KR20140133663A - System and method for products recommendation service, and apparatus applied to the same - Google Patents

Abstract

Disclosed are a system and a method for a product recommendation service and an apparatus applied to the same. In other words, by generating at least two recommended product lists for a buyer based on an individual recommendation algorithm, selecting a recommended product appropriate to the buyer from at least one among the recommended product lists based on the suitability of each individual recommendation algorithm for the buyer, and providing the selected product; the reliability of the recommended product is improved. As a result, the satisfaction of the buyer can be increased and the product seller′s revenue increased at the same time.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a product recommendation service system,

The present invention relates to a method for organizing and offering a list of recommended products by selecting and recommending an individual recommendation algorithm suitable for a buyer among various individual recommendation algorithms for determining a recommended product in relation to a product recommendation service.

In a service page including the Internet or a mobile portal site, a product recommendation service that recommends a product suitable for a buyer (or a customer who visits a purchase) is provided.

In this product recommendation service, various recommendation algorithms for determining recommended products are generally applied to construct and provide a list of recommended products suitable for the buyer.

Accordingly, in order to provide a highly satisfied commodity recommendation service for purchasers, it is important to select a recommendation algorithm that is most suitable for a purchaser from among various individual recommendation algorithms for determining a recommended commodity, and to provide a list of recommended commodities accordingly something to do.

The object of the present invention is to provide a method and apparatus for generating at least two recommendation product lists for a purchaser based on an individual recommendation algorithm, By selecting and providing a recommended product suitable for the purchaser from at least one of the at least two recommended product catalogs based on the goodness of fit with the purchaser, it is possible to increase the reliability of the recommended product to increase the satisfaction of the purchaser, To increase profits.

According to a first aspect of the present invention, there is provided a service apparatus comprising: a generation unit configured to generate at least two recommended product lists for a purchaser based on an individual recommendation algorithm; A determination unit for determining a fitness of each of the individual recommendation algorithms with the purchaser; And a selection unit for selecting and providing a recommended product suitable for the buyer from at least one of the at least two recommended product lists based on the fitness with the buyer for each of the individual recommendation algorithms.

More specifically, the service apparatus may further include a collecting unit for collecting a list of purchased goods purchased by the purchaser, wherein the determining unit determines, from each of the recommended goods list created based on the individual recommendation algorithm, And a recommendation item included in the list is checked to determine the fitness with the purchaser for each of the recommendation algorithms.

More specifically, for each of the individual recommendation algorithms, the determination unit may determine, for each of the recommendation algorithms, the number of all purchased products included in the purchased product list, And determines the fitness with the purchaser based on the ratio between the number of customers and the number of customers.

More specifically, each of the recommended products included in the at least two recommended product lists is given an individual recommendation score calculated by each of the individual recommendation algorithms, A recommended recommendation score from at least one of the at least two recommended recommendation scores based on a final recommendation score calculated by further applying a goodness-of-fit with the purchaser to each of the individual recommendation scores for each of the individual recommendation scores, And selects a product.

More specifically, each of the recommended products included in the at least two recommended product lists is assigned a priority according to the final recommendation score size, and the selection unit selects one of the recommended products among the recommended products And the recommendation commodity having the high priority is selected as a recommended commodity suitable for the purchaser.

According to a second aspect of the present invention, there is provided a merchandise recommendation service method comprising: a request step of a terminal device requesting a recommended merchandise for a buyer; A generating step of generating a list of at least two recommended products for the purchaser based on an individual recommendation algorithm of the service device; And a selection step of selecting a recommended product suitable for the buyer from at least one of the at least two recommended product lists based on the goodness of fit with the buyer for each of the individual recommendation algorithms and providing the recommended product to the terminal .

According to a third aspect of the present invention, there is provided a method of operating a service device, the method comprising: generating a list of at least two recommended goods for a buyer based on an individual recommendation algorithm; A determination step of determining a fitness of each of the individual recommendation algorithms with the purchaser; And a selection step of selecting and providing a recommended product suitable for the buyer from at least one of the at least two recommended product lists based on the fitness with the buyer for each of the individual recommendation algorithms .

More specifically, the method may further include a collecting step of collecting a list of purchased goods purchased by the purchaser before the determining step, wherein the determining step comprises: And checking a recommendation product included in the purchased product list in the product list to determine a fitness with the purchaser for each of the recommendation algorithms.

More specifically, the determining step may include: determining, for each of the recommendation algorithms, the number of all purchased products included in the purchased product list, The degree of fitness with the purchaser is determined based on the ratio between the number of the buyers.

More specifically, each recommended product included in the at least two recommended product lists is assigned an individual recommendation score calculated by each of the individual recommendation algorithms, and the selecting step includes: Wherein the recommendation score is calculated from at least one of the at least two recommended product lists based on a final recommendation score calculated by further applying a fitness degree with each of the individual recommendation algorithms to each of the individual recommendation scores of each product, And selects a recommended product.

More specifically, each of the recommended products included in the at least two recommended product lists is assigned a priority according to the size of the final recommendation score, and the selecting step selects all recommendation products included in the at least two recommended product lists The recommended product having the highest priority is selected as a recommended product suitable for the purchaser.

According to a fourth aspect of the present invention, there is provided a computer-readable medium including: a generating step of generating at least two recommended product lists for a purchaser based on an individual recommendation algorithm; A determination step of determining a fitness of each of the individual recommendation algorithms with the purchaser; And a selection step of selecting and providing a recommendation item suitable for the buyer from at least one of the at least two recommendation goods lists based on the fitness with the buyer for each of the individual recommendation algorithms .

According to a fifth aspect of the present invention, there is provided a merchandise recommendation service system comprising: a terminal device for requesting and receiving a recommendation item for a purchaser; And generating a list of at least two recommendation products for the buyer based on the individual recommendation algorithm, and based on the fitness with the buyer for each of the recommendation algorithms, And a service device that selects a recommended product suitable for the terminal device and provides the selected device to the terminal device.

According to the commodity recommendation service system and method of the present invention, and the apparatus to which the present invention is applied, among the various recommendation algorithms for determining recommended commodities, a recommendation commodity list is prepared Thus, the reliability of the recommended product can be increased, the satisfaction of the purchaser can be increased, and the profit of the seller of the product can be increased.

1 is a schematic configuration diagram of a commodity recommendation service system according to an embodiment of the present invention;
2 is a configuration diagram of a service apparatus according to an embodiment of the present invention;
3 is a schematic flowchart for explaining an operation flow in a goods recommendation service system according to an embodiment of the present invention;
4 is a schematic flowchart for explaining an operational flow in a service apparatus according to an embodiment of the present invention;

Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.

1 is a diagram illustrating a product recommendation service system according to an embodiment of the present invention.

1, a product recommendation service system according to an embodiment of the present invention includes a terminal device 100 for requesting and receiving a recommended product suitable for a buyer from a service device 200, And a service device 200 for providing the service.

The terminal device 100 refers to a user device for connecting to the service device 200 and receiving a recommendation item suitable for the purchaser, for example, a smart phone, a personal computer, a notebook, and a tablet PC And any device that can be connected to the service apparatus 200 may be included.

The service device 200 refers to a server for providing a recommendation product to a purchaser who accesses a service page through the terminal device 100. The service device 200 is intended for a purchaser who accesses a service page through the Internet or a mobile web Or may be configured as a separate server for providing a recommended product at the same time as selling the product or providing only the recommended product.

Here, in the case of a product sold to a purchaser, for example, it is possible to classify a product category according to its characteristics such as an application and contents, and furthermore, various categories of goods (e.g., games, , Drama, books, etc.).

Meanwhile, according to an embodiment of the present invention, the service apparatus 200 provides a product recommendation service by constructing a recommended product list suitable for a buyer by applying various individual recommendation algorithms for determining a recommended product.

Here, in the case of the individual recommendation algorithm, for example, similar to the above, the collaborative filtering (Collaborative Filtering), which is a method of recommending items purchased by a large number of purchasers, Association Rule, which is a method of recommending B to users who purchased A, is applied.

In addition, various individual recommendation algorithms such as continuity recommendation, genre recommendation, sex / age best recommendation, and person similarity recommendation according to service or product characteristics may be additionally applied.

In the case of the individual recommendation algorithm, it is difficult to directly evaluate the performance between the algorithms because it has its own performance measure.

For example, an individual recommendation algorithm used in a product recommendation service creates a list of recommended products according to individual characteristics, and the recommended products included in the recommended product list are determined according to the criteria of the individual recommendation algorithm itself .

However, since it is impossible to compare different individual recommendation algorithms as described above, an artificial work for integrating the results derived from each individual recommendation algorithm is required.

As described above, in order to maintain a high service satisfaction of a buyer in a product recommendation service applying a plurality of individual recommendation algorithms, it is more important than selecting an individual recommendation algorithm most suitable for a buyer among a plurality of individual recommendation algorithms .

Accordingly, in an embodiment of the present invention, among the various individual recommendation algorithms for determining a recommendation commodity in relation to a commodity recommendation service, a recommendation list for recommending commodities is prepared by selecting a recommendation algorithm suitable for a purchaser Which will be described in detail below.

The terminal device 100 accesses the service device 200 and performs a function of requesting a recommended product.

More specifically, the terminal device 100 accesses the service page through the Internet or the mobile web, for example, by transmitting personal information (e.g., ID / PW) of the purchaser to the service device 200, ) To determine a recommended product suitable for the buyer based on the personal information of the purchaser.

The service device 200 performs a function of generating a plurality of recommendation product lists for a purchaser.

More specifically, when the personal information of the purchaser is received from the terminal apparatus 100 connected to the service page, and the recommendation commodity is requested, the service apparatus 200 transmits a plurality of recommendations A list of goods is generated.

In addition, the service apparatus 200 performs a function of determining the fitness of each individual recommendation algorithm with the purchaser.

More specifically, when the generation of the plurality of recommended goods lists for the buyer is completed based on the individual recommendation algorithm, the service device 200 generates a fitness for each of the individual recommendation algorithms .

To this end, the service device 200 collects a list of purchase items purchased by past buyers based on the buyer's personal information, in order to determine the fitness with the buyer for each of the recommendation algorithms.

On the basis of this, the service apparatus 200 confirms the recommendation commodities included in the purchase commodity list in each recommendation commodity list generated on the basis of the past recommendation algorithm, The fitness is judged.

In addition, the service device 200 performs a function of selecting a recommended product suitable for the purchaser.

More specifically, when the fitness of the purchaser with each of the individual recommendation algorithms is determined, the service apparatus 200 generates recommendation items And provides it to the terminal device 100.

At this time, the service apparatus 200 further applies the fitness with the purchaser to each of the recommendation scores of all the recommended products included in the plurality of recommended product lists, And a recommendation commodity with a size of the final recommendation score is preferentially selected as a recommendation commodity suitable for the buyer.

Hereinafter, the configuration of the service apparatus 200 according to an embodiment of the present invention will be described in more detail with reference to FIG.

That is, the service apparatus 200 according to an exemplary embodiment of the present invention includes a generator 210 for generating a list of recommended products, a determination unit 220 for determining a fitness with the purchaser, And a selection unit (230).

The service device 200 according to an embodiment of the present invention may further include a receiver 240 for receiving buyer identification information from the terminal device 100, a collecting unit 250 for collecting the purchased goods list And a providing unit 260 for providing a recommended product to the terminal apparatus 100. [

Here, each configuration of the service apparatus 200 including the above-described generating unit 210, the determining unit 220, the selecting unit 230, the receiving unit 240, the collecting unit 250, and the providing unit 260 May be embodied in software modules executed by a processor, or a combination thereof.

The generation unit 210 performs a function of generating a plurality of recommended product lists for the purchaser.

More specifically, when the personal information of the purchaser is received from the terminal device 100 connected to the service page through the reception unit 240 and a recommendation item is requested, the generation unit 210 generates a buyer A plurality of recommended product lists to be targeted are generated.

For example, the generating unit 210 generates a list of recommended products based on a similarity-based recommendation algorithm (algorithm 1) as shown in Table 1 below and a list of recommended products based on the reliability [Confidence, P A list of recommended goods based on the individual recommendation algorithm (algorithm 2) based on B (A) = P (A∩B) / P B) = P (A) * P (B)] based on the individual recommendation algorithm (algorithm N).

At this time, the generating unit 210 assigns individual recommendation scores (Score) according to the self-reference for each recommendation algorithm to each of the recommendation products included in each recommended product list, and assigns a priority (Rank) [Score (i, j): Individual recommendation score (1? I? N) of the j-th result of the i-th individual recommendation algorithm].

Rank Item Score (1, rank) One AA01 0.98 2 AA02 0.94 ... ... 99 AA99 0.33 100 AA100 0.31

Rank Item Score (2, rank) One BB01 0.99 2 BB02 0.92 ... ... 99 BB99 0.12 100 BB100 0.11

Rank Item Score (N, rank) One ZZ01 3.84 2 ZZ02 2.32 ... ... 99 ZZ99 1.14 100 ZZ100 1.08

Here, both the Confidence and Lift-based individual recommendation algorithms refer to an algorithm for evaluating association rules (ARs) between A and B.

 For example, when the reliability and the degree of improvement of (A, B) are respectively '0.3' and '2', and the reliability and the degree of improvement of (A, C) are respectively 0.5 and 1.5, In the case of recommending a product based on the reliability, C is recommended to the purchaser who purchases A, whereas when recommending the product based on the degree of improvement, recommendation of B to the purchaser who purchases A can be recommended.

The generating unit 210 normalizes individual recommendation scores of the recommendation commodities included in the recommendation commodity list when a plurality of recommendation commodity lists based on the recommendation algorithm are generated.

For example, the generating unit 210 may store each of a plurality of recommended product lists targeted at the purchaser, such as those shown in [Table 1] to [Table 3] The individual recommendation score of the recommendation product included in the list is normalized so that the recommendation score may be between '0' and '100'.

Rank Item Score (1, rank) One AA01 100 2 AA02 98 ... ... 99 AA99 2 100 AA100 0

 Rank Item Score (2, rank) One BB01 100 2 BB02 91 ... ... 99 BB99 One 100 BB100 0

Rank Item Score (N, rank) One ZZ01 100 2 ZZ02 83 ... ... 99 ZZ99 2 100 ZZ100 0

The determination unit 220 performs a function of determining fitness with each purchaser for each of the recommendation algorithms.

More specifically, when the generation of the plurality of recommended product lists for the buyer is completed based on the individual recommendation algorithm, the determination unit 220 determines whether or not the fitness for each of the individual recommendation algorithms .

To this end, the collecting unit 250 collects a list of purchase items purchased by the past buyers based on the buyer's personal information, in order to determine the fitness with the buyer for each of the recommendation algorithms.

On the basis of this, the determination unit 220 identifies a recommended product included in the purchased product list in each of the recommended product lists created on the basis of the past individual recommendation algorithm, The fitness is judged.

That is, for each of the recommendation algorithms, the determination unit 220 determines, for each of the recommendation algorithms, the number of all purchased products included in the purchased product list as shown in Equation 1 below, Which is a ratio of the number of recommended products included in the purchase order to the buyer.

[Equation 1]

Goodness of fit with buyer = list of recommended items ∩ Purchase product list / All purchase items * 100

For example, the judging unit 220 judges whether or not a list of recommended products prepared based on the recommendation algorithm for the past one day (for example, yesterday) By comparing the lists, it is possible to judge the fitness with the purchaser for each of the individual recommendation algorithms (algorithm '1' to algorithm 'N') as shown in Table 7 below.

Algorithm 1 Algorithm 2 ... Algorithm N 0.32 0.40 0.25

For reference, in the case of the algorithm 2 in Table 7, it means that there are five purchasers of the buyer, and two of the purchased commodities correspond to the recommended commodity list.

The selecting unit 230 performs a function of selecting a recommended product suitable for the purchaser.

More specifically, when the fitness of each of the individual recommendation algorithms with respect to each purchaser is determined, the selecting unit 230 selects a recommended recommendation item suitable for the purchaser from a plurality of recommended product lists based on the fitness with the purchaser, And provides the providing unit 260 with the terminal device 100. [

At this time, the selecting unit 230 further applies the fitness with the purchaser to each of the individual recommendation scores of all the recommendation goods included in the plurality of recommendation goods lists, as shown in Equation (2) below, Calculate the final recommendation score for each referral product.

[Equation 2]

Final score = Individual recommendation score (Score) * Fit with buyer

When the calculation of the final recommendation score is completed, the selection unit 230 assigns priority to each of the recommended products included in the list of recommended products as shown in [Table 8] according to the size of the final recommendation score, And a product list.

Item Score (i, j) Final_score (i, j) BB01 100 40 BB02 91 36.4 AA01 100 32 AA02 98 31.36 ZZ01 100 25 ZZ02 83 20.75 ... ... ... AA99 2 0.64 ZZ99 2 0.5 BB99 One 0.4 AA100 0 0 BB100 0 0 CC100 0 0

In addition, the selecting unit 230 preferentially selects a recommendation commodity having a higher priority from the total recommended commodity list having a predetermined priority according to the size of the final recommendation score, as a recommendation commodity suitable for the purchaser.

For example, if the number of recommended products to be provided to the purchaser is 100, each recommended product having the order of priority of the top 100 can be selected as a recommended product suitable for the purchaser and provided to the terminal device 100 have.

As described above, according to the merchandise recommendation service system according to the embodiment of the present invention, among the various individual recommendation algorithms for determining the recommended merchandise, Thus, the reliability of the recommended product can be increased, the satisfaction of the purchaser can be increased, and the profit of the seller of the product can be increased.

Hereinafter, a method of recommending a product according to an embodiment of the present invention will be described with reference to FIGS. 3 and 4. FIG. Here, for convenience of explanation, the configurations shown in FIGS. 1 and 2 will be described with reference to corresponding reference numerals.

First, referring to FIG. 3, an operation flow in a goods recommendation service system according to an embodiment of the present invention will be described.

First, the terminal device 100 accesses a service page via the Internet or the mobile web and transmits the personal information (e.g., ID / PW) of the purchaser to the service device 200, for example, to the service device 200 (S110) to select a recommended product suitable for the buyer based on the personal information of the purchaser.

Then, when the personal information of the purchaser is received from the terminal device 100 connected to the service page and the recommendation commodity is requested, the service apparatus 200 acquires a plurality of recommendation commodities And generates a list (S120).

At this time, the service apparatus 200 assigns individual recommendation scores (Score) according to the self-reference for each recommendation algorithm to each of the recommendation products included in the recommendation product list, and assigns a priority rank .

Then, when the generation of the plurality of recommended product lists for the buyer is completed based on the individual recommendation algorithm, the service apparatus 200 transmits the personal information of the buyer (S130). The purchase history information of the purchaser is stored in the list of purchased products.

Then, the service apparatus 200 confirms the fitness with the purchaser for each of the recommendation algorithms by confirming the recommendation commodities included in the purchase commodity list in each recommendation commodity list created based on the past individual recommendation algorithm (S140).

Thereafter, the service apparatus 200 selects a recommended product suitable for the buyer from a plurality of recommended product lists based on the fitness with the purchaser for each of the individual recommendation algorithms, And provides it to the terminal device 100 (S150 - S160).

At this time, the service apparatus 200 further applies the fitness with the purchaser to each of the recommendation scores of all the recommended products included in the plurality of recommended product lists, And a recommendation commodity with a size of the final recommendation score is preferentially selected as a recommendation commodity suitable for the buyer.

Hereinafter, the operation of the service device 200 according to an embodiment of the present invention will be described in detail with reference to FIG.

First, when the personal information of the purchaser is received from the terminal device 100 connected to the service page through the receiving unit 240 and the recommendation commodity is requested, the generating unit 210 transmits the personal information of the purchaser to the purchaser (S210). ≪ / RTI >

At this time, the generating unit 210 assigns individual recommendation scores (Score) according to the self-reference for each recommendation algorithm to each of the recommendation products included in the recommendation product list, and assigns a priority rank .

When a plurality of recommended product lists based on the individual recommendation algorithm are generated, the generating unit 210 generates an individual recommendation score of the recommended product included in each recommended product list, for example, from 0 to 100 Normalization is performed.

Then, the collecting unit 250 collects the list of purchase items purchased by the past buyers based on the buyer's personal information (S220) to determine the fitness with the buyer for each of the recommendation algorithms.

Then, the determination unit 220 identifies the recommended products included in the purchased product list in each of the recommended product lists created on the basis of the past individual recommendation algorithm, thereby determining the fitness with the purchaser for each of the recommended algorithms (S230).

At this time, the determination unit 220 determines, for each of the recommendation algorithms, the number of all purchased products included in the purchased product list, the number of recommended products included in the purchased product list Which is a ratio between the buyer and the buyer.

Next, a final recommendation score for each of the recommended products is calculated by further applying the fitness with the purchaser to each of the individual recommendation scores of all the recommended products included in the plurality of recommended product lists S240).

At this time, when the calculation of the final recommendation score is completed, the selection unit 230 forms all the recommended products included in the plurality of recommended product lists with a priority order according to the size of the final recommendation score, and constitutes a list of all recommended products.

Then, the selecting unit 230 preferentially selects a recommendation commodity having a higher priority from the entire recommendation commodity list having a predetermined priority according to the size of the final recommendation score, as a recommendation commodity suitable for the purchaser, And provides the selected recommended product to the terminal device 100 (S250 - S260).

As described above, according to the product recommendation service method according to an embodiment of the present invention, among the various individual recommendation algorithms for determining a recommendation product, a recommendation product list is prepared Thus, the reliability of the recommended product can be increased, the satisfaction of the purchaser can be increased, and the profit of the seller of the product can be increased.

Meanwhile, the steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, or may be embodied in a computer readable medium, in the form of a program instruction, which may be carried out through various computer means. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

According to the commodity recommendation service system and method of the present invention, and the apparatuses applied thereto, the recommendation commodity suitable for a purchaser is determined through a combination of a plurality of algorithms for determining a recommendation commodity, It is not only the use of the related technology but also the possibility of commercialization or operation of the applied device is sufficient and it is practically possible to carry out it clearly.

100: terminal device
200: Service device
210: Generation unit 220:
230: selection unit 240:
250: collecting unit 260:

Claims (13)

A generating unit for generating a list of at least two recommended products for a buyer based on an individual recommendation algorithm;
A determination unit for determining a fitness of each of the individual recommendation algorithms with the purchaser; And
And a selecting unit for selecting and providing a recommended product suitable for the purchaser from at least one of the at least two recommended product lists based on the fitness with the purchaser for each of the individual recommendation algorithms, . The method according to claim 1,
The service device comprises:
And a collecting unit for collecting a list of purchased goods purchased by the purchaser,
Wherein,
And a recommendation commodity selection unit for recommending a recommendation commodity included in the purchase commodity list from the recommendation commodity list based on the individual recommendation algorithm to determine a goodness of fit with the purchaser for each recommendation algorithm. . 3. The method of claim 2,
Wherein,
For each of the recommendation algorithms, based on a ratio between the number of all purchased products included in the purchased product list and the ratio between the number of recommended products included in the purchased product list in each of the previously created recommended product lists, And determines the fitness of the service device. The method according to claim 1,
Each of the recommendation products included in the at least two recommendation product lists,
An individual recommendation score calculated by each individual recommendation algorithm is given,
Wherein the selection unit comprises:
Based on a final recommendation score calculated by further applying a goodness of fit with the purchaser to each of the individual recommendation algorithms to the individual recommendation scores of each of the recommendation products included in the at least two recommended product lists, And selects a recommended product suitable for the purchaser from at least one of the recommended product lists. 5. The method of claim 4,
Each of the recommendation products included in the at least two recommendation product lists,
A priority is set according to the final recommendation score size,
Wherein the selection unit comprises:
And selects a recommendation commodity having a higher priority among all recommendation commodities included in the at least two recommendation commodities as a recommendation commodity suitable for the purchaser. A requesting step of the terminal device requesting a recommendation item for the buyer;
A generating step of generating a list of at least two recommended products for the purchaser based on an individual recommendation algorithm of the service device; And
And a selection step of selecting a recommended product suitable for the buyer from at least one of the at least two recommended product lists based on the fitness with the buyer for each of the individual recommendation algorithms and providing the selected recommendation product to the terminal device How to recommend a service. A generating step of generating at least two recommended product lists for a buyer based on an individual recommendation algorithm;
A determination step of determining a fitness of each of the individual recommendation algorithms with the purchaser; And
And a selection step of selecting and providing a recommended product suitable for the buyer from at least one of the at least two recommended product lists based on the fitness with the buyer for each of the individual recommendation algorithms Method of operation of the device. 8. The method of claim 7,
The method comprises:
Further comprising a collecting step of collecting a list of purchased goods purchased by the purchaser before the determining step,
Wherein,
And a recommendation commodity selection unit for recommending a recommendation commodity included in the purchase commodity list from the recommendation commodity list based on the individual recommendation algorithm to determine a goodness of fit with the purchaser for each recommendation algorithm. Lt; / RTI > 9. The method of claim 8,
Wherein,
For each of the recommendation algorithms, based on a ratio between the number of all purchased products included in the purchased product list and the ratio between the number of recommended products included in the purchased product list in each of the previously created recommended product lists, And determining the fitness of the service device. 8. The method of claim 7,
Each of the recommendation products included in the at least two recommendation product lists,
An individual recommendation score calculated by each individual recommendation algorithm is given,
In the selecting step,
Based on a final recommendation score calculated by further applying a goodness of fit with the purchaser to each of the individual recommendation algorithms to the individual recommendation scores of each of the recommendation products included in the at least two recommended product lists, And selects a recommended product suitable for the purchaser from at least one of the recommended product catalogs. 11. The method of claim 10,
Each of the recommendation products included in the at least two recommendation product lists,
A priority is set according to the final recommendation score size,
In the selecting step,
And selects a recommendation commodity having a high priority among all recommendation commodities included in the at least two recommendation commodities as a recommendation commodity suitable for the purchaser. A generating step of generating at least two recommended product lists for a buyer based on an individual recommendation algorithm;
A determination step of determining a fitness of each of the individual recommendation algorithms with the purchaser; And
Selecting and providing a recommendation item suitable for the buyer from at least one of the at least two recommendation goods lists based on the fitness with the buyer for each of the individual recommendation algorithms, ≪ / RTI > A terminal device for requesting and receiving a recommended product for a buyer; And
Generating at least two recommended product lists for the buyer based on the individual recommendation algorithms and generating at least one recommended product list from at least one of the at least two recommended product lists based on the goodness of fit with the buyer for each of the individual recommendation algorithms, And a service device that selects a recommended product suitable for the user and provides the selected recommended product to the terminal device.

Images