Program recommendation apparatus and program recommendation method

The present invention relates to a program recommendation device and a program recommendation method, and is suitable for application to, for example, a program recommendation device that recommends a program required by a viewer from a large number of television programs distributed via a broadcasting satellite. It is a thing.

In a satellite broadcasting system in which television programs are distributed to viewers via broadcasting satellites, television signals are digitized and a huge number of programs are simultaneously distributed. In such a system, the number of programs selected by the viewer is significantly increased.

Further, in a system in which various information is provided from the host side to the computer terminal via a telephone line or a dedicated line, the user on the terminal side selects necessary information from a huge amount of information and requests it from the host side. It will be.

<p> When a viewer or a user wants to select information or the like using such a television program or a computer, a desired program or information must be searched from a huge amount of programs or information. In this case, the viewer or the user selects a program to be selected, a word related to the information to be selected, or the like as a keyword, and searches for the desired program or information.</p><p> However, in the method in which the viewer or user directly inputs the keyword into the search system, the viewer or user constantly learns the knowledge about the latest keyword or information classification method as the program or information is updated. It is necessary to keep updating, and it is difficult to easily select the desired keyword.</p><p> In addition, there is a method in which a keyword selected by a viewer or a user in the past and a keyword are stored as a selection history and used as a keyword at a later search. However, in this method, when the search system is used for the first time, there is no history information, and the viewer or user directly selects and inputs the keyword. In this case as well, as in the above case, It is difficult to easily select keywords. Thus, the search operation of the viewer or the user becomes complicated, and there is a problem that it is difficult to easily select the required program or information.</p><p> The present invention has been made in consideration of the above points, and an object of the present invention is to propose a program recommendation device and a program recommendation method capable of presenting a program according to a user's taste to a user by a simple operation.</p>

<p> In order to solve such a problem, in the present invention, a program according to the user's preference is searched using the preference information and the program guide information input by the user, and the list of the searched programs is used as a recommended program. Changed to display on the display means.</p><p> As a result, the user can obtain a list of programs according to the preference as a recommended program simply by inputting the preference information.</p>

<p> According to the present invention, a program according to a user's preference is searched using the preference information and program guide information input by the user, and a list of the searched programs is displayed on a display means as a recommended program. By doing so, the user can obtain a list of programs according to the preference as a recommended program simply by inputting the preference information, and thus presents the program according to the user's preference to the user with a simple operation. It is possible to realize a program recommendation device to obtain.</p>

An embodiment of the present invention will be described in detail below with reference to the drawings.

(1) Overall configuration of satellite broadcast reception system In Fig. 1, 1 shows the satellite broadcast reception system as a whole, and the broadcast signal received by the parabolic antenna 3 is demodulated and decoded by the reception decoding device (IRD: Integrated Receiver / Decoder) 2. It is compressed and decrypted. The video / audio signal SV1 obtained as a result is transmitted by the following VHS VCR (Video Cassette Recorder) 6.

The VCR6 monitors and displays the video / audio signal SV1 by recording it on a video tape loaded inside, or by transmitting the video / audio signal SV1 as it is from the output line to the monitor device 4.

When the viewer operates the remote commander 5, the command corresponding to the operation is converted into an infrared signal IR and sent to the reception decoding device 2. The reception decoding device 2 switches channels based on the command, registers / reads user data, sends a control signal CONT to each device (VCR6, VCR7, DVD8, MD9) connected to the reception decoding device 2, etc. Perform various actions. The control signal CONT is sent to VCR6 via the control line. When VCR6 is designated as a control target by this control signal CONT, VCR6 is controlled by the control signal CONT. On the other hand, the devices (8 mm VCR7, digital video disc player (DVD: Digital Video Disc) 8) and mini disc player (MD: Mini) sequentially connected to the VCR 6 via the control line are controlled by the control signal CONT. When either Disc) 9 or monitoring device 4) is specified, VCR6 sends the control signal CONT to the following 8mm VCR7.

When the control signal CONT is input, the VCR7 determines the device designated by the control signal CONT. When this determination result is VCR7, VCR7 executes the operation specified by the control signal CONT. When this instruction is, for example, an instruction to play an 8 mm videotape loaded in the VCR7, the VCR7 displays the playback video signal SV3 by sending the reproduced video signal SV3 to the monitoring device 4 by playing the videotape. Further, when the instruction by the control signal CONT is an instruction to record the broadcast signal (video / audio signal SV1) received and decoded by the reception decoding device 2 on the VCR7, the VCR7 is a VHS system from the reception decoding device 2. The video / audio signal SV1 input via the VCR6 and the monitor device 4 of the above is recorded. On the other hand, when the control target of the control signal CONT is not VCR7, VCR7 sends the control signal CONT to the subsequent DVD8 as it is.

When the control signal CONT is input to the DVD8, the device specified by the control signal CONT is determined. When the determination result is DVD8, the DVD8 executes the operation specified by the control signal CONT. If this instruction is, for example, an instruction to play video and / or audio from a disk loaded in DVD 8, the DVD 8 sends the video / audio signal SV4 to the monitor device 4 by playing the disk. indicate. On the other hand, when the control target of the control signal CONT is not DVD8, DVD8 sends the control signal CONT to the subsequent MD9 as it is.

When the MD9 inputs the control signal CONT, the MD9 determines the device designated by the control signal CONT. When this determination result is MD9, MD9 executes the operation specified by the control signal CONT. When this instruction is, for example, an instruction to reproduce the disk loaded in the MD9, the MD9 audibly displays the audio signal SV5 by transmitting the audio signal SV5 to the monitoring device 4 by reproducing the disk. Further, when the instruction by the control signal CONT is an instruction to record the broadcast signal (video / audio signal SV1) received and decoded by the reception decoding device 2 on the MD9, the MD9 is subjected to the VHS method from the reception decoding device 2. The audio signal of the video / audio signal SV1 input via the VCR6 and the monitor device 4 of the above is recorded. On the other hand, when the control target of the control signal CONT is not MD9, MD9 sends the control signal CONT to the subsequent monitoring device 4 as it is. At this time, the monitoring device 4 executes the operation specified by the control signal CONT.

(2) Configuration of reception decoding device In Fig. 2, the reception decoding device 2 supplies the RF signal output by the LNB (Low Noise Block downconverter) 3A of the parabolic antenna 3 to the tuner 21 of the front end 20 and demodulates it. .. The output of the tuner 21 is supplied to the QPSK demodulation circuit 22 and demodulated by QPSK. The output of the QPSK demodulation circuit 22 is supplied to the error correction circuit 23, and errors are detected, corrected, and corrected as necessary.

The CAM (Conditional Access Module) 33, which consists of an IC card consisting of a CPU, ROM, and RAM, stores the keys necessary for decrypting the code together with the decryption program. Since the signal transmitted via the broadcasting satellite is encrypted, a key and a decryption process are required to break this code. Therefore, this key is read from the CAM 33 via the card reader interface 32 and supplied to the demultiplexer 24. The demultiplexer 24 uses this key to decrypt the encrypted signal.

The demultiplexer 24 receives the signal output from the error correction circuit 23 of the front end 20, supplies the decoded video signal to the MPEG video decoder 25, and supplies the decoded audio signal to the MPEG audio decoder 26.

The MPEG video decoder 25 stores the input digital video signal in the DRAM 25A and executes decoding processing of the video signal compressed by the MPEG method. The decoded video signal is supplied to the NTSC encoder 27 and converted into an NTSC luminance signal (Y), a chroma signal (C) and a composite signal (V). The luminance signal and chroma signal are output as S-video signals via the buffer amplifiers 28Y and 28C, respectively. Further, the composite signal is output via the buffer amplifier 28V.

The MPEG audio decoder 26 stores the digital audio signal supplied from the demultiplexer 24 in the DRAM 26A, and executes decoding processing of the audio signal compressed by the MPEG method. The decoded audio signal is digital-to-analog converted by the D / A converter 30, the left channel audio signal is output via the buffer amplifier 31L, and the right channel audio signal is output via the buffer amplifier 31R.

The RF modulator 41 converts the composite signal output by the NTSC encoder 27 and the audio signal output by the D / A converter 30 into an RF signal and outputs the signal. In addition, this RF modulator 41 passes through the NTSC RF signal input from another device and outputs it to the other device as it is. In the case of this embodiment, these video signals and audio signals are supplied to the VCR 6 via the AV line.

The CPU 29 executes various processes according to the program stored in the ROM 37. Further, the CPU 29 controls the AV device control signal transmission / reception unit 2A, outputs a predetermined control signal to another device via the control line, and also receives a control signal from the other device.

A predetermined command can be directly input to the CPU 29 by operating the operation button switch on the front panel 40. Further, when the operation key of the remote commander 5 is operated, an infrared signal is output by the IR transmitting unit of the remote commander 5, the infrared signal is received by the IR receiving unit 39, and the received light receiving result is supplied to the CPU 29. Therefore, a predetermined command can be input to the CPU 29 by operating the remote commander 5.

In addition, the CPU 29 takes in, for example, EPG (Electronic Program Guide) information other than the video signal and audio signal output by the demultiplexer 24, creates EPG data from it, supplies it to the SRAM (Static Random Access Memory) 36, and stores it. Let me. The EPG information includes information about the program of each broadcast channel from the current time to several tens of hours later (for example, the channel of the program, the broadcast time, the title, the channel, the program commentary, etc.). Since this EPG information is transmitted frequently, the SRAM 36 always holds the latest EPG information.

The CPU 29 can transfer the data stored inside the SRAM 36 to an external device via the modem 34 via a communication means. Incidentally, as a method of transferring the data of the SRAM 36 to an external device (floppy disk, card-shaped recording medium, etc.), in addition to communication using a modem, an output line dedicated to the data may be provided.

In addition, EEPROM (Electrically Erasable Programable Read Only Memory) 38 contains data that you want to keep even after the power is turned off (for example, the reception history of the rewritable channel 21 for the past 4 weeks and the database (11A, 11B, 11C) described later. ) Data) etc. are stored. In addition, the CPU 29 compares the time information output by the calender timer 35 with the time stamp separated and output by the demultiplexer 24 from the received signal so that the decoding process can be performed at the correct timing according to the comparison result. , Controls the MPEG video decoder 25 and the MPEG audio decoder 26.

Further, the CPU 29 controls the MPEG video decoder 25 when it wants to generate predetermined OSD (On-Screen Display) data. The MPEG video decoder 25 generates predetermined OSD data in response to this control, writes it to the DRAM 25A, and further reads and outputs it. As a result, predetermined characters, figures, and the like can be output to the monitoring device 4 and displayed.

Here, when the operation key of the program guide is selected on the remote commander 5 or the front panel 40, the CPU 29 as the display control means controls the MPEG video decoder 25, and the monitor device 4 displays the broadcast program selection screen. The user can select and specify a desired program by moving the cursor to a desired program position on this screen and clicking the remote commander 5. At this time, a list of programs suitable for the user is displayed from among a large number of programs, using the program range corresponding to the user's preference generated in advance in the keyword generation function block provided in the reception decoding device 2 as a keyword. Will be done.

In this way, Fig. 3 shows a keyword generation function block used when searching for a program desired by the user based on the EPG information. That is, in FIG. 3, the user interface processing unit 12 corresponds to the remote commander 5, the IR receiving unit 39, and the front panel 40 (FIG. 2) of the receiving decoding device 2, the answer analysis processing unit 13, and the situation-specific preference keyword generation unit. 14 and the specific situation preference keyword generation processing unit 15 and the package title search processing unit 16 as a search means correspond to CPU29 (Fig. 2), and correspond to the preference group cluster dictionary 11A, the preference group keyword group database 11B, and the package title database. 11C corresponds to EEPROM 38 as a storage means.

(3) Keyword generation by the reception decoding device FIG. 3 shows the functional block of the part related to the keyword generation of the reception decoding device 2 described above with respect to FIG. 2, and the user interface processing unit 12 operates the remote commander 5 by the user. Therefore, the dialogue screen for keyword generation is displayed on the display screen 4A of the monitoring device 4 (Fig. 1). The user inputs the user profile for keyword generation while specifying the answer to each question using the cursor on this dialogue screen.

The first input item is "advancement", "employment", and "marriage", which are the growth stages of the individual user, taking into account the relationship with the user's family and society, as the life stage in which the user is currently placed. , "Child-rearing", "retirement", etc., and in this case, the dialogue screen as shown in FIG. 4 is displayed on the display screen 4A of the monitor device 4.

Secondly, as an input item, there is an item for inputting age / gender. In this case, the display screen 4A displays a dialogue screen as shown in FIG. In addition, as an input item, there is a third item related to the user's preference tendency. In this case, the display screen 4A displays a dialogue screen for specifying a plurality of preference tendencies as shown in FIG.

In addition, as input items, fourthly, as the user's life scene / selection site environment phase, life scenes such as "breakfast", "lunch", "dinner", "weekday relaxation" and "holiday relaxation" are entered. There is an item to enter. In this case, the user inputs his / her actual time range (this is called environmental numerical value / area data) corresponding to each life scene for each day of the week on the dialogue screen as shown in FIG. As a result, for example, as a life scene of "breakfast", data such as "Monday 7:00 to 7:30", "Saturday 7:30 to 8:00", ..., etc. are obtained. ..

When the user's answer is input in this way, the user interface processing unit 12 sends the answer to the answer analysis processing unit 13. The answer analysis processing unit 13 uses a time zone identifier (situation identifier) representing each life scene input by the user as a different identifier, and each time zone identifier obtained based on the user's response. The user's habit status conversion data is obtained by pairing the corresponding user-specific daytime time range data (environmental numerical area data) for each life scene.

Figure 8 shows an example of this habit situation conversion data. That is, Fig. 8 (A) is a data array in which the day of the week and the time correspond to the time zone identifier (situation identifier) representing "breakfast time". In this case, breakfast is in the same time range from Monday to Friday. Therefore, these data are represented by the product of the data representing the range of the day of the week (Monday to Friday) and the data representing the range of the time (7:00 to 7:30), and at a time different from these weekdays. The Saturday when breakfast is taken is represented by the product of the day of the week data (Saturday) and the data representing the time range (7:30 to 8:00). The daytime time range data (area data of the environmental numerical value) is obtained by the sum of each data represented by the product of the day range data and the time range data, and the day time range data and the time are obtained. The habit situation conversion data is obtained by the combination with the band identifier (situation identifier).

In addition, Fig. 8 (B) shows the habit situation conversion data by combining the time zone identifier (situation identifier) indicating "when relaxing on a holiday" and the day of the week time range data, and both Saturday and Sunday are from 8:00 to 11:30. It means that the life scene of "when relaxing on a holiday" corresponds to the minute. In this way, the time zone identifier as the situation identifier set according to the characteristics of the user is a name or number that distinguishes typical life scenes that affect when selecting a program, and is a user's preference tendency. It independently influences program selection and is a factor that should be selected from time to time. Incidentally, as the situation identifier, in addition to the time zone identifier, for example, there is a companion situation identifier set according to the person who shares the situation with the user, and the situation sharing partner includes a friend, a lover, or the like. This companion status identifier is used to generate keywords when selecting music for music programs and music software.

Thus, the habit-situation conversion data representing the user habit obtained by the combination of the time zone identifier and the area data of the environmental numerical value is temporarily stored in the EEPROM 38 (Fig. 2) as a storage means.

In addition, the response analysis processing unit 13 obtains a preference attribute attribution data array as data representing the user's preference tendency that changes from time to time or depending on the case. In this case, the item of the preference tendency input by the user to the user interface processing unit 12 is used. This item was input by the above-mentioned dialogue screen for Fig. 6, and according to this answer, the user's attitude values toward watching TV include "knowledge orientation", "active orientation", and "entertainment orientation". Multiple preference attributes that affect program selection, such as "knowledge orientation", can be obtained. By the way, when generating keywords when selecting music, items for obtaining directional tendencies such as "specific genre-oriented", "song-oriented", "wide-range-oriented", and "fashion-oriented" are given as questions to the user. ..

Therefore, the response analysis processing unit 13 first obtains the preference attribute of the user based on the response regarding the user's preference tendency input to the user interface processing unit 12. That is, the response analysis processing unit 13 has each of the preference attributes "knowledge orientation", "active orientation", "entertainment orientation", "jizukuri orientation", etc. obtained from the user's response. The degree of orientation is set as a value on the attribute classification axis representing each orientation. As a result, in the preference attribute classification space formed by each attribute classification axis, the coordinates determined by each orientation become the user's preference attribute vector, and one point on the preference space determined by this preference attribute vector. Is a preference attribute point representing the preference tendency of this user.

By the way, FIG. 9 shows an example of a preference attribute classification space formed by three attribute classification axes, an age hierarchy axis (Z axis), an active orientation axis (X axis), and a knowledge orientation axis (Y axis). The preference attribute point P is obtained from the age, active orientation, and knowledge orientation obtained by inputting.

Here, when a large number of users are used as a population and a plurality of preference attribute points are plotted in one preference attribute classification space, a dense group (hereinafter referred to as a cluster) may appear in several places. Each cluster corresponds to a group of users with similar preference attributes, and there will be a finite number of clusters in the preference attribute classification space that are not necessarily exclusive. As an example of the cluster, the knowledge attitude cluster determined by the knowledge-oriented axis, the active-oriented axis, and the age group axis shown in Fig. 9 corresponds to the knowledge attitude cluster CL1 corresponding to the entertainment group and the knowledge desire group. There are knowledge attitude cluster CL2 and knowledge attitude cluster CL3 corresponding to fashion pursuit groups. Further, the cluster may be formed in a projective subspace using only a part of the preference attribute classification axis. In this case, for example, an age hierarchy cluster is formed in the projective space using the age hierarchy axis. By the way, in the preference attribute classification space when selecting music, clusters corresponding to mood euphoria and screaming divergence are formed.

The name or number used to distinguish these clusters is called a cluster identifier, and the center of each cluster is called a cluster representative point. Here, the preference attribute point P corresponding to one user generally does not always match the representative point of the cluster. In addition, one user is considered to have some preference attributes of the proximity cluster. Therefore, the degree to which one user's preference attribute points belong to each adjacent cluster is represented by a numerical array, and this numerical array is used as the preference attribute attribution data array of the user.

Here, when the data of the user's preference attribute point P is determined, the degree of attribution to each cluster is determined from the preference attribute point P, the representative point of the cluster, the spread, and the shape. Of these, the cluster representative point and the way the cluster spreads do not depend on the user's preference attribute point P at all, and are unique to each cluster. Therefore, it is possible to determine in advance a method of calculating the degree of attribution (preference attribute attribution) to each cluster from the cluster representative points and the spread aspect for each cluster.

The calculation method of the degree of attribution to the cluster (degree of attribution of preference attribute) will be described below. When the preference attribute point P of one user is determined, in order to obtain the degree of attribution to a certain cluster (the degree of attribution of the preference attribute), first, the error vector between the preference attribute point P and the cluster representative point is obtained. Next, the value is calculated using a function that decreases monotonically as this error vector increases (that is, a function determined by how the cluster spreads).

The function used to obtain the degree of attribution of the preference attribute is normal with the standard deviation of the spread (variation of the spread), etc., when the spread aspect of the cluster is isotropic regardless of the preference attribute classification axis direction. The inverse value of the number obtained by adding 1.0 to the square of the length of the converted error vector (representing the distance to the cluster representative point) is defined as the preference attribute attribution. In this case, the site block distance, the maximum absolute value component, or the eclipse distance can also be used as the length of the error vector.

If there is a difference for each preference attribute classification axis as a way of spreading the cluster, the reciprocal of the standard deviation value for each preference attribute classification axis is used as the load coefficient of that axis instead of the above-mentioned isotropic distance. The reciprocal of the number obtained by adding approximately 1.0 to the square of the separately loaded (square) norm (that is, when the cluster is regarded as a square) is defined as the preference attribute attribution.

In addition, as a way of spreading the cluster, when it spreads in a direction inclined with respect to the preference attribute classification axis, a quadratic form ellipsoidal norm using a coefficient obtained from the covariance coefficient etc. (that is, the cluster is an ellipsoid) The quotient obtained by dividing another constant by the number obtained by adding a certain number to (when regarded as) is defined as the preference attribute attribution degree.

By the way, when the spread of the cluster is complicated and a general function is required, a function using a convex polyhedron norm using the maximum value of a finite number of linear expressions instead of the above-mentioned city block distance, or It is possible to use a function that uses a Nieuro or a lucup table.

The various functions set as the attribution calculation method in this way are stored in advance in the cluster dictionary 11A (Fig. 3), and are designated by the attribution calculation method designation data stored in advance in the cluster dictionary 11A. This attribution calculation method specification data to be read is data that specifies the function to be used for each cluster when calculating the attribution of the cluster and what parameters should be executed for that function, and is represented by a function pointer. It is a combination of the calculated calculation function identifier and the calculation parameters such as the cluster representative point and the degree of cluster spread. Calculation parameters are represented by data arrays, pointers to data structures, and the like.

The calculation of the preference attribute attribution data array using the functions and parameters set in this way is performed in the cluster dictionary 11A when the user preference attribute point P is determined by the analysis of the user response in the response analysis processing unit 13. It is executed in the response analysis processing unit 13 while referring to the attribution calculation method designation data corresponding to each stored cluster.

That is, for the attribution value to one cluster, the attribution calculation method designation data of the cluster is extracted from the cluster dictionary 11A, and the parameters that are a part of the calculation method designation data and the preference attribute point data that is the answer analysis result. Is used as an argument to read the function specified by the attribution calculation method specification data and execute the function. The function value obtained as a result of this function execution is the cluster attribution value. By repeating this sequentially for all clusters, the obtained attribution value is sequentially assigned to the array elements, thereby obtaining the preference attribute attribution data array of the user.

Incidentally, the cluster dictionary 11A can be provided in the EEPROM 38 (FIG. 2), read from a predetermined recording medium, or downloaded from a communication line and stored in the EEPROM 38 for use. In this case, the type of cluster and the calculation method can be updated, and further, a new calculation method can be realized by updating the cluster dictionary together with the addition of registration of a new function program.

By the way, FIG. 10 shows an example of the preference attribute attribution data array, and in the attribution sequence to each age group, each number arranged is each age group (for example, teens, 20s, 30s, .. ....) represents the degree of attribution to each preference group, and in the degree of attribution array to each preference group, each number arranged is for each preference group (knowledge desire group, fashion pursuit group, ...). Represents the degree of attribution. In this case, limiting each of the arranged numbers to "0" or "1" means whether the user is completely in each cluster or completely irrelevant.

Thus, when the response analysis processing unit 13 obtains the user's preference attribute attribution data array, the attribution data array is sent to the situation-specific preference keyword generation unit 14 (FIG. 3) together with the above-mentioned habit-situation conversion data. The preference keyword generation unit 14 for each situation sets the preference attribute clusters (identifiers) corresponding to the top several high attribution degree data sequences of the preference attribute attribution degree data array as the strong attribution clusters of the user.

The situation-specific preference keyword generation unit 14 extracts the keywords corresponding to the strongly belonging cluster from the preference group situation-specific keyword group database 11B. In this preference group database 11B for each situation, the keywords included in the preference titles (favorite program Jiangru) of people with various tendencies in various situations are classified and stored.

That is, in general, a typical user belonging to each preference cluster prefers a title (program janle) with a certain tendency under a typical situation. Therefore, in the keyword group database 11B for each situation according to preference group, frequently-used keyword groups are prepared in advance for each situation classification and each preference cluster in the favorite title (program janru) and the article of the introduction summary. By the way, when a keyword is generated when a TV program is selected, a program genre name is prepared as a frequently-used keyword. Each keyword prepared in the keyword group database 11B according to the preference group and the situation is given a preference degree.

The keyword group database 11B for each situation according to preference group can extract a group of keywords and a pair of preference degree divided for each situation classification when at least one preference attribute cluster is specified. As an actual configuration, a database and a search server (subroutine, thread, process), etc. are used.

Therefore, the situation-specific preference keyword generation unit 14 sequentially selects the situation-specific keyword group corresponding to the strong attribution cluster of the user for each typical situation represented by each situation classification identifier. Extract from database 11B. In general, there are a plurality of strong attribution clusters, and a plurality of preference keyword groups can be obtained even for a single situation. Merge (combine) this into one set for each situation. As a method of this merging, first, a set of keywords is obtained by performing a collective merger of keyword groups of each cluster. Next, the preference for each keyword is the preference attached to the cluster keyword and the attribution of this user to the cluster when the keyword comes from the preference keyword group of only one cluster. Calculate from degree. The condition of the function of this calculation is a function that has a weak monotonic increase in both the original preference and the attribution.

For example, there are a method using the product of the degree of preference and the degree of attribution, a method using the arithmetic mean, a method using the minimum value, and the like. Furthermore, a monotonically increasing function using the Ruthkuuptable method may be used.

Next, assuming that the same keyword is included in the preference keyword group of a plurality of clusters, first, the preference degree is obtained as a single cluster by the above method, and the sum or the maximum value thereof is set as the synthesis preference degree. By repeating these processes for each situation classification, a favorite keyword group (program genre name group) for each situation regarding a specific user can be obtained.

The keyword group thus obtained is stored in the EEPROM 38 (Fig. 2). In addition, the strongly attributed cluster data for each user is also stored in EEPROM38, and when the preference keyword database for each preference cluster status (keyword group database for each preference wave status in Fig. 3) is updated, the updated database is displayed. By re-searching, the preference keyword group for each situation for each user can be updated by synthesizing by the above method.

By the way, FIG. 11 is an example of the situation-specific preference keywords of a specific user generated by the situation-specific preference keyword generation unit 14, and the program genre name group in each situation (breakfast, rest, ...) is shown. Generated for each situation.

Thus, the situation-specific preference keyword group (FIG. 11) generated by the situation-specific preference keyword generation unit 14 is sent to the subsequent specific situation preference keyword generation processing unit 15. Here, the specific situation represents a situation at a specific time point, and is typically represented by one situation identifier, but is a composite of situations represented by a plurality of situation identifiers according to each situation. Therefore, as an expression of a specific situation, an array of numerical values representing the degree (situation attribution degree) close to each of the typical situations represented by the situation identifier is used. This situation attribution degree array is called a situation attribution degree data array.

This status attribution data array can be automatically generated by the system, or can be input to the system on the spot by the user via the input means (user interface processing unit 12). For example, CPU29 (Fig. 2) automatically generates the degree of time zone attribution that determines the vicinity of the time zone boundary based on the time. On the other hand, as for the situation of friends in the field, the degree of attribution to the situation is determined as a result of input by the user using the dialogue screen to specify the situation.

The specific situation preference keyword generation processing unit 15 receives the preference keyword group of the specific user corresponding to the specific situation represented by the situation attribution data array thus obtained from the situation-specific preference keyword generation unit 14. Based on the situation-specific preference keyword group corresponding to a typical situation, it is obtained by load synthesis using the degree of situation attribution. In the load composition calculation for obtaining the preference degree to be associated with each keyword, the product-sum composition of the situation attribution degree and the preference degree of the typical situation can be simply used. The keyword set with a preference degree obtained in this way becomes a specific situation preference keyword group of the specific user. Incidentally, as a method of load synthesis calculation for obtaining the degree of preference, a function having monotonous increasing property for all variables may be selected and used for synthesis.

Thus, as shown in FIG. 12, the specific situation keyword group generated in the specific situation preference keyword generation process 15 is sent to the package title search processing unit 16 as the subsequent search means, and the specific situation keyword group causes the specific situation keyword group. The corresponding title is searched from the package title database 11C. In the case of this embodiment, the package title database 11C stores the EPG data transmitted by satellite broadcasting, and the EPG data specified by the program genre generated as a specific situation keyword group is searched. Will be done. Based on this EPG data, a plurality of characters representing the searched program are displayed as recommended programs on the display screen 4A of the monitor device 4, and the user selects the program by designating one of the characters. Can be done.

By the way, the contents of the package title database 11C are updated every time new EPG data is imported, and the latest data is always retained.

(4) Operation and effect of the embodiment In the above configuration, the user uses the dialogue screen displayed on the monitor screen to display the life stage, age / gender, user preference tendency, and user's preference. Life scene / selection When inputting the on-site environment and daily matters, the keyword generation block section (Fig. 3) of the receiving / decoding device 2 contains the habit status conversion data related to the user's habit status and the user's preference attribute. The preference attribute attribution degree data related to is generated, and thereby, a search keyword group reflecting the preference tendency of the user under a specific situation in a specific field is generated.

Therefore, even if the user does not have specialized knowledge about searching such as constantly updated keywords and the latest knowledge about the Janle classification method, he / she can ask simple daily questions about the user's habits and preferences. After answering once, programs that match the user's unique situation and taste are continuously searched.

In addition, the latest keywords can be handled immediately by simply rewriting the keyword database for each preference group and situation stored in a storage means such as EPPROM38. As a result, the user can always respond to the update of the keyword without remembering the latest keyword.

Thus, according to the above configuration, the burden on the user's search can be significantly reduced.

(5) Other Examples In the above-described embodiment, the case where the life stage, age / gender, preference tendency, and life scene are input as user input items has been described, but the present invention is not limited to this. It may be limited to one of these items, or other items may be added.

Further, in the above-described embodiment, the case where the keyword generation block for information retrieval is provided inside the reception decoding device 2 for receiving satellite broadcasting is described, but the present invention is not limited to this, and the keyword generation device is separate. It may be provided in.

Further, in the above-described embodiment, the case where the present invention is applied to a device for searching a program of digital satellite broadcasting has been described, but the present invention is not limited to this, for example, searching for a huge amount of information by an internet, a compact disk, or the like. It can be widely applied to a keyword generator of various information retrieval devices such as a package information retrieval.

<figref num="1">It is a block diagram which shows the satellite broadcasting reception system using the keyword generator by this invention.</figref><figref num="2">It is a block diagram which shows the structure of the reception decoding apparatus including the keyword generation apparatus.</figref><figref num="3">It is a block diagram which shows the keyword generation function block of the reception decoding apparatus.</figref><figref num="4">It is a schematic diagram which shows the dialogue screen for a user.</figref><figref num="5">It is a schematic diagram which shows the dialogue screen for a user.</figref><figref num="6">It is a schematic diagram which shows the dialogue screen for a user.</figref><figref num="7">It is a schematic diagram which shows the dialogue screen for a user.</figref><figref num="8">It is a schematic diagram which shows an example of a habit situation conversion data.</figref><figref num="9">It is a schematic diagram which shows the simplified example of a preference attribute space.</figref><figref num="10">It is a schematic diagram which shows the preference attribute attribution degree data arrangement example.</figref><figref num="11">It is a schematic diagram which shows the example of a user's situation preference keyword.</figref><figref num="12">It is a schematic diagram which shows a specific situation keyword group.</figref>

Code description

2 ...... Receive decoding device, 4 ...... Monitor device, 5 ...... Remote commander, 11A ...... Preference cluster dictionary, 11B ...... Keyword group database by preference group situation, 11C ...... Package title database, 12 ...... User interface processing department, 13 ...... Answer analysis processing department, 14 ... ... Situation-specific preference keyword generation unit, 15 ...... Specific situation preference keyword generation processing unit, 16 ...... Package title search processing unit.