This application is based on and claims priority under 35 USC 119 from Japanese Patent Application No. 2017-059234 filed Mar. 24, 2017.
The present invention relates to an information processing apparatus.
According to an aspect of the invention, an information processing apparatus includes a first acquisition unit, a second acquisition unit, an evaluation unit, and an output unit. The first acquisition unit acquires first information indicating groups of plural articles. The second acquisition unit acquires second information indicating purchase histories of the respective articles. The evaluation unit evaluates, for each group, a relationship between the articles based on the purchase histories of the articles. The output unit performs output in accordance with the evaluation by the evaluation unit.
Exemplary embodiments of the present invention will be described in detail based on the following figures, wherein:
1-1. Overall Configuration of Information Processing System
Each terminal 2 shown in
The information processing apparatus 1 illustrated in
1-2. Configuration of Terminal
The controller 21 includes a central processing unit (CPU), a read only memory (ROM), and a random access memory (RAM). The CPU reads and executes computer programs (hereinafter simply referred to as programs) stored in the ROM and the memory unit 22, thereby controlling the respective units of the terminal 2.
The memory unit 22 is a large-capacity memory unit such as a solid state drive or a hard disk drive, and stores various programs to be read by the CPU of the controller 21.
The communication unit 23 is a communication circuit that is connected to the communication line 4 in a wireless or wired manner. The terminal 2 exchanges information with the information processing apparatus 1 via the communication line 4 by the communication unit 23 thereof.
The operation unit 25 includes an operator such as an operation button for giving various instructions. The operation unit 25 receives an operation by a user and supplies a signal corresponding to an operation content to the controller 21. In addition, the operation unit 25 may include a touch panel which detects the contact, pressure, or the like of an operation body such as a user's finger or a stylus pen.
The display 24 includes a liquid crystal display, and displays an image under the control of the controller 21. The transparent touch panel of the operation unit 25 may be disposed to overlap the liquid crystal display of the display 24.
1-3. Configuration of Information Processing Apparatus
The controller 11 includes a CPU, a ROM, and a RAM. The CPU reads and executes programs stored in the ROM and the memory unit 12, thereby controlling the respective units of the information processing apparatus 1.
The communication unit 13 is a communication circuit that is connected to the communication line 4 in a wireless or wired manner. The information processing apparatus 1 exchanges information with the terminal 2 via the communication line 4 by the communication unit 13 thereof.
The operation unit 15 includes an operator such as an operation button for giving various instructions. The operation unit 15 receives an operation by a user and supplies a signal corresponding to the operation content to the controller 11. In addition, the operation unit 15 may include a touch panel which detects an operation body such as a user's finger or a stylus pen.
The display 14 includes a display screen such as a liquid crystal display, and displays an image under the control of the controller 11. The transparent touch panel of the operation unit 15 may be disposed to overlap the display screen of the display 14.
The memory unit 12 is a large-capacity memory unit such as a hard disk drive, and stores various programs to be read by the CPU of the controller 11. In addition, the memory unit 12 stores an article DB 121, a purchase history DB 122, and a sales period DB 123.
The article DB 121 is a database that stores information on articles such as article identification information and classification of the articles. The purchase history DB 122 is a database that stores a history (purchase history) indicating when an article was purchased. The sales period DB 123 is a database that stores a period (sales period) during which an article has been sold.
An example of a category illustrated in
A group G1 illustrated in
Here, this relationship of the group G1 is referred to as a “relationship having complementarity.” “Cream powder” is an article that complements “instant coffee” premised on the presence of “instant coffee,” that is, a complementary article of “instant coffee.”
In addition, a group G2 illustrated in
Here, this relationship in the group G2 is referred to as a “substitutional relationship.” “Plain milk,” “high-density milk,” and “low fat milk” are referred to mutually substitutive articles.
The information processing apparatus 1 specifies, for each group of two or more articles, in what manner the articles belonging to each group are purchased using the databases stored in the memory unit 12 and evaluates a relationship between the articles.
For example, the article DB 121 illustrated in
For example, the purchase history DB 122 illustrated in
For example, the sales period DB 123 illustrated in
1-4. Functional Configuration of Information Processing Apparatus
The first acquisition unit 111 acquires first information indicating groups of multiple articles from the article DB 121 stored in the memory unit 12. The number of articles constituting a group may be two or may be three or more as long as the articles are plural different articles. The first acquisition unit 111 may acquire, for example, all groups of two articles among the articles stored in the article DB 121. In addition, the groups of articles acquired by the first acquisition unit 111 may be article category groups.
The second acquisition unit 112 acquires second information indicating purchase histories of the respective multiple articles from the purchase history DB 122 stored in the memory unit 12.
The evaluation unit 114 evaluates, for each group indicated by the first information acquired by the first acquisition unit 111, a relationship between the articles based on the purchase histories of the respective articles. The evaluation unit 114 illustrated in
In addition, the second acquisition unit 112 may acquire second information indicating a purchase history for each user. In this case, the controller 11 may serve as the correction unit 116 which specifies multiple users having similar article purchase histories based on the acquired second information and corrects the above-described second information for these users. Then, the evaluation unit 114 may evaluate a relationship between the above-described articles based on the corrected second information.
The output unit 115 performs output in accordance with the evaluation by the evaluation unit 114. For example, the output unit 115 may output, for each group of articles, information indicating which article included in the group is superior. In addition, the output unit 115 may output, based on the superiority and inferiority of articles included in a group, information indicating the display mode, arrangement, quantity, advertisement layout of the articles in a real store, the timing and frequency of voice guidance in the store, and the like.
The third acquisition unit 113 acquires third information indicating periods during which the respective multiple articles were sold. When the controller 11 serves as the third acquisition unit 113, the evaluation unit 114 specifies, using the third information acquired by the third acquisition unit 113, a period during which all of the multiple articles included in the groups which are indicated by the first information acquired by the first acquisition unit 111 were sold. Then, the evaluation unit 114 may evaluate a relationship between the articles based on the purchase histories of the articles in the specified period.
1-5. Operation of Information Processing Apparatus
When the controller 11 serves as the above-described correction unit 116, the controller 11 specifies multiple users having similar article purchase histories by, for example, a collaborative filtering method based on the second information acquired in step S102. This similarity determination may be made based on, for example, a Euclidean distance between multidimensional vectors indicated by respective purchase histories of the users. Then, the controller 11 corrects a purchase history based on the purchase histories of the specified similar users (step S104).
Subsequently, the controller 11 specifies a difference between the respective purchase histories of the user “L” and the user “N.” In the example illustrated in
Therefore, the controller 11 makes a correction to replace the purchase history used for the evaluation by data corresponding to a case where the user “L” purchases the article “C,” for example, at the same level as the user “N.” Thereby, the information processing apparatus 1 excludes the difference from another user having a similar purchase history as an abnormal value. Therefore, for example, the prediction of a future purchase behavior is corrected for an article that has never been purchased due to the lack of knowledge or experience of the user, but is highly likely to be purchased (so-called a “thoughtlessly avoided” article) when there is a chance.
In addition, the number of purchases may be corrected based on a statistical representative value such as the arithmetic mean, mode, or intermediate value of the number of purchases of multiple users having similar purchase histories. In this case, data to be corrected may be specified based on whether or not the difference from the statistical representative value exceeds a threshold value.
As illustrated in
When evaluating the group of the article “A” and the article “B,” the controller 11 performs evaluation based on the purchase histories thereof for the period during which the respective sales periods thereof overlap each other.
For example, the controller 11 calculates a rate of the number of accounts in which the article “A” was purchased among accounts performed for the above-described period, as a probability P(A1) of the article “A” being purchased. Specifically, the controller 11 calculates, as the probability P(A1), a numerical value obtained by dividing the number of accounts in which the article “A” was purchased by the total number of accounts for this period. In this case, the probability P(A0) of the article “A” being not purchased is indicated by the following formula (1).
P(A0)=1−P(A1) (1)
Here, the “number of accounts” may be obtained by regarding accounts performed by a common user for a fixed period such as within 2 hours or for the same day as one account.
The controller 11 calculates a probability P(B1) of the article “B” being purchased and a probability P(B0) of the article “B” being not purchased.
In addition, the controller 11 calculates a probability P(A1, B0) of the article “A” being purchased and the article “B” being not purchased. Specifically, the controller 11 calculates the probability P(A1, B0) by obtaining the rate of the number of accounts in which the article “A” is purchased and the article “B” is not purchased among accounts performed for the above-described period.
Then, the controller 11 calculates a probability P(A0, B1) of the article “A” being not purchased and the article “B” being purchased and a probability P(A1, B1) of the article “A” being purchased and the article “B” being purchased.
The controller 11 obtains mutual information amount I(A, B) indicated by the purchase histories of the article “A” and the article “B,” using the calculated probabilities. The mutual information amount I(A, B) is represented by the following formula (2).
In the formula (2), subscripts i and j represent “not purchased” in the case of “0” and “purchased” in the case of “1.”
The mutual information amount I(A, B) represents how much information on the other article (here, assumed as the article “B”) is known when it is found that one article (here, assumed as the article “A”) is purchased. That is, a pair of articles having a low mutual information amount is a highly independent article pair, and the purchase of one article has no effect on the purchase of the other article. On the other hand, for an article pair having a high mutual information amount, the purchase of one article has an effect on the purchase of the other article. The effect at this time includes a case where the other article is always purchased when one article is purchased and a case where the other article is never purchased when one article is purchased.
In addition, the controller 11 evaluates a relationship between the article “A” and the article “B” based on a relative rates of the mutual information amount I(A, B), the probability P(A1, B0), the probability P(A0, B1), and the probability P(A1, B1).
In this case, the controller 11 compares these rates with a predetermined threshold value, determines that the article “A” and the article “B” are in a “relationship having complementarity,” and rates the article “B” as a complementary article of the article “A.”
In addition, in the example illustrated in
In this case, the controller 11 compares these rates with the predetermined threshold value, determines that the article “A” and the article “B” are in a “relationship having substitutivity,” and rates the article “A” and the article “B” as mutually substitutive articles.
In addition, in the example illustrated in
In this case, the controller 11 compares the rates of these with the predetermined threshold value, determines that the article “A” and the article “B” are in a “relationship having simultaneity,” and rates the article “A” and the article “B” as articles that are likely to be purchased at the same time.
Based on the comparison of the mutual information amount I(A, B) and the probabilities described above, the controller 11 evaluates, for each group of articles, a relationship between the articles included in the group and outputs the evaluation result (step S106). Then, the controller 11 determines whether or not there exists an unevaluated group (step S107). If determining that there exists an unevaluated group, the controller 11 returns the process to step S101 (step S107; YES). On the other hand, if determining that there exists no unevaluated group (step S107; NO), the controller 11 terminates the process.
With the above operation, the information processing apparatus 1 of the information processing system 9 evaluates, for each group of articles, a relationship between the articles based on the respective purchase histories of the articles. Thus, the user determines, for example, which article is to be advertised or displayed conspicuously in consideration of the mutual influence of multiple articles on the purchase behaviors of customers.
For example, if a rate at which the article “B” (second article) is purchased when the article “A” (first article) is purchased is larger by more than a threshold value than a rate at which the article “A” (first article) is purchased when the article “B” (second article) is purchased, the information processing apparatus 1 determines that the article “A” (first article) and the article “B” (second article) are in a “relationship having complementarity” and rates the article “B” (second article) as a complementary article of the article “A” (first article).
At this time, the information processing apparatus 1 rates the article “B” (second article) as being superior to the article “A” (first article), and proposes, for example, a sales plan in which the article “B” (second article) is advertised or displayed more conspicuously than the article “A” (first article).
In addition, when handling a new article, the user may specify an article category of the new article and may determine a method of advertising or displaying the new article based on a relationship between the new article and an existing article which is evaluated by the information processing apparatus 1 for each article category group.
The exemplary embodiment has been described above. The contents of the exemplary embodiment may be modified as follows. In addition, the following modifications may be combined with each other.
2-1. Modification 1
In the above-described exemplary embodiment, the information processing apparatus 1 of the information processing system 9 stores the article DB 121, the purchase history DB 122, and the sales period DB 123 in the memory unit 12. These databases may be managed by a server device other than the information processing apparatus 1. In this case, the information processing apparatus 1 may transmit a query requesting data to the server device via the communication line 4, and the server device may supply data corresponding to the query. In addition, these databases may not be generated by accumulating information transmitted from the terminal 2, but may be generated, for example, by a user's operation in the server device or in the information processing apparatus 1. In this case, the information processing system 9 may not include the terminal 2.
2-2. Modification 2
In the above-described exemplary embodiment, the evaluation unit 114 implemented by the controller 11 of the information processing apparatus 1 evaluates the relationship based on the rate at which when one of articles included in a group is purchased, another article included in the group is purchased. It should be noted that the evaluation unit is not limited to this example. For example, in a case where three or more articles are included in a group, when one article included in the group is purchased and another article among the other articles included in the group is further purchased, the evaluation unit 114 may evaluate a relationship between the three or more articles based on a rate at which an article which is included in the group but is not any of the above described articles is purchased.
2-3. Modification 3
In the above-described exemplary embodiment, when the controller 11 of the information processing apparatus 1 serves as the third acquisition unit 113 which acquires third information indicating the sales periods of the multiple articles, the controller 11 serves as the evaluation unit 114 which evaluates the relationship between the articles based on the purchase histories of the articles for the period during which the sales periods of multiple articles included in the group overlap each other. Alternatively, the controller 11 may evaluate the relationship without considering the sales periods. In this case, the controller 11 may not serve as the third acquisition unit 113, and the memory unit 12 may not store the sales period DB 123.
2-4. Modification 4
In the above-described exemplary embodiment, the controller 11 corrects the purchase history based on the purchase histories of the specified similar users. Alternatively, the controller 11 may not make this correction and may not specify users having similar purchase histories.
2-5. Modification 5
In the above-described exemplary embodiment, the evaluation unit 114 implemented by the controller 11 of the information processing apparatus 1 evaluates the relationship between the first article and the second article as the second article being superior to the first article, if a rate at which the second article is purchased when the first article is purchased is larger by more than the threshold value than a rate at which the first article is purchased when the second article is purchased. It should be noted that an evaluation unit is not limited thereto.
For example, if the rate at which the first article and the second article are purchased at the same time is larger by more than the predetermined threshold value than a rate at which only one of the first and second articles is purchased, the controller 11 may propose a sales plan in which a third article which is neither the first article nor the second article is recommended to the user. This is because, when there is a high possibility of the first article and the second article being purchased together, it is predicted that there is no change in the number of purchases even if one of the first article and the second article is recommended.
In addition, for example, if a rate at which only one of the first article and the second article is purchased is larger by more than a predetermined threshold value than a rate at which the first article and the second article are purchased at the same time, the controller 11 may propose a sales plan in which one of the first article and the second article is randomly selected and is recommended to the user. The first article and the second article are in a relationship having substitutability. Therefore, even if one of the first article and the second article is purchased in response to the recommendation, there is a low possibility that the other is purchased.
2-6. Modification 6
A program executed by the controller 11 of the information processing apparatus 1 may be provided in a state in which it is stored in a computer-readable recording medium including a magnetic recording medium such as a magnetic tape or a magnetic disc, an optical recording medium such as an optical disc, a magneto-optical recording medium, a semiconductor memory or the like. In addition, the program may be downloaded via a communication line such as the Internet. In addition, as a control device exemplified by the controller 11 described above, various devices other than the CPU may be applied, and for example, a dedicated processor or the like is used.
The foregoing description of the exemplary embodiments of the present invention has been provided for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Obviously, many modifications and variations will be apparent to practitioners skilled in the art. The embodiments were chosen and described in order to best explain the principles of the invention and its practical applications, thereby enabling others skilled in the art to understand the invention for various embodiments and with the various modifications as are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the following claims and their equivalents.
Number | Date | Country | Kind |
---|---|---|---|
2017-059234 | Mar 2017 | JP | national |