keyword as the target keyword is time definition information, weighting is performed only when the current time is within the time zone indicated by the time information indicated by the definition information. Here, specific weighting is performed by multiplying each element of the interest vector by a weighting coefficient W (vi · W → vi; i = 1 to n).
(4-3) Search Control Processing Subsequently, the processing procedure of the search control processing that is repeatedly executed in parallel with other processing by the control unit 71 after the information collection / analysis apparatus 70 is started will be described with reference to FIG. .

  When this search control process is started, first, it is checked whether or not an unprocessed record that is not a process target in the subsequent process is registered in the search information management table described above (s610). If no unprocessed record is registered, the process waits until an unprocessed record is newly registered (s610: NO).

  On the other hand, when an unprocessed record is registered in s610 (s610: YES), the time indicated by “source generation time information” is the latest (most current) among the registered records. (S620) and having the largest “number of search keys” is selected as a processing target record (hereinafter referred to as “target record” in this section) (s630).

  Next, based on the search program management table, a search program that is not currently activated is selected (s640). In this embodiment, as shown in FIG. 12, as a “search program table”, a search program indicating address information of a server (in this embodiment, the information search server 20) having a plurality of search programs to be executed. A data table in which address generation, query generation information indicating a search expression and query generation method, category information indicating a category of content searched by the search program, and the like are registered in the storage unit 73 in advance. In s640, the search program address registered in the search program management table is selected as a search program that has not been activated as a search program address that is not referenced when issuing a query for the target record.

  Next, based on the target record, a search process to be described later for starting the search program selected in s640 is started (s650). Note that the search information management table described above is a queuing data table (queue), and the target records selected in s620 and s630 are processed from the search information management table when the search process is started. Will be deleted as a completed record.

  Then, based on the search program management table, whether or not there is a search program that has not been started at present, that is, issuance of a query for all search program addresses registered in the search program management table for the target record Is checked (s660), and if it is determined that there is a search program that has not been started (s660: YES), the process returns to s640, and the search program is searched for the target record. Steps s640 to s660 are repeated until the reference for issuing the query for all the search program addresses registered in the management table is completed.

Thereafter, when it is determined in s660 that there is no search program that has not been activated (s660: NO), the process returns to s610, and the processes after s610 are repeatedly performed on unprocessed records.
(4-4) Search Process Next, the process procedure of the search process activated in s650 in FIG. 11 will be described based on FIG. This search process is a process executed in parallel with the search control process described above.

  When this search processing is started, first, a content search formula and a query that stores this are based on the “query generation information” in the record selected from the search program table in s640 of FIG. A query is generated (s710). Here, based on the above-described query generation information, each of the content search formulas including the “search key” of the target record (record in the search information management table) passed from the search control process when starting this search process is Each query that is generated and stored is generated.

  Next, the query generated in s710 is transmitted to the address information indicated by “search program address” of the target record passed from the search control process, so that the search program corresponding to the target record is sent to the server. (In the present embodiment, the information search server 20) is activated (s720). From the server that has received this query, the address information of one or more contents detected as a result of the search of the contents in accordance with the search expression stored in the query is returned as a response to the query.

  In this way, after transmitting the query in s720, it waits until there is a response from this server (s730: NO). If there is a response (s730: YES), the response is not a response to the query (s740: NO), the search process is terminated.

  On the other hand, if the response received in s730 is a response to the query, that is, a content search result (s740: YES), the “category information” of the target record passed from the search control process is checked (s750).

  If the category information indicates “term search” in s750 (s750: YES), the search result is information related to one piece of address information among the address information of one or more contents as a response from the server. Register in the management table (s760). In the present embodiment, as a “search result management table”, a data table that registers source generation time information, category information, content storage destination addresses, and evaluation values in association with each other is stored in the storage unit 73 as shown in FIG. Has been. In s760, information related to any one address information among one or more address information that is a response from the server is registered in one record in the search result management table. Specifically, in the record, the address information of the target record passed from the search control process is registered as “content storage destination address”, and the source generation time information of the target record is registered as “source generation time information”. The information indicating the term search is registered as “category information”.

  In addition, when the category information indicates “general search” in s750 described above (s750: NO), a predetermined number (in this embodiment) of the address information of one or more contents that are responses from the server. Register information on the address information of 5 cases (the number of cases where the address information included in the response from the server is 4 or less) in the search result management table (s770). Here, among the one or more address information that is a response from the server, information related to a maximum of five address information is registered in different records in the search result management table. Specifically, in each different record, the address information of the target record passed from the search control process is registered as “content storage destination address”, and the source generation time information of the target record is registered as “source generation time information” Then, information indicating general search is registered as “category information”, and the corresponding address information is registered as “content storage destination address”.

In this way, after registering in the search result management table in s760 and s770, the search process is terminated.
(4-5) Search Result Evaluation Process Subsequently, the processing procedure of the search result evaluation process that is repeatedly executed in parallel with other processes by the control unit 71 after the information collection / analysis apparatus 70 is started based on FIG. explain.

  When the search result evaluation process is started, first, it is checked whether or not an unprocessed record not to be processed in the subsequent processes is registered in the search result management table (s1110), and an unprocessed record is checked. If it is determined that no record is registered, the process waits until an unprocessed record is registered in the above-described search process (s1110: NO).

If it is determined that an unprocessed record is registered in s1110 (s1110: YES), the record is set as a record to be processed (hereinafter referred to as “target record” in this section), and the interest vector in this record and the current An inner product with the interest vector in the interest table is calculated, and the value is registered as an evaluation value in the target record (s1120). In this embodiment, every time an unprocessed record is registered in the search result management table as the “current interest table”, as shown in FIG. 16A, the “interest vector” in the unprocessed record and An update time related to the update is updated by the control unit 73 (s1250, s1260 in FIG. 17), and a data table is stored in the storage unit 73. In s1120, first, the interest vector V in the target record is divided by the norm || V || of the interest vector V to obtain a normal vector E (= V / || V ||) for the interest vector V. calculate. Similarly, the interest vector Vc in the current interest table is also divided by the norm || Vc || to calculate a normal vector Ec (= Vc / || Vc ||) of the interest vector Vc. Then, an inner product R (= E · Ec) for these normal vectors E and Ec is calculated, and the value of the inner product R is registered as an evaluation value in the target record.

  Next, among the interest vectors in the memory interest table, an interest vector having the largest inner product with the interest vector in the current interest table is specified, and a storage area corresponding to the interest vector is specified (s1130). In this embodiment, as a “memory interest table”, as shown in FIG. 16B, a storage area (reference address) for storing content in the subsequent processing is a content to be stored in the storage area. A data table in which the interest vectors are registered in association with each other is stored in the storage unit 73. In s1130, for each interest vector registered in the memory interest table, an inner product with the interest vector registered in the current interest table is calculated by the same method as described above, and a value indicating the inner product is calculated. A storage area corresponding to the largest interest vector is specified.

  The plurality of storage areas used when outputting the contents are storage areas allocated in the storage unit 73 in advance, and are respectively a term display data queue, a video playback data queue, and an audio output data queue. , A data queue for displaying text images.

  Next, content is acquired based on the “content storage destination address” in the target record (s1140). Here, by sending a request to the access destination indicated by the content storage destination address in the target record, the content returned from the access destination that has received this request is acquired.

  Next, the type of content acquired in s1140 is checked (s1150). If the content is data (text data) indicating information such as a term (s1150: YES), the content is set to “source” in the target record. Along with the “generation time information” and “evaluation value (registered in s1120)”, data for term display in the storage area registered in the memory interest table in association with the interest vector specified in s1130 After storing in the queue (s1160), the process proceeds to the next process (s1220).

  Further, when the content acquired in s1140 is data (moving image data) indicating a moving image (s1150: NO, s1170: YES), the content is converted into “source generation time information” and “evaluation value (s1120) in the target record. In addition, after being stored in the data queue for moving image reproduction in the storage area registered in the memory interest table in association with the interest vector specified in s1130 (s1180), The process proceeds to processing (s1220).

  Further, when the content acquired in s1140 is data (audio data) indicating audio (s1170: NO, s1190: YES), this content is set to “source generation time information” and “evaluation value (s1120) in the target record. Registered in the memory interest table in association with the interest vector specified in s1130, and then stored in the data queue for audio output (s1200), and then The process proceeds to (s1220).

  When the content acquired in s1140 is data (text data, image data) indicating information other than the term (s1190: NO), the content is converted into “source generation time information” and “evaluation value ( In the storage area selected in s1130 and stored in the data queue for text image display (s1210), the process proceeds to the next process (s1220).

  Next, among the contents stored in all the data queues in each of the plurality of storage areas used when outputting the contents, the time indicated by the “source generation time information” stored together with the contents is determined from the current time. The content (and information stored together with the content) whose time has passed for a predetermined time (10 minutes in the present embodiment) or more is deleted from the data queue (s1220).

Then, the contents stored in all the data queues in each of the plurality of storage areas used when outputting the contents are sorted in descending order of the values indicated by the evaluation values stored together with the contents (the larger the order of reading, After rearranging (to be faster) (s1230), the process returns to s1110 and waits until an unprocessed record is registered (s1110: NO).
(4-6) Output Control Processing Subsequently, the processing procedure of the output control processing that is repeatedly executed in parallel with other processing by the control unit 71 after the information collection / analysis apparatus 70 is started will be described with reference to FIG. .

  When this output control process is started, first, the interest vector having the largest inner product with the interest vector in the current interest table is specified from the interest vectors in the memory interest table, as in s1130 in FIG. 15 (s1310). .

  Next, the storage area to be referred to when outputting the content is switched to the storage area registered in the memory interest table in association with the interest vector specified in s1310 (s1320).

  Next, it is checked whether or not there is content that has not been output in the subsequent processing in the storage area switched in s1320 (s1330), and when it is determined that there is no content that has not been output (s1330: NO). , Return to s1310.

  On the other hand, if it is determined that there is content that has not been output (s1330: YES), the content is read from the data queue of the storage area in which the content is stored and displayed on the content provision screen (see FIG. 2) (s1340).

  Here, content is output by displaying the content stored in the data queue in the storage area switched in s1320 on the content providing screen. At this time, unoutput content is stored in different data queues. If all the contents are output, and if a plurality of contents are stored in the same data queue, only one content is output according to the arrangement order of the contents in the data queue. .

  As for a specific content output mode, for example, if a data queue storing unoutput content is a term display data queue, the content is displayed in a term display area on the content providing screen. At this time, if other content is already displayed in the term display area, the display of this content is erased, and the content that has not been output is preferentially displayed.

  Further, if the data queue storing the unoutput content is a data queue for moving image reproduction, based on the content, moving image reproduction to the moving image reproduction area on the content providing screen, and from the sound output unit 79 Start audio output. At this time, if the reproduction and output of other content has already been started, the reproduction and output of this content are stopped, and the content that has not been output is preferentially reproduced and output.

  If the data queue storing the unoutput content is a data queue for audio output, the audio output unit 79 starts to output the audio indicated by the content. At this time, if the output of audio indicated by other content has already been started, the audio output is interrupted, and the output of audio indicated by the content not yet output is preferentially started.

  Further, if the data queue storing unoutput content is a data queue for text image display, the content is displayed so as to scroll from the bottom to the top of the text / image display area on the content providing screen. At this time, when other content is already displayed in the text / image display area, the content is scrolled as the content is scrolled from the state in which the unoutput content is arranged below the content. Display.

Thus, after outputting unoutput content, the process returns to s1330, and thereafter, s1330 and s1340 are repeated until there is no unoutput content, that is, until output of all unoutput content is completed.
(5) Operation and Effect The information collection / analysis device 70 of the content providing system 1 as described above extracts a keyword group from the transmission target file acquired from the personal computer 50 or the telephone device 50 (s420 in FIG. 7), and this keyword group. After searching the content based on the content (s710 to s730 in FIG. 13), the content is stored in the content output data queue (s1150 to s1210 in FIG. 15), and then displayed on the content providing screen or the sound output unit 79. (S1340 in FIG. 18).

  Here, the content searched based on the keyword group is an interest vector calculated based on the keyword group referred to in the search among a plurality of storage areas used when the content is output, that is, an interest indicating the user's interest. Information is stored in a data queue that constitutes a storage area corresponding to a vector obtained by digitizing information (s1130 in FIG. 15).

  When the content thus stored is actually output, the interest vector (FIG. 16 (b)) corresponding to the storage region of the storage region used when the content is output and the interest vector of the current interest table. The storage area having the highest inner product with the interest vector, that is, the highest similarity to the interest vector is switched to the reference destination (s1310 and s1320 in FIG. 18), and the content stored in the data queue of the storage area is output. (S1340 in the figure).

  This current interest table is updated every time an unprocessed record is registered in the search result management table. Therefore, the latest interest vector reflecting the interest of the user who changes over time is updated. Can be said to be registered.

  Therefore, as described above, the latest interest information is stored in the associated recording area by outputting the content stored in the data queue of the storage area having the highest similarity to the interest vector of the current interest table. Content can be output.

  As described above, when the content is actually output, the content stored in the storage area associated with the latest interest information is the target, and therefore, until the content is output after the keyword group is acquired. In the meantime, even if the interest information, that is, the user's interest changes, it is possible to provide content that is estimated to be highly interested in the user at the appropriate time following such a change.

  Further, in actually outputting the content searched based on the keyword group, after assigning an evaluation value indicating the inner product of the interest vector in the content and the interest vector of the current interest table (s1120 in FIG. 15), The contents are stored in a storage area used for outputting (s1150 to s1210 in the figure), and then rearranged so that the higher the evaluation value assigned earlier is in the earlier order of reading (s1230 in the figure), sequentially. The content is output (s1340 in FIG. 18). As described above, the larger the evaluation value indicating the inner product with the interest vector of the current interest table, that is, the content that can be evaluated (estimated) that the similarity with the latest interest vector reflecting the user's interest is higher, from the data queue. Such content can be preferentially provided because it is read out and output first.

  In addition, since the storage area used to output the contents is configured with data queues stored according to the queue method, the contents are normally read and output from these data queues in the stored order. However, these contents are rearranged so that the contents that can be evaluated as having high similarity to the latest interest vector reflecting the user's interest are read out first (s1230 in FIG. 15). By simply reading in order, content that can be evaluated as having high similarity can be read sequentially. Therefore, when outputting the content, it is only necessary to read the content according to the cue method from the storage area corresponding to the latest interest vector reflecting the user's interest. There is no need to specify content that can be evaluated as having high similarity to the latest reflected interest vector.

Further, when the content searched based on the keyword group is actually output, after each content is stored in the data queue (s1150 to s1210 in FIG. 15), “source generation time information” stored together with the content is stored. Since the content whose time indicated by is a time when a predetermined time or more has elapsed from the current time is deleted (s1220 in the figure), such content is not output. This “source generation time information” indicates the time at which the transmission target data from which the keyword group is extracted is generated in the personal computer 30 or the telephone device 50 (s170 in FIG. 3, s250 in FIG. 4). As described above, for content whose time has passed a predetermined time or more from the current time, it is highly likely that the user's interest has faded or lost as the time elapses. . Therefore, by deleting such content before output, it is possible to output content that is presumed that such user's interest has faded or has been lost. Can be prevented.
2. Second Embodiment The content providing system according to the present embodiment has the same configuration as the content providing system 1 according to the first embodiment, and only part of the processing content is different. Therefore, only the differences will be described in detail. To do.
(1) Processing by Information Collection Analysis Device 70 Various processes executed by the information collection analysis device 70 will be described below.
(1-1) Search Result Evaluation Processing A processing procedure of search result evaluation processing that is repeatedly executed in parallel with other processing by the control unit 71 after the information collection / analysis apparatus 70 is started will be described with reference to FIG. In addition, the same step number is attached | subjected about the process similar to the search result evaluation process in 1st Embodiment.

  When this search result evaluation process is started, first, as in the first embodiment, after performing s1110 to s1140, based on the target record of the search result management table, an evaluation value update process to be described later is started (s1410). ). This evaluation value update process is a process for updating the evaluation value for the content acquired in s1140 with the passage of time, and is a process performed in parallel with other processes including the search result evaluation process. .

Next, as in the first embodiment, after performing s1150 to 1410, among the contents stored in all the data queues in each of the plurality of storage areas used when outputting the contents, the contents are stored together with the contents. The content (and information stored together with the content) whose value indicated by “evaluation value” is equal to or less than a predetermined threshold (“0” in the present embodiment) is deleted from the data queue (s1420). Control goes to s1230.
(1-2) Evaluation Value Update Processing Next, the processing procedure of the evaluation value update processing that is s1410 in FIG. 19 will be described with reference to FIG.

  When this evaluation value update process is started, first, the time indicated by “source generation time information” among the target records in the search result management table passed from the search result evaluation process when starting this evaluation value update process The elapsed time from the current time to the current time is calculated (s1510).

  Next, it is checked whether or not the elapsed time calculated in s1510 has reached the time to update the “evaluation value” in the target record (s1520). In the present embodiment, as shown in FIG. 21, an elapsed time coefficient table, which is a data table in which an elapsed time coefficient to be added to the “evaluation value” of the target record is associated with each elapsed time and registered. It is stored in the storage unit 73. Therefore, here, when the elapsed time coefficient corresponding to the elapsed time calculated in s1510 is registered in the elapsed time coefficient table, the elapsed time is set to the time when the “evaluation value” in the target record is to be updated. Judge that it has reached.

  If it is determined in s1520 that the time to be updated has not been reached (s1520: NO), the process returns to s1510. On the other hand, if it is determined that the time to be updated has been reached (s1520: YES), the “evaluation value” in the target record is registered in the elapsed time coefficient table in association with the elapsed time calculated in s1510. The value obtained by integrating the elapsed time coefficients is updated to a new “evaluation value” in the target record (s1530). In this s1530, the “evaluation value” in the target record is updated to a function indicating the product of the original evaluation value X and the elapsed time coefficient Y calculated in s1510, and thereafter, this evaluation value is referred to. In other processing, a value calculated by this function is referred to.

Thus, after updating the “evaluation value” in the target record, it is checked whether or not the evaluation value is equal to or lower than a predetermined threshold value (“0” in the present embodiment) (s1540). If it is not below (s1540: NO), the process returns to s1510. On the other hand, if it is below the threshold (s1540: YES), this evaluation value update process is terminated.
(2) Actions and Effects According to the content providing system as described above, the following actions and effects can be obtained in addition to the actions and effects obtained from the same configuration as in the first embodiment.

For example, the evaluation value for the content acquired based on the keyword group indicates the inner product of the interest vector corresponding to the content and the interest vector of the current interest table, but is only calculated upon acquisition. For this reason (s1120 in FIG. 15), if the interest vector of the current interest table changes with the user's interest over time, the content is highly similar to the latest interest vector reflecting the user's interest. Therefore, content that is not necessarily estimated to be of high user interest may be preferentially provided. However, in this embodiment, since the evaluation value can be updated by the evaluation value update process so that the evaluation value becomes lower (s1510 to s1540 in FIG. 20), each content is reflected with the user's interest along with the elapsed time. The similarity to the latest interest vector can be lowered. For this reason, it is possible to prevent content with less interest from the user being preferentially provided over time.
3. The embodiment of the present invention has been described above. However, the present invention is not limited to the above-described embodiment, and various forms can be taken as long as they belong to the technical scope of the present invention. Nor.

  For example, in the above-described embodiment, the information collection / analysis apparatus 70 is configured to acquire content from an external server. However, for example, in a case where the information collection / analysis apparatus 70 itself has various contents, a configuration may be adopted in which a search according to a keyword is retrieved from these contents. Further, the information collection / analysis apparatus 70 may be provided with a content search function in the server, and the information collection / analysis apparatus 70 itself may be configured to search for content by this function. The device 70 itself can search and acquire content based on the keyword.

  Further, for example, the personal computer and the information collecting apparatus may be integrated, and in this case, the input unit 41 and the display unit 43 in FIG. 1 correspond to the client device of the present invention. Further, the display unit 77 for outputting content may not be integrated with the information collecting device, but may be another device (for example, the display unit 43 of the personal computer 30 in FIG. (It may be a device).

In the second embodiment, in the evaluation value changing process, the “evaluation value” in the target record is updated according to the elapsed time from the time indicated by “source generation time information” in the target record to the current time. What was comprised was illustrated. However, the time used when specifying the elapsed time to be referred to when updating this “evaluation value” may be other than “source generation time information”. The acquired time (or stored in the storage unit 73), the date and time, or a specific time thereafter can be employed.
In the above embodiment, the content searched based on the keyword group is stored in the storage area corresponding to the latest interest information (s1150 to s1210 in FIG. 15), and then sorted in descending order of evaluation value. (S1230 in the figure) exemplifies a configuration in which such a content with a high evaluation value is output with the highest priority. However, in addition to rearranging in the descending order of evaluation values, for example, rearranging in the order stored most recently or the order acquired most recently (that is, the order in which the time indicated by the source generation time information is close), The content may be preferentially output in such an order, or the content may be preferentially output in an order rearranged randomly. Further, in the above embodiment, when outputting the content searched based on the keyword group, the content to be read out first is read out from among the content stored in each data queue constituting the storage area. However, it may be configured so that only the priority is output.
4). Correspondence with the Present Invention In the embodiment described above, the personal computer 30 and the telephone device 50 are client devices in the present invention, the information collection device 10 is a content providing device in the present invention, and the information search server 20 is It is a content search part in this invention. A system including these devices is the content providing system of the present invention.

The text data collected at s150 in FIG. 3 and the voice data acquired at s240 in FIG. 4 are input information and output information in the present invention.
Further, s310 and s320 in FIG. 5 are information acquisition means in the present invention.

Further, s1120 in FIG. 15 and s1250 in FIG. 17 are interest estimation means in the present invention.
Further, s1220 in FIG. 15 and s1420 in FIG. 19 are content deletion means in the present invention.

  Further, s1140 in FIG. 15 is content acquisition means in the present invention, s1150 to s1220 in FIG. 15 are content storage means in the present invention, s1130 in FIG. 15 is similarity evaluation means in the present invention, and s1230 in FIG. These are order rearranging means in the present invention.

Further, s1340 in FIG. 18 is content output means in the present invention.
Further, s1510 to s1540 in FIG. 20 are evaluation result changing means in the present invention.

Block diagram showing the configuration of the content provision system Figure showing the content provision screen Flow chart showing PC monitoring process Flow chart showing telephone monitoring process Flow chart showing source information registration processing Diagram showing data structure of source control table Flow chart showing keyword extraction processing The figure which shows the data structure of a search information management table Flow chart showing interest vector calculation processing Interest estimation table Flow chart showing search control processing The figure which shows the data structure of the retrieval program management table Flow chart showing search processing Search result management table Flowchart showing search result evaluation processing in the first embodiment Diagram showing data structure of current interest table and memory interest table Flow chart showing current interest table update processing Flow chart showing output control processing Flowchart showing search result evaluation processing in the second embodiment Flow chart showing evaluation value change processing Flow chart showing elapsed time coefficient table

DESCRIPTION OF SYMBOLS 1 ... Content provision system, 10 ... Web server, 20 ... Information retrieval server, 21 ... Control part, 23 ... Memory | storage part, 25 ... Communication part, 30 ... Personal computer, 31 ... Control part, 33 ... Memory | storage part, 35 ... Monitoring Unit 37, communication unit 40, user interface 41 input unit 43 display unit 50 personal computer 50 digital telephone device 51 control unit 53 storage unit 55 monitoring unit 57 communication , 60 ... user interface, 61 ... sound input part, 63 ... sound output part, 70 ... information collection and analysis device, 71 ... control part, 73 ... storage part, 73 ... control part, 75 ... communication part, 77 ... display part 79 ... Sound output section.

Claims (9)

Information acquisition means for acquiring at least one or both of text data including a keyword input to the user interface and text data including a keyword output from the user interface from a client device having a user interface;
Each text data Ru is acquired by the information acquisition means, a keyword extracted from the text data, the are regarded as keywords that reflects the interests of the user of the user interface, interest in the information corresponding to the keyword user Interest estimation means for estimating as interest information indicating
Each time text data is acquired by the information acquisition means, content acquisition means for acquiring content according to the acquired text data ;
The storage corresponding to the interest information estimated by the interest estimation means based on the text data used in the acquisition of the content acquired by the content acquisition means in the storage unit associated with each of the plurality of interest information Content storage means for storing in the unit;
Content output means for providing content to the user of the client device by causing the content storage means to output the content stored in the storage section to the output section,
The content output unit selects a storage unit corresponding to the latest interest information estimated by the interest estimation unit from the plurality of storage units when outputting the content to the output unit, and the selected storage unit A content providing apparatus that causes the output unit to output content stored in the device. The content acquisition unit, to the content retrieval unit having a function to search for content based on the interest information estimated by the interest estimating means and acquired text data by the information acquisition unit, which is the acquired text The content providing apparatus according to claim 1, wherein a content search based on data is requested, and the searched content is acquired in response to the request. For each of the contents acquired by the content acquisition means, it comprises similarity evaluation means for evaluating how similar the content is to the latest interest information estimated by the interest estimation means,
The content storage means stores the content acquired by the content acquisition means in a storage unit together with an evaluation result by the similarity evaluation means for the content,
The content output unit selects a storage unit corresponding to the latest interest information estimated by the interest estimation unit and then resembles the content stored in the storage unit according to the evaluation result by the similarity evaluation unit. The content providing apparatus according to claim 1, wherein content having a higher degree of content is preferentially output to the output unit. In the case where the storage unit is configured to sequentially store contents in a queue method,
Each time content is stored in the storage unit by the content storage unit, the content that is stored in the storage unit is read more quickly as the content is more similar in the evaluation result by the similarity evaluation unit. Order rearrangement means for rearranging so that
The content output unit outputs a content stored in the storage unit according to the order rearranged by the order rearrangement unit after selecting the storage unit corresponding to the latest interest information estimated by the interest estimation unit. The content providing apparatus according to claim 3, wherein the content providing apparatus outputs the content to a unit. The content storage means is configured to store the content acquired by the content acquisition means in a storage unit together with time information indicating a specific time after the content is acquired,
Furthermore, according to the specific time indicated by the time information stored in the storage unit together with the content by the content storage unit, an evaluation result changing unit is provided for changing the evaluation result stored in the storage unit together with the content. The content providing apparatus according to claim 3, wherein: The evaluation result changing unit is configured to store the content with the content by the content storage unit so that the degree of similarity decreases as the elapsed time from the time indicated by the time information stored in the storage unit together with the content increases to the current time. The content providing apparatus according to claim 5, wherein the evaluation result stored in is changed. Content deletion means for deleting, from the storage unit, content stored in the storage unit by the content storage unit, wherein the content indicated by the time information stored together with the content is earlier than a predetermined threshold value. The content providing device according to claim 5, wherein the content providing device is provided.   A content providing system comprising the content providing apparatus according to claim 1.   A program for causing a computer system to execute various processing procedures for causing the content providing apparatus according to any one of claims 1 to 7 to function as all means.

Images