Information processing device, information processing method, and computer program

Abstract

An information processing device (10) comprising: an acquisition means (50) that acquires a plurality of elements included in sequential data; a calculation means (100) that calculates, on the basis of at least two elements among the plurality of elements, a classification index indicating the likelihood of a class to which the sequential data belongs; and a selection means that, on the basis of the classification index, selects and outputs K (K is an integer that is no more than N and no less than 1) classes having a high potential that the sequential data belongs thereto from among N (N is an integer) classes that are classification candidates for the sequential data. By using such an information processing device, it is possible to appropriately select a plurality of classes having a high potential that the sequential data belongs thereto.

Description

[Technical field]

[0001] The present disclosure relates to the technical fields of an information processing device, an information processing method, and a computer program for processing information regarding class classification, for example. [Background technology]

[0002] As this type of device, there is known a device that performs class classification of input data. For example, Patent Document 1 discloses a technique for outputting a plurality of class candidates based on a classification score.

[0003] As another related technique, for example, Patent Document 2 discloses a technique for pruning candidates for authentication targets whose posterior probability is lower than a threshold value. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2013-025496 A [Patent Document 2] JP 2009-289253 A Summary of the Invention [Problem to be solved by the invention]

[0005] This disclosure aims to improve upon the related art mentioned above. [Means for solving the problem]

[0006] One aspect of the information processing device disclosed herein includes an acquisition means for acquiring a plurality of elements included in sequence data, a calculation means for calculating a classification index indicating the likelihood of a class to which the sequence data belongs based on at least two of the plurality of elements, and a selection means for selecting and outputting K classes (where K is a natural number equal to or less than N and equal to 1) to which the sequence data is likely to belong based on the classification index from N classes (where N is a natural number) that are classification candidates for the sequence data.

[0007] One aspect of the information processing method disclosed herein includes obtaining multiple elements included in sequence data, calculating a classification index indicating the likelihood of a class to which the sequence data belongs based on at least two of the multiple elements, and selecting, based on the classification index, K classes (where K is a natural number equal to or less than N and equal to 1 or greater) to which the sequence data is likely to belong from N classes (where N is a natural number) that are candidates for classification of the sequence data, and outputting the K classes.

[0008] One aspect of the computer program of the present disclosure operates a computer to obtain multiple elements included in sequence data, calculate a classification index indicating the likelihood of a class to which the sequence data belongs based on at least two of the multiple elements, and select and output K classes (where K is a natural number equal to or less than N and equal to 1) to which the sequence data is likely to belong based on the classification index from N classes (where N is a natural number) that are candidates for classification of the sequence data. [Brief description of the drawings]

[0009] [Figure 1] 1 is a block diagram showing a hardware configuration of an information processing device according to a first embodiment. [Diagram 2] 1 is a block diagram showing a functional configuration of an information processing device according to a first embodiment. [Diagram 3] 4 is a flowchart showing a flow of operations of the information processing device according to the first embodiment. [Figure 4]10 is a flowchart showing the flow of operations of the information processing device according to the second embodiment. [Diagram 5] 13 is a flowchart showing the flow of a classification index calculation process in an information processing device according to a third embodiment. [Figure 6] 13 is a flowchart showing the flow of operations of the information processing device according to the fourth embodiment. [Figure 7] 13 is a graph showing an example of class selection in the information processing device according to the fourth embodiment. [Figure 8] 13 is a flowchart showing a flow of operations of an information processing device according to a modified example of the fourth embodiment. [Figure 9] 13 is a flowchart showing the flow of operations of the information processing device according to the fifth embodiment. [Figure 10] 13 is a graph showing an example of class selection in an information processing device according to the fifth embodiment. [Figure 11] FIG. 13 is a block diagram showing a functional configuration of an information processing device according to a sixth embodiment. [Figure 12] 13 is a flowchart showing the flow of a classification index calculation process in an information processing device according to a sixth embodiment. [Figure 13] 13 is a flowchart showing the flow of operations of the information processing device according to the seventh embodiment. [Figure 14] FIG. 23 is a diagram showing an example of display of output classes by the information processing device according to the seventh embodiment. [Figure 15] 20 is a flowchart showing the flow of operations of the information processing device according to the eighth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, embodiments of an information processing device, an information processing method, and a computer program will be described with reference to the drawings.

[0011] First Embodiment An information processing device according to a first embodiment will be described with reference to FIGS. 1 to 3. FIG.

[0012] (Hardware configuration) First, the hardware configuration of the information processing device according to the first embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing the hardware configuration of the information processing device according to the first embodiment.

[0013] 1, an information processing device 10 according to the first embodiment includes a processor 11, a RAM (Random Access Memory) 12, a ROM (Read Only Memory) 13, and a storage device 14. The information processing device 10 may further include an input device 15 and an output device 16. The processor 11, the RAM 12, the ROM 13, the storage device 14, the input device 15, and the output device 16 are connected via a data bus 17.

[0014] The processor 11 reads a computer program. For example, the processor 11 is configured to read a computer program stored in at least one of the RAM 12, the ROM 13, and the storage device 14. Alternatively, the processor 11 may read a computer program stored in a computer-readable storage medium using a storage medium reading device (not shown). The processor 11 may acquire (i.e., read) a computer program from a device (not shown) disposed outside the information processing device 10 via a network interface. The processor 11 controls the RAM 12, the storage device 14, the input device 15, and the output device 16 by executing the computer program that the processor 11 reads. In particular, in this embodiment, when the processor 11 executes the computer program that the processor 11 reads, a functional block for selecting and outputting a class to which sequence data may belong is realized in the processor 11. Note that the processor 11 may be one of a central processing unit (CPU), a graphics processing unit (GPU), a field-programmable gate array (FPGA), a demand-side platform (DSP), and an application specific integrated circuit (ASIC), or may be a plurality of devices in parallel.

[0015] The RAM 12 temporarily stores a computer program executed by the processor 11. The RAM 12 temporarily stores data that is temporarily used by the processor 11 when the processor 11 is executing the computer program. The RAM 12 may be, for example, a D-RAM (Dynamic RAM).

[0016] The ROM 13 stores computer programs executed by the processor 11. The ROM 13 may also store other fixed data. The ROM 13 may be, for example, a programmable ROM (P-ROM).

[0017] The storage device 14 stores data that is to be stored long-term by the information processing device 10. The storage device 14 may operate as a temporary storage device for the processor 11. The storage device 14 may include, for example, at least one of a hard disk device, a magneto-optical disk device, an SSD (Solid State Drive), and a disk array device.

[0018] The input device 15 is a device that receives an input instruction from a user of the information processing device 10. The input device 15 may include, for example, at least one of a keyboard, a mouse, and a touch panel. The input device 15 may also include a terminal owned by the user (for example, a smartphone, a tablet terminal, etc.).

[0019] The output device 16 is a device that outputs information related to the information processing device 10 to the outside. For example, the output device 16 may be a display device (e.g., a display) that can display information related to the information processing device 10. The display device here may be a television monitor, a personal computer monitor, a smartphone monitor, a tablet terminal monitor, digital signage, or the like. The output device 16 may also be a speaker that outputs information related to the information processing device 10 by voice.

[0020] (Functional configuration) Next, the functional configuration of the information processing device 10 according to the first embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram showing the functional configuration of the information processing device according to the first embodiment.

[0021] 2, the information processing device 10 according to the first embodiment includes, as processing blocks for realizing its functions, an acquisition unit 50, an index calculation unit 100, and a class selection unit 200. Each of the acquisition unit 50, the index calculation unit 100, and the class selection unit 200 may be realized by the above-mentioned processor 11 (see FIG. 1).

[0022] The acquisition unit 50 is configured to be able to acquire multiple elements included in the sequence data. The acquisition unit 50 may acquire data directly from an arbitrary data acquisition device (e.g., a camera, a microphone, etc.), or may read data acquired in advance by a data acquisition device and stored in a storage or the like. The elements of the sequence data acquired by the acquisition unit 50 are configured to be output to the index calculation unit 100. Note that sequence data is data including multiple elements arranged in a predetermined order, and one example is time-series data. More specific examples of sequence data include, but are not limited to, video data and audio data.

[0023] The index calculation unit 100 is configured to be able to calculate a classification index based on at least two elements among the multiple elements acquired by the acquisition unit 50. The index calculation unit 100 calculates a classification index based on, for example, two elements successively acquired by the acquisition unit 50. Note that the "classification index" here is an index indicating the likelihood of a class to which sequence data belongs. Specific examples of the classification index and a specific calculation method will be described in detail in other embodiments described later.

[0024] The class selection unit 200 is configured to select and output K classes (where K is a natural number equal to or less than N) to which the sequence data is likely to belong from among N classes (where N is a natural number) that are classification candidates for the sequence data, based on the classification index calculated by the index calculation unit 100. The N classes of the classification candidates may be preset. Alternatively, the N classes of the classification candidates may be appropriately set by a user, or may be appropriately set based on the type of sequence data to be handled. Meanwhile, the K classes are a plurality of classes selected by the class selection unit 200 from among the N classes of the classification candidates. However, when a plurality of classes cannot be selected, the class selection unit 200 may exceptionally select only one class (i.e., K≧2 in principle, but K=1 may be an exception). More specific processing contents of the class selection unit 200 (i.e., a method of selecting K classes) will be described in detail in another embodiment described later.

[0025] (Operation flow) Next, the flow of operations of the information processing device 10 according to the first embodiment will be described with reference to Fig. 3. Fig. 3 is a flowchart showing the flow of operations of the information processing device according to the first embodiment.

[0026] 3, when the operation of the information processing device 10 according to the first embodiment is started, the acquisition unit 50 first acquires one element from sequence data (step S11). The acquisition unit 50 outputs the acquired element of the sequence data to the index calculation unit 100. Then, the index calculation unit 100 calculates a classification index based on the two or more acquired elements (step S12).

[0027] Next, the class selection unit 200 determines whether or not it is possible to select K classes (i.e., classes to which the sequence data is likely to belong) based on the calculated classification index (step S13). The criteria for this determination will be described in detail in another embodiment described later. If the class selection unit 200 determines that it is not possible to select K classes based on the calculated classification index (step S13: NO), the information processing device 10 starts the process again from step S11. That is, the information processing device 10 obtains the next element of the sequence data and calculates a new classification index.

[0028] On the other hand, when it is determined that K classes can be selected based on the calculated classification index (step S13: YES), the class selection unit 200 selects K classes based on the calculated classification index (step S14). That is, K classes are selected from N classes that are classification candidates. Then, the class selection unit 200 outputs information about the selected K classes (step S15). The class selection unit 200 may assign labels corresponding to the selected K classes to the sequence data.

[0029] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the first embodiment will be described.

[0030] As described in FIG. 1 to FIG. 3, in the information processing device 10 according to the first embodiment, a classification index is calculated based on at least two elements of the sequence data. Then, K classes are selected from N classes that are classification candidates of the sequence data based on the classification index. In this way, it is possible to appropriately determine which class the sequence data may belong to, taking into account at least two elements contained in the sequence data. In addition, since the K classes to be selected are multiple classes, the class determination can be performed more flexibly compared to the case where only one class is selected. For example, this has a beneficial effect in the case where the correct answer is multiple classes or the user himself wants to determine one correct answer from multiple classes selected as candidates. Note that specific application examples of the information processing device 10 according to this embodiment will be described in detail in other embodiments described later.

[0031] <Second embodiment> The information processing device 10 according to the second embodiment will be described with reference to Fig. 4. The second embodiment differs from the first embodiment only in some operations, and may be similar to the first embodiment in terms of, for example, the device configuration (see Figs. 1 and 2). Therefore, hereinafter, the parts that differ from the first embodiment will be described in detail, and the explanation of the other overlapping parts will be omitted as appropriate.

[0032] (Operation flow) First, the flow of operations of the information processing device 10 according to the second embodiment will be described with reference to Fig. 4. Fig. 4 is a flowchart showing the flow of operations of the information processing device according to the second embodiment. In Fig. 4, the same processes as those shown in Fig. 3 are denoted by the same reference numerals.

[0033] 4, when the operation of the information processing device 10 according to the first embodiment is started, the acquiring unit 50 first acquires one element from the sequence data (step S11). The acquiring unit 50 outputs the acquired element of the sequence data to the index calculation unit 100.

[0034] Particularly in the second embodiment, the index calculation unit 100 calculates a classification index for each of the N classes that are classification candidates (step S21). For example, if the N classes that are classification candidates include class A, class B, class C, class D, ..., the index calculation unit 100 calculates N classification indexes corresponding to each class, such as the classification index of class A, the classification index of class B, the classification index of class C, the classification index of class D, ....

[0035] Next, the class selection unit 200 determines whether or not it is possible to select K classes (i.e., classes to which the sequence data is likely to belong) based on the calculated N classification indices (step S13). If the class selection unit 200 determines that it is not possible to select K classes based on the calculated N classification indices (step S13: NO), the information processing device 10 starts the process again from step S11. That is, the information processing device 10 obtains the next element of the sequence data and calculates a new classification index.

[0036] On the other hand, if it is determined that K classes can be selected based on the calculated N classification indices (step S13: YES), the class selection unit 200 selects K classes based on the calculated N classification indices (step S22), and outputs information about the selected K classes (step S15).

[0037] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the second embodiment will be described.

[0038] As described in FIG. 4, in the information processing device 10 according to the second embodiment, a classification index is calculated for each of N classes (i.e., N classification indexes are calculated). By calculating N classification indexes in this manner, it is possible to appropriately select K classes to which sequence data is likely to belong from among N classes that are classification candidates. For example, whether or not the sequence data is likely to belong to class A included in the N classes (in other words, whether or not class A should be selected as one of the K classes) can be determined based on the classification index calculated for class A. Similarly, whether or not the sequence data is likely to belong to class B included in the N classes (in other words, whether or not class B should be selected as one of the K classes) can be determined based on the classification index calculated for class B.

[0039] <Third embodiment> An information processing device 10 according to the third embodiment will be described with reference to Fig. 5. The third embodiment shows a specific method for calculating the N classification indexes in the second embodiment described above, and the device configuration and the overall operation flow may be similar to those of the second embodiment. Therefore, in the following, only the parts that differ from the second embodiment will be described in detail, and the description of the other overlapping parts will be omitted as appropriate.

[0040] (Operation flow) First, the flow of the classification index calculation process in the information processing device 10 according to the third embodiment (i.e., the process of step S21 in Fig. 4) will be described with reference to Fig. 5. Fig. 5 is a flowchart showing the flow of the classification index calculation process in the information processing device according to the third embodiment.

[0041] 5, when the classification index calculation process is started in the information processing device 10 according to the third embodiment, the index calculation unit 10 first calculates a first likelihood for one class included in the N classes that are classification candidates (step S211). The first likelihood here is a value indicating the likelihood that sequence data belongs to a certain class, and can be expressed as p(X|y=1) (y=1 means a state of belonging to a certain class).

[0042] Next, the index calculation unit 100 calculates a second likelihood for one class for which the first likelihood has been calculated (step S212). The second likelihood here is a value indicating the likelihood that the sequence data does not belong to a certain class, and can be expressed as p(X|y=0) (y=0 means that the sequence data does not belong to a certain class).

[0043] Next, the index calculation unit 100 calculates a likelihood ratio, which is a ratio between the first likelihood and the second likelihood already calculated (step S213). This likelihood ratio can be expressed as p(X|y=1) / p(X|y=0), for example. Note that the calculation method of the likelihood ratio so far is only an example, and the likelihood ratio may be calculated using other methods. The likelihood ratio will be described in detail again in other embodiments described later.

[0044] Next, the index calculation unit 100 judges whether the likelihood ratios have been calculated for all the N classes that are classification candidates (step S214). If the likelihood ratios have not been calculated for all the N classes that are classification candidates (step S214: NO), the index calculation unit 100 repeats the same process (i.e., the process from step S211 to step S213) for other classes for which the likelihood ratios have not yet been calculated. On the other hand, if the likelihood ratios have been calculated for all the N classes that are classification candidates (step S214: YES), the index calculation unit 100 outputs the calculated likelihood ratios (specifically, the N likelihood ratios calculated for each of the N classes) as classification indices (step S215).

[0045] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the third embodiment will be described.

[0046] 5, the information processing device 10 according to the third embodiment calculates a likelihood ratio, which is a ratio between a first likelihood indicating the likelihood that sequence data belongs to a certain class and a second likelihood indicating the likelihood that sequence data does not belong to a certain class, as a classification index. By using such a likelihood ratio as a classification index, it becomes possible to appropriately select K classes to which sequence data is likely to belong from N classes that are classification candidates.

[0047] <Fourth embodiment> An information processing device 10 according to the fourth embodiment will be described with reference to Figures 6 and 7. Note that the fourth embodiment differs from the first to third embodiments in only some of its operations, and other operations and device configurations may be similar to those of the first to third embodiments. Therefore, hereinafter, only the parts that differ from the first to third embodiments will be described in detail, and explanations of other overlapping parts will be omitted as appropriate.

[0048] (Operation flow) First, the flow of operations of the information processing device 10 according to the fourth embodiment will be described with reference to Fig. 6. Fig. 6 is a flowchart showing the flow of operations of the information processing device according to the fourth embodiment. In Fig. 6, the same processes as those shown in Figs. 3 and 4 are denoted by the same reference numerals.

[0049] 6, when the operation of the information processing device 10 according to the fourth embodiment starts, the acquiring unit 50 first acquires one element from sequence data (step S11). The acquiring unit 50 outputs the acquired element of the sequence data to the index calculating unit 100. Then, the index calculating unit 100 calculates a classification index for each of N classes that are classification candidates (step S21).

[0050] Next, the class selection unit 200 determines whether K classification indices out of the calculated N classification indices have reached a predetermined threshold (step S31). The "predetermined threshold" here is a threshold for determining that the sequence data is highly likely to belong to the corresponding class based on the classification indices, and may be set in advance, for example, by prior simulation. Alternatively, the predetermined threshold may be appropriately set by a user. The predetermined threshold may be set as one threshold common to all of the N classification indices, or may be set as N thresholds corresponding to the N classification indices, respectively.

[0051] If K classification indices among the calculated N classification indices have not reached the predetermined threshold (step S31: NO), the information processing device 10 starts the process again from step S11. That is, the information processing device 10 acquires the next element of the sequence data and calculates new classification indices. On the other hand, if K classification indices among the calculated N classification indices have reached the predetermined threshold (step S31: YES), the class selection unit 200 selects classes corresponding to the K classification indices that have reached the predetermined threshold (i.e., K classes) (step S32). Then, the class selection unit 200 outputs information related to the selected K classes (step S15).

[0052] (Specific operation example) Next, a specific operation example (a specific selection example of K classes) of the information processing device 10 according to the fourth embodiment will be described with reference to Fig. 7. Fig. 7 is a graph showing a class selection example in the information processing device according to the fourth embodiment. Note that in the following example, a case will be described in which the classification index is calculated as the likelihood ratio described in the third embodiment.

[0053] As shown in Fig. 7, N classification indices (7 in the figure) corresponding to N classes that are classification candidates vary from their initial values ​​over time. Specifically, the N classification indices vary gradually each time the next element in the sequence data is acquired and calculated. Here, if it is determined that the sequence data is highly likely to belong to that classification index, the classification index gradually increases (i.e., varies upward in the figure). On the other hand, if it is determined that the sequence data is low likely to belong to that classification index, the classification index gradually decreases (i.e., varies downward in the figure).

[0054] If the classification index continues to increase, it will eventually reach the predetermined threshold. The information processing device 10 according to the fourth embodiment selects K classes in the order of the earliest timing at which the classification index reaches the predetermined threshold. For example, in the example shown in FIG. 7, K=4 is set. Therefore, when four of the seven classification indexes reach the predetermined threshold, the classes corresponding to the four classification indexes are selected as the K classes to which the sequence data is likely to belong. On the other hand, the three classes corresponding to the other three classification indexes that do not reach the predetermined threshold are not selected as the K classes to which the sequence data is likely to belong. Note that the classes not selected as the K classes are not usually output, but information about the classes not selected as the K classes can be utilized. For example, the class farthest from the predetermined threshold among the classes not selected as the K classes may be output as reference information (i.e., presented to the user) or may be excluded when processing similar sequence data. Alternatively, among the classes not selected as the K classes, the class that reaches a second threshold value set on the negative side (for example, see the rejection threshold described below) earliest may be output as reference information or may be excluded when processing similar sequence data.

[0055] When K classes are selected as described above, the calculation process of the classification index may be performed until the K classification indexes reach a predetermined threshold. In other words, when the K classification indexes reach the predetermined threshold, the calculation process of the classification index may be stopped. Furthermore, for the classification index that has reached the predetermined threshold, the calculation process may be stopped at that point.

[0056] A rejection threshold may be set separately from the above-mentioned predetermined threshold. The rejection threshold is a threshold for determining that the sequence data is unlikely to belong to the corresponding class based on the classification index, and may be set in advance by, for example, prior simulation. Alternatively, the rejection threshold may be set appropriately by a user. The rejection threshold may be set as one threshold common to all of the N classification indexes, or may be set as N thresholds corresponding to each of the N classification indexes. The rejection threshold is typically set to a value sufficiently lower than the initial value.

[0057] If the classification index continues to decrease, it will eventually reach the rejection threshold. For the classification index that has reached the rejection threshold, the calculation process may be stopped at that point. In other words, it may be determined that the classification index is unlikely to increase and reach the predetermined threshold, and the calculation process thereafter may be stopped.

[0058] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the fourth embodiment will be described.

[0059] As described in FIG. 6 and FIG. 7, in the information processing device 10 according to the fourth embodiment, K classes to which sequence data is likely to belong are selected in order of the earliest time when the classification index reaches a predetermined threshold. In this way, it is possible to easily and accurately select the K classes. Furthermore, if the calculation process of the classification index is stopped when the K classification indexes reach the predetermined threshold, the processing load can be appropriately reduced. Furthermore, for classification indexes that have reached the predetermined threshold or the rejection threshold, the calculation process of the classification index is stopped when each threshold is reached, so that the subsequent processing load can be appropriately reduced.

[0060] <Modification of the Fourth Embodiment> Next, an information processing device 10 according to a modification of the fourth embodiment will be described with reference to Fig. 8. Note that the information processing device 10 according to the modification of the fourth embodiment differs from the fourth embodiment described above only in some operations, and other operations and device configurations may be similar to those of the fourth embodiment. Therefore, hereinafter, the parts that differ from the fourth embodiment already described will be described in detail, and the explanation of other overlapping parts will be omitted as appropriate.

[0061] (Operation flow) First, the flow of operations of the information processing device 10 according to the modified example of the fourth embodiment will be described with reference to Fig. 8. Fig. 8 is a flowchart showing the flow of operations of the information processing device according to the modified example of the fourth embodiment. In Fig. 8, the same processes as those shown in Fig. 6 are denoted by the same reference numerals.

[0062] 8, when the operation of the information processing device 10 according to the modification of the fourth embodiment is started, the acquiring unit 50 first acquires one element from the sequence data (step S11). The acquiring unit 50 outputs the acquired element of the sequence data to the index calculating unit 100. Then, the index calculating unit 100 calculates a classification index for each of N classes that are classification candidates (step S21).

[0063] Next, the class selection unit 200 determines whether K classification indices among the calculated N classification indices have reached a predetermined threshold value (step S31). If K classification indices among the calculated N classification indices have reached a predetermined threshold value (step S31: YES), the class selection unit 200 selects classes corresponding to the K classification indices that have reached the predetermined threshold value (i.e., K classes) (step S32). Then, the class selection unit 200 outputs information related to the selected K classes (step S15).

[0064] On the other hand, if K classification indices out of the calculated N classification indices have not reached the predetermined threshold (step S31: NO), the class selection unit 200 judges whether or not a time limit has elapsed since the series of processes shown in the flowchart was started (step S41). The "time limit" here is a value set as the longest time for executing the series of processes, and an optimal value is determined, for example, by a prior simulation. The time limit may be appropriately changed by an operation by a user or the like.

[0065] If it is determined that the time limit has not elapsed (step S41: NO), the information processing device 10 determines that the series of processes may be continued, and starts the process again from step S11. That is, the information processing device 10 obtains the next element of the sequence data and calculates a new classification index. On the other hand, if it is determined that the time limit has elapsed (step S41: YES), the information processing device 10 determines that the series of processes should not be continued any further, stops the calculation process of the classification index, and proceeds to a process of outputting K classes.

[0066] Specifically, the class selection unit 200 sets the number of classification indices that have reached the predetermined threshold when the time limit has elapsed as K (step S42). For example, if the original value of K is 4 and there are only two classification indices that have reached the predetermined threshold when the time limit has elapsed, the class selection unit 200 changes the value of K from 4 to 2. Similarly, if there is only one classification indices that have reached the predetermined threshold when the time limit has elapsed, the class selection unit 200 changes the value of K from 4 to 1. That is, the class selection unit 200 here executes a process of forcibly reducing the value of K to a value that can be selected at the present time.

[0067] After that, the class selection unit 200 selects classes corresponding to the K classification indices (changed values) that have reached the predetermined threshold (step S32), and outputs information about the selected K classes (step S15).

[0068] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the modified example of the fourth embodiment will be described.

[0069] As described in FIG. 8, in the information processing device 10 according to the modified example of the fourth embodiment, when the time limit has elapsed, even if the classification index that has reached the predetermined threshold at that time is less than the original K, K is forcibly made smaller, and classes corresponding to the classification index that has already reached the predetermined threshold are selected as the K classes. In this way, it is possible to avoid the inconvenience that K classes are not output even after a long time has elapsed (i.e., classes cannot be selected because the classification index that has reached the predetermined threshold is less than K). Note that the fourth embodiment is an example of a configuration in which the value of K is variably set. If the value of K is made variable in this way, class selection can be performed more flexibly compared to the case in which the value of K is fixed.

[0070] <Fifth embodiment> An information processing device 10 according to the fifth embodiment will be described with reference to Figures 9 and 10. The fifth embodiment differs from the first to fourth embodiments in only some of its operations, and other operations and device configurations may be similar to those of the first to fourth embodiments. Therefore, hereinafter, only the parts that differ from the first to fourth embodiments will be described in detail, and explanations of other overlapping parts will be omitted as appropriate.

[0071] (Operation flow) First, the flow of operations of the information processing device 10 according to the fifth embodiment will be described with reference to Fig. 9. Fig. 9 is a flowchart showing the flow of operations of the information processing device according to the fifth embodiment. In Fig. 9, the same processes as those shown in Fig. 4 are denoted by the same reference numerals.

[0072] 9, when the operation of the information processing device 10 according to the fifth embodiment is started, first, the acquiring unit 50 acquires one element from the sequence data (step S11). The acquiring unit 50 outputs the acquired element of the sequence data to the index calculating unit 100. Then, the index calculating unit 100 calculates a classification index for each of N classes that are classification candidates (step S21).

[0073] Next, the class selection unit 200 judges whether a predetermined time has elapsed since the series of processes shown in the flowchart was started (step S51). The "predetermined time" here is a value set as the longest time for executing the series of processes, and an optimal value is determined, for example, by a prior simulation. The predetermined time may be appropriately changed by an operation of a user, etc.

[0074] If it is determined that the predetermined time has not elapsed (step S51: NO), the information processing device 10 restarts the process from step S11. That is, the information processing device 10 acquires the next element of the sequence data and calculates a new classification index. On the other hand, if it is determined that the predetermined time has elapsed (step S51: YES), the class selection unit 200 selects K classes in descending order of the classification index at the time when the predetermined time has elapsed (step S52). Then, the class selection unit 200 outputs information related to the selected K classes (step S15).

[0075] (Specific operation example) Next, a specific operation example (a specific selection example of K classes) of the information processing device 10 according to the fifth embodiment will be described with reference to Fig. 10. Fig. 10 is a graph showing a class selection example in the information processing device according to the fifth embodiment. Note that in the following example, a case will be described in which the classification index is calculated as the likelihood ratio described in the third embodiment.

[0076] As shown in Fig. 10, N classification indices (7 in the figure) corresponding to N classes that are classification candidates vary from their initial values ​​over time. Specifically, the N classification indices vary gradually each time the next element of the sequence data is acquired and calculated. Here, if it is determined that the sequence data is highly likely to belong to that classification index, the classification index gradually increases (i.e., varies upward in the figure). On the other hand, if it is determined that the sequence data is low likely to belong to that classification index, the classification index gradually decreases (i.e., varies downward in the figure).

[0077] In the information processing device 10 according to the fifth embodiment, when a predetermined time has elapsed since the start of processing, the calculation process of the classification index is stopped at that time, and K classes are selected in descending order of the classification index. That is, the classes corresponding to the K largest classification indexes are selected as the K classes to which the sequence data is likely to belong. In the example shown in FIG. 10, K=4 is set. Therefore, of the seven classification indexes, the classes corresponding to the first four largest classification indexes are selected as the K classes to which the sequence data is likely to belong. On the other hand, the three classes corresponding to the other three classification indexes that do not reach the predetermined threshold are not selected as the K classes to which the sequence data is likely to belong. Note that a predetermined threshold may be set as in the fourth embodiment already described (see, for example, FIG. 7), and the K classes may be selected when the classification index reaches the predetermined threshold. Specifically, the calculation process of the classification index may be stopped when any classification index reaches the predetermined threshold, and the K classes may be selected in descending order of the classification index at that time.

[0078] If the K largest classification indices include a classification indice that should not be selected if possible (for example, a classification indice smaller than the initial value), the class corresponding to that classification indice may not be selected. In this case, the K classes that are output will be smaller than the value of K that was originally set. This configuration is an example of a configuration in which the value of K is variably set.

[0079] Furthermore, in order to prevent K from becoming small as described above, a plurality of predetermined times may be set in stages. For example, if the K largest classification indices include a classification indice that should not be selected when a first predetermined time has elapsed, the processing may be continued without selecting K classes until a second predetermined period longer than the first predetermined time has elapsed. Then, when the second predetermined time has elapsed after the processing has continued, the K largest classification indices may be selected.

[0080] For example, assume that K=4 is set. In this case, if no classification indices that should not be selected are included when the first specified time has elapsed, the class selection unit 200 selects and outputs the four classification indices as they are. On the other hand, if a classification indice that should not be selected is included when the first specified time has elapsed, the class selection unit 200 does not output a classification indice at that time and continues processing. Then, when a second specified time has elapsed, the class selection unit 200 may select the K largest classification indices in descending order.

[0081] In addition, when multiple predetermined times are set, the user may be asked whether or not to continue processing when the first predetermined time has elapsed. In this case, if the user answers that the processing is OK, the processing continues even after the first predetermined time has elapsed until the next predetermined time has elapsed depending on the situation. On the other hand, if the user answers that the processing should not be continued, K classes may be selected and output at that point.

[0082] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the fifth embodiment will be described.

[0083] 9 and 10, in the information processing device 10 according to the fifth embodiment, K classes are selected in descending order of the classification index after a predetermined time has elapsed. In this way, it is possible to reliably output K classes within a preset time. As a result, it is possible to avoid the inconvenience of not outputting K classes even after a long time has elapsed.

[0084] Sixth embodiment An information processing device 10 according to the sixth embodiment will be described with reference to Fig. 11 and Fig. 12. The sixth embodiment differs in some configurations and operations from the first to fifth embodiments described above, and may be similar to the first to fifth embodiments in terms of, for example, the hardware configuration and the overall flow of operations. Therefore, hereinafter, the parts that differ from the first to fifth embodiments described above will be described in detail, and the explanation of other overlapping parts will be omitted as appropriate.

[0085] (Functional configuration) First, the functional configuration of the information processing device 10 according to the sixth embodiment will be described with reference to Fig. 11. Fig. 11 is a block diagram showing the functional configuration of the information processing device according to the sixth embodiment. In Fig. 11, the same reference numerals are used to designate the same elements as those shown in Fig. 2.

[0086] As shown in FIG. 11, the information processing device 10 according to the sixth embodiment includes an acquisition unit 50, an index calculation unit 100, and a class selection unit 200 as processing blocks for realizing the functions thereof. In particular, in the sixth embodiment, the acquisition unit 50 is configured to be able to sequentially acquire elements included in the sequence data. The index calculation unit 100 includes a first calculation unit 110 and a second calculation unit 120. The first calculation unit 110 includes a likelihood ratio calculation unit 111 and a first storage unit 112. The second calculation unit 120 includes an integrated likelihood ratio calculation unit 121 and a second storage unit 122. Each of the likelihood ratio calculation unit 111 and the integrated likelihood ratio calculation unit 121 may be realized by the above-mentioned processor 11 (see FIG. 1). Each of the first storage unit 112 and the second storage unit 122 may be realized by the above-mentioned storage device 14 (see FIG. 1).

[0087] The likelihood ratio calculation unit 111 is configured to be able to calculate a likelihood ratio for each of the elements acquired by the acquisition unit 50. The likelihood ratio is a value indicating the likelihood that each element belongs to a certain class among a plurality of classes. The likelihood ratio calculation unit 111 calculates a likelihood ratio based on the acquired element and past data stored in the first storage unit 112. The information stored in the first storage unit 112 is configured to be readable by the likelihood ratio calculation unit 111. When the first storage unit 112 stores past likelihood ratios, the likelihood ratio calculation unit 111 may read the stored past likelihood ratios and calculate the likelihood ratio for the acquired element. On the other hand, when the first storage unit 112 stores elements acquired in the past, the likelihood ratio calculation unit 111 may calculate the past likelihood ratios from the stored past elements and calculate the likelihood ratio for the acquired element. A specific example of the likelihood ratio will be described below.

[0088] The N elements that make up the sequence data are denoted as x1,…,x N Let R and F be the multiple classes. If these two classes are respectively a state of being one class and a state of not being one class, then it can be considered similar to the case described in the third embodiment. For convenience of explanation, the explanation will proceed by referring to them as class R and class F.

[0089] Here, element x i The probability that belongs to class R is calculated without considering past data, and the result is p(R|x i ) and the element x i The probability that belongs to class F is calculated without considering past data and is called p(F|x i ) In this case, these likelihood ratios are expressed by the following equation (1).

[0090]

number

[0091] The likelihood ratio of the above formula (1) is the element x i The probability that element x belongs to class R i This shows the ratio of likelihood to the probability that the class F belongs to the class. For example, if the likelihood ratio is greater than 1, then p(R|x i )>p(R|x i ), so element x i should be classified into class R rather than class F. In this way, the likelihood ratio in equation (1) functions as an index showing whether the input element belongs to class R or class F.

[0092] Furthermore, the likelihood ratio calculation unit 111 can perform calculations taking into consideration a plurality of elements (i.e., the continuity between the input element and past data) as described above. In this case, for example, two elements x i ,x i-1 The likelihood ratio calculated taking into account the above is expressed as the following equation (2).

[0093]

number

[0094] The integrated likelihood ratio calculation unit 121 is configured to be able to calculate an integrated likelihood ratio, which is a classification index. The integrated likelihood ratio is a value indicating the likelihood that sequence data belongs to a certain class among a plurality of classes. The integrated likelihood ratio calculation unit 121 calculates an integrated likelihood ratio using the likelihood ratio calculated by the likelihood ratio calculation unit 111 and past integrated likelihood ratios stored in the second storage unit 122. The information stored in the second storage unit 122 (i.e., past integrated likelihood ratios) is configured to be readable by the integrated likelihood ratio calculation unit 121. A specific example of the integrated likelihood ratio will be described below. As with the above description of the likelihood ratio, a case of two-class classification in which the number of classes is two will be described.

[0095] When N elements are input at the time of calculating the integrated likelihood ratio, the N elements are x1,…,x N Here, the probability that the entire sequence data belongs to class R is expressed as p(x1,…,x N |R). Let p(x1,…,x N In this case, these likelihood ratios are expressed by the following equation (3). Equation (3) is called the integrated likelihood ratio.

[0096]

number

[0097] When it is assumed that each element of the sequence data is independent, the integrated likelihood ratio can be calculated by decomposing it into terms for each element as shown in the following equation (4).

[0098]

number

[0099] In the above formula (4), the logarithm of the likelihood ratio is used to decompose each element into a sum in order to simplify the calculation, but this is not essential. Note that, below, the term likelihood ratio or integrated likelihood ratio may also be used for such logarithmic likelihood ratios. Also, although the notation of the base of the logarithm is omitted, the value of the base is arbitrary.

[0100] However, as described above, in this embodiment, the likelihood ratio and the integrated likelihood ratio are calculated taking into account two or more elements, and therefore the assumption that each element is independent is often not valid. Therefore, it is not possible to decompose into terms for each element as in formula (4), and the integrated likelihood ratio is calculated using a different formula depending on the number of elements whose relationships are considered.

[0101] For example, when two elements, a certain element and the element immediately before it, are taken into consideration, the integrated likelihood ratio can be calculated using the following formula (5).

[0102]

number

[0103] (Indicator calculation process flow) Next, a flow of index calculation processing (i.e., processing for calculating likelihood ratios and integrated likelihood ratios) by the information processing device 10 according to the sixth embodiment will be described with reference to Fig. 12. Fig. 12 is a flowchart showing the flow of classification index calculation processing in the information processing device according to the sixth embodiment.

[0104] 12, in the index calculation process according to the sixth embodiment, first, the likelihood ratio calculation unit 111 of the first calculation unit 110 reads out past data from the first storage unit 112 (step S201). The past data may be, for example, a processing result by the likelihood ratio calculation unit 111 of an element acquired immediately before an element currently acquired by the acquisition unit 50 (in other words, a likelihood ratio calculated for the immediately previous element). Alternatively, the past data may be the element itself acquired immediately before the element acquired in the acquisition.

[0105] Next, the likelihood ratio calculation unit 111 calculates a new likelihood ratio (i.e., a likelihood ratio for the element currently acquired by the acquisition unit 50) based on the element acquired by the acquisition unit 50 and the past data read out from the first storage unit 112 (step S202). The likelihood ratio calculation unit 111 outputs the calculated likelihood ratio to the second calculation unit 120. The likelihood ratio calculation unit 111 may store the calculated likelihood ratio in the first storage unit 112.

[0106] Next, the integrated likelihood ratio calculation unit 121 in the second calculation unit 120 reads out a past integrated likelihood ratio from the second storage unit 122 (step S203). The past integrated likelihood ratio may be, for example, a processing result by the integrated likelihood ratio calculation unit 121 for an element acquired immediately before the element currently acquired by the acquisition unit 50 (in other words, an integrated likelihood ratio calculated for the immediately previous element).

[0107] Next, the integrated likelihood ratio calculation unit 121 calculates a new integrated likelihood ratio (i.e., an integrated likelihood ratio for the element currently acquired by the acquisition unit 50) based on the likelihood ratio calculated by the likelihood ratio calculation unit 111 and the past integrated likelihood ratio read out from the second storage unit 122 (step S204). The integrated likelihood ratio calculation unit 121 outputs the calculated integrated likelihood ratio to the class selection unit 200. The integrated likelihood ratio calculation unit 121 may store the calculated integrated likelihood ratio in the second storage unit 122.

[0108] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the sixth embodiment will be described.

[0109] As described in Fig. 11 and Fig. 12, in the information processing device 10 according to the sixth embodiment, a likelihood ratio and an integrated likelihood ratio are calculated from elements included in the sequence data. Then, the calculated integrated likelihood ratio is used as a classification index. As already described, the integrated likelihood ratio is a value that takes into account the continuity of multiple elements included in the sequence data. Therefore, by using the integrated likelihood ratio as a classification index, it becomes possible to appropriately select K classes to which the sequence data is likely to belong.

[0110] Seventh embodiment An information processing device 10 according to the seventh embodiment will be described with reference to Fig. 13 and Fig. 14. Note that the seventh embodiment differs in some operations from the above-mentioned first to sixth embodiments, and other operations and device configurations may be similar to those of the first to sixth embodiments. Therefore, hereinafter, the parts that differ from the above-mentioned first to sixth embodiments will be described in detail, and explanations of other overlapping parts will be omitted as appropriate.

[0111] (Operation flow) First, the flow of operations of the information processing device 10 according to the seventh embodiment will be described with reference to Fig. 13. Fig. 13 is a flowchart showing the flow of operations of the information processing device according to the seventh embodiment. In Fig. 13, the same processes as those shown in Fig. 3 are denoted by the same reference numerals.

[0112] 13, when the operation of the information processing device 10 according to the seventh embodiment is started, first, the acquiring unit 50 acquires one element from the sequence data (step S11). The acquiring unit 50 outputs the acquired element of the sequence data to the index calculation unit 100. Then, the index calculation unit 100 calculates a classification index based on the two or more acquired elements (step S12).

[0113] Next, the class selection unit 200 determines whether or not it is possible to select K classes (i.e., classes to which the sequence data is likely to belong) based on the calculated classification index (step S13). If the class selection unit 200 determines that it is not possible to select K classes based on the calculated classification index (step S13: NO), the information processing device 10 starts the process again from step S11. That is, the information processing device 10 obtains the next element of the sequence data and calculates a new classification index.

[0114] On the other hand, if it is determined that K classes can be selected based on the calculated classification index (step S13: YES), the class selection unit 200 selects K classes based on the calculated classification index (step S14). That is, the class selection unit 200 selects K classes from the N classes that are classification candidates.

[0115] Next, particularly in the seventh embodiment, the class selection unit 200 ranks the selected K classes in descending order of the likelihood that the sequence data belongs to them (step S71). For example, when selecting K classes based on the timing at which a predetermined threshold is reached as in the already-described fourth embodiment, the class selection unit 200 may rank the K classes in descending order of the timing at which the predetermined threshold is reached. Alternatively, when selecting K classes based on the magnitude of the classification index at a timing after a predetermined time has elapsed as in the already-described fifth embodiment, the class selection unit 200 may rank the K classes in descending order of the magnitude of the classification index.

[0116] After that, the class selection unit 200 outputs information about the ranked K classes (step S72). That is, the class selection unit 200 outputs information about the selected K classes together with information about their rankings.

[0117] (Example of output result) Next, a display example of an output result of the information processing device 10 according to the seventh embodiment will be specifically described with reference to Fig. 14. Fig. 14 is a diagram showing a display example of an output class by the information processing device according to the seventh embodiment.

[0118] As shown in Fig. 14, the output result of the information processing device 10 according to the seventh embodiment may be displayed on a display device such as a display. In this case, the display device may be the above-mentioned output device 16 (see Fig. 1) or may be an external display or the like. The output result of the information processing device 10 may be output as sound via a speaker or the like.

[0119] The output result of the information processing device 10 according to the seventh embodiment may be displayed in a list format so that the order can be understood. In the example (K=4) shown in Fig. 14, the order and the name of the class are displayed in a ranking format. Specifically, it is displayed that the first place is class B, the second place is class D, the third place is class F, and the fourth place is class A. Note that, for the K ranked classes, numerical values ​​indicating the magnitude of their classification indexes (in other words, quantitative values ​​indicating the likelihood that the sequence data belongs to them) may be displayed.

[0120] The selected K classes may be displayed in different display modes depending on their ranking. For example, only the first-ranked class may be displayed in a conspicuous color (e.g., red or yellow), while the non-selected K classes may be displayed in a relatively inconspicuous color (e.g., black or brown). Also, only the first-ranked class may be displayed blinking, while the classes of other rankings are displayed without blinking. Also, the higher the ranking of a class, the larger the font or the bolder the characters may be displayed.

[0121] The display mode of the screen may be appropriately changed by a user operation or the like. For example, the display mode may be switched from a state in which the ranks are displayed in descending order to a state in which the ranks are displayed in ascending order by a user operation. Also, the display mode according to the above-mentioned rank may be switched ON / OFF by a user operation. An arbitrary class out of the selected K classes may be prevented from being displayed (i.e., deleted from the screen) by a user operation. Alternatively, it may be possible to mark only an arbitrary class out of the selected K classes by a user operation. If K is a relatively large value and the K classes cannot be displayed on one screen (i.e., the number of classes is too large to fit within the screen), the user may operate to turn the page so that all the classes can be confirmed. Alternatively, the display magnification may be changed by a user operation. When the user selects one of the K classes, detailed information (e.g., information registered in advance) regarding the selected class may be displayed in a pop-up.

[0122] Furthermore, in addition to the K selected classes, the classes that were not selected may be displayed. In this case, all the classes that were not selected may be displayed, or only the classes that have the largest classification index among the classes that were not selected (for example, the classes that are not selected, up to the third largest in terms of classification index) may be displayed. The classes that were not selected may also be displayed in order. Furthermore, for the K ranked classes, numerical values ​​indicating the magnitudes of their classification indexes may be displayed. Whether or not the classes that were not selected are displayed may be switchable by a user operation.

[0123] When displaying the classes not selected as described above, the K selected classes and the classes not selected may be displayed in different display modes so that they can be distinguished from each other. For example, the K selected classes may be displayed in a relatively conspicuous color (e.g., red or yellow), while the K non-selected classes may be displayed in a relatively inconspicuous color (e.g., black or brown). Furthermore, only the K selected classes may be displayed in a blinking manner, while the non-selected classes may be displayed without blinking. Furthermore, the K selected classes may be displayed in a large font or bold characters, while the non-selected classes may be displayed in a small font or thin characters.

[0124] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the seventh embodiment will be described.

[0125] 13 and 14, in the information processing device 10 according to the seventh embodiment, the selected K classes are ranked and then output. In this way, it becomes possible to easily grasp the likelihood that the sequence data belongs to the class from the output result.

[0126] Eighth embodiment An information processing device 10 according to the eighth embodiment will be described with reference to Fig. 15. Note that the eighth embodiment differs in some operations from the first to seventh embodiments described above, and other operations and device configurations may be similar to those of the first to seventh embodiments. Therefore, hereinafter, the parts that differ from the first to seventh embodiments described above will be described in detail, and explanations of other overlapping parts will be omitted as appropriate.

[0127] (Operation flow) First, the flow of operations of the information processing device 10 according to the eighth embodiment will be described with reference to Fig. 15. Fig. 15 is a flowchart showing the flow of operations of the information processing device according to the eighth embodiment. In Fig. 15, the same processes as those shown in Fig. 3 are denoted by the same reference numerals.

[0128] 15, when the operation of the information processing device 10 according to the eighth embodiment is started, the acquiring unit 50 first acquires one element from the sequence data (step S11). The acquiring unit 50 outputs the acquired element of the sequence data to the index calculation unit 100. Then, the index calculation unit 100 calculates a classification index based on the two or more acquired elements (step S12).

[0129] Next, the class selection unit 200 determines whether or not it is possible to select K classes (i.e., classes to which the sequence data is likely to belong) based on the calculated classification index (step S13). If the class selection unit 200 determines that it is not possible to select K classes based on the calculated classification index (step S13: NO), the information processing device 10 starts the process again from step S11. That is, the information processing device 10 obtains the next element of the sequence data and calculates a new classification index.

[0130] On the other hand, when it is determined that K classes can be selected based on the calculated classification index (step S13: YES), the class selection unit 200 weights the N classes that are classification candidates (step S81) and selects the K classes taking the weights into consideration (step S82). The weighting here is a weighting related to the ease of selection of the class. For example, of the N classes, a class that is weighted heavily will be more likely to be selected as one of the K classes, and a class that is weighted lightly will be less likely to be selected as one of the K classes. The weights may be set appropriately based on information acquired in advance about each class, etc. The weights may be adjustable by the user.

[0131] More specifically, when a classification index is calculated for each of N classes as in the second embodiment, weighting may be performed on each of the N classification indexes. In this case, weighting may be performed on the classification index that has already been calculated, or may be performed when the classification index is calculated (i.e., at the time of step S12). Alternatively, when thresholds for the N classification indexes are used to select K classes, weighting may be performed on the thresholds. In this case, different thresholds are used for different classes.

[0132] Finally, the class selection unit 200 outputs information about the selected K classes (step S15).

[0133] (Technical effect) Next, technical effects obtained by the information processing device 10 according to the eighth embodiment will be described.

[0134] 15, in the information processing device 10 according to the eighth embodiment, weighting is performed on the N classes that are classification candidates, and the K classes are selected taking the weighting into consideration. In this way, it is possible to adjust the ease with which the N classes are selected, and to more appropriately select the K classes.

[0135] <Ninth embodiment> An information processing device 10 according to the ninth embodiment will be described. The ninth embodiment describes a specific application example of the information processing device according to the first to eighth embodiments described above, and the system configuration and operation flow may be similar to those of the first to eighth embodiments. Therefore, hereinafter, the parts that differ from the first to eighth embodiments described above will be described in detail, and the explanation of other overlapping parts will be omitted as appropriate.

[0136] (Product proposal) The information processing device 10 according to the ninth embodiment may be used to suggest products that a user may be interested in on a web shopping site. Specifically, the information processing device 10 may select K products (i.e., K classes) that the user may be interested in from N products (i.e., N classes) that are handled products, and output them to the user. In this case, examples of the input sequence data include past purchase history, browsing history, etc.

[0137] Similarly, it may be used to suggest products and stores in digital signage, etc. Digital signage may be capable of capturing an image of a user using a camera mounted thereon. In this case, the user's emotions may be estimated from the user's image, and stores and products may be suggested according to the emotions. In addition, the user's line of sight may be estimated from the user's image (i.e., the part the user is looking at may be estimated), and stores and products that the user may be interested in may be suggested. Alternatively, the user's attributes (e.g., gender, age, etc.) may be estimated from the user's image, and stores and products that the user may be interested in may be suggested. When estimating information about a user as described above, weighting may be applied to the N classes according to the estimated information (see the eighth embodiment).

[0138] (Criminal Investigation) The information processing device 10 according to the ninth embodiment can also be used in criminal investigations. For example, when finding the real culprit from among multiple suspects, if only one person who is most likely to be the culprit is selected from among them, a major problem will occur if the selection is incorrect. However, the information processing device 10 according to this embodiment can select and output the top K suspects who are most likely to be the culprit. Specifically, it is sufficient to select and output a class corresponding to the top K suspects who are most likely to be the culprit from sequence data that includes information about each of the multiple suspects as elements. In this way, it is possible to appropriately find the real culprit by investigating, for example, multiple suspects who are most likely to be the culprit.

[0139] (Radar image analysis) The information processing device 10 according to the ninth embodiment can also be applied to the analysis of radar images. Radar images, by their nature, tend to have low clarity, so it is difficult for a machine alone to accurately determine, for example, what is in the image. However, the information processing device 10 according to this embodiment can select and output K candidates that are likely to be in the radar image. Therefore, it is possible to first output K candidates and have the user make a judgment from among them. For example, if "dogs," "cats," "ships," and "tanks" are presented as candidates that appear in a radar image of a port, the user can easily determine that a "ship," which is highly related to the port, appears in the radar image.

[0140] Note that the above-mentioned application example is just one example, and in a situation where it is required to select K candidates from N candidates, beneficial effects can be achieved by applying the information processing device 10 of this embodiment.

[0141] <Additional Notes> The above-described embodiment may be further described as follows, but is not limited to the following.

[0142] (Appendix 1) The information processing device described in Supplementary Note 1 comprises an acquiring means for acquiring a plurality of elements included in sequence data, a calculating means for calculating a classification index indicating the likelihood of a class to which the sequence data belongs based on at least two of the plurality of elements, and a selecting means for selecting and outputting K classes (where K is a natural number equal to or less than N and equal to 1) to which the sequence data is likely to belong based on the classification index from N classes (where N is a natural number) that are classification candidates for the sequence data.

[0143] (Appendix 2) The information processing device according to Supplementary Note 2 is the information processing device according to Supplementary Note 1, characterized in that the calculation means calculates the classification index for each of the N classes.

[0144] (Appendix 3) The information processing device described in Supplementary Note 3 is the information processing device described in Supplementary Note 2, characterized in that the calculation means calculates, as the classification index, a likelihood ratio which is a ratio between a first likelihood indicating a likelihood that the sequence data belongs to a certain class and a second likelihood indicating a likelihood that the sequence data does not belong to a certain class.

[0145] (Appendix 4) The information processing device described in Supplementary Note 4 is the information processing device described in Supplementary Note 2 or 3, characterized in that the selection means selects the K classes in order of the earliest time at which the classification index reaches a predetermined threshold, based on the classification index calculated for each of the N classes.

[0146] (Appendix 5) The information processing device described in Supplementary Note 5 is the information processing device described in Supplementary Note 2 or 3, characterized in that the selection means selects the K classes in order of magnitude of the classification index in a predetermined time range, based on the classification index calculated for each of the N classes.

[0147] (Appendix 6) The information processing device described in Appendix 6 is the information processing device described in any one of Appendixes 1 to 5, characterized in that the acquisition means sequentially acquires the multiple elements, and the calculation means calculates the classification index by taking into account the continuity of the at least two elements.

[0148] (Appendix 7) The information processing device described in Supplementary Note 7 is the information processing device described in any one of Supplementary Notes 1 to 6, characterized in that the selection means ranks the K selected classes in order of likelihood to which the sequence data belongs, and outputs the ranked classes.

[0149] (Appendix 8) The information processing device described in Appendix 8 is the information processing device described in any one of Appendixes 1 to 7, characterized in that the selection means weights each of the N classes in terms of how likely it is to be selected, and then selects the K classes.

[0150] (Appendix 9) The information processing method described in Supplementary Note 9 is an information processing method characterized by obtaining a plurality of elements included in sequence data, calculating a classification index indicating a likelihood of a class to which the sequence data belongs based on at least two of the plurality of elements, and selecting and outputting K classes (where K is a natural number equal to or smaller than N and equal to 1) to which the sequence data is likely to belong from N classes (where N is a natural number) that are classification candidates for the sequence data based on the classification index.

[0151] (Appendix 10) The computer program described in Supplementary Note 10 is a computer program that causes a computer to obtain a plurality of elements included in sequence data, calculate a classification index indicating a likelihood of a class to which the sequence data belongs based on at least two of the plurality of elements, and select and output K classes (where K is a natural number equal to or smaller than N and equal to 1) to which the sequence data is likely to belong based on the classification index from N classes (where N is a natural number) that are classification candidates for the sequence data.

[0152] (Appendix 11) The recording medium according to claim 11 is a recording medium having the computer program according to claim 10 recorded thereon.

[0153] This disclosure may be modified as appropriate within the scope that does not contradict the gist or idea of ​​the invention that can be read from the claims and the entire specification, and information processing devices, information processing methods, and computer programs that involve such modifications are also included in the technical idea of ​​this disclosure. [Explanation of symbols]

[0154] 10. Information processing device 11 Processors 50 Acquisition Department 100 Indicator calculation section 110 First Calculation Section 111 Likelihood ratio calculation unit 112 1st memory section 120 Second Calculation Section 121 Integrated Likelihood Ratio Calculation Unit 122 2nd memory section 200 Class Selection Section

Claims

1. An acquisition means for acquiring a plurality of elements included in the sequence data; a calculation means for calculating a classification index indicating a likelihood of a class to which the sequence data belongs based on at least two of the plurality of elements; a selection means for selecting and outputting K classes (where K is a natural number equal to or smaller than N and equal to 1) to which the sequence data is likely to belong from N classes (where N is a natural number) that are candidates for classification of the sequence data based on the classification index; Equipped with The selection means selects the K classes in order of the earliest time at which the classification index reaches a predetermined threshold based on the classification index calculated for each of the N classes.

23. An information processing apparatus comprising:

2. 2. The information processing apparatus according to claim 1, wherein the calculation means calculates the classification index for each of the N classes.

3. 3. The information processing device according to claim 2, wherein the calculation means calculates, as the classification index, a likelihood ratio which is a ratio between a first likelihood indicating a likelihood that the sequence data belongs to a certain class and a second likelihood indicating a likelihood that the sequence data does not belong to a certain class.

4. The acquiring means acquires the plurality of elements sequentially, 4. The information processing apparatus according to claim 1, wherein the calculation means calculates the classification index by taking into account continuity of the at least two elements.

5. 5. The information processing apparatus according to claim 1, wherein the selection means ranks the K selected classes in order of the likelihood that the sequence data belongs to the classes, and outputs the ranked classes.

6. 6. The information processing apparatus according to claim 1, wherein the selection means selects the K classes after weighting each of the N classes in terms of how likely it is to be selected.

7. 7. The information processing apparatus according to claim 1, wherein the selection means reduces the value of K when the K classification indices do not reach a predetermined threshold within a time limit.

8. 7. The information processing apparatus according to claim 1, wherein the selection means selects the K classes in descending order of the classification index when the K classification indexes do not reach a predetermined threshold within a time limit.

9. 9. The information processing device according to claim 1 , wherein the selection means outputs, based on the classification index calculated for each of the N classes, a class that reaches a rejection threshold earliest among the classes not selected as the K classes.

10. 9. The information processing apparatus according to claim 1, wherein the calculation means stops calculation of the classification index for a classification index that has reached the predetermined threshold or a rejection threshold before the K classes are selected.

11. by at least one computer, Obtain multiple elements contained in the sequence data, calculating a classification index indicating a likelihood of a class to which the sequence data belongs based on at least two of the plurality of elements; selecting and outputting K classes (where K is a natural number equal to or smaller than N and equal to 1) to which the sequence data is likely to belong from N classes (where N is a natural number) that are candidates for classification of the sequence data based on the classification index; and When selecting the K classes, the K classes are selected in order of the earliest time at which the classification index reaches a predetermined threshold based on the classification index calculated for each of the N classes.

23. An information processing method comprising:

12. Obtain multiple elements contained in the sequence data, calculating a classification index indicating a likelihood of a class to which the sequence data belongs based on at least two of the plurality of elements; selecting and outputting K classes (where K is a natural number equal to or smaller than N and equal to 1) to which the sequence data is likely to belong from N classes (where N is a natural number) that are candidates for classification of the sequence data based on the classification index; and When selecting the K classes, the K classes are selected in order of the earliest time at which the classification index reaches a predetermined threshold based on the classification index calculated for each of the N classes. A computer program for causing a computer to operate in such a manner.

Images