KR101215301B1 - Goods or contents recommendation device and method - Google Patents

Abstract

An apparatus and method for recommending a product or content is provided. The product or content recommendation device is an input unit for receiving information about n (n is a natural number of two or more) specifications for selection, and is useful for determining a recommendation order based on information on the n (n is a natural number of two or more) specifications. It includes an operation unit for determining the value, and an output unit for outputting the recommendation rank. In addition, the product or content recommendation method is to receive the option, option specification, reference specification information from the input unit to calculate the utility value that is the basis of the recommendation ranking decision, and receives the utility value from the calculation unit to determine and output the recommendation rank It includes a step.

Description

Goods or contents recommendation device and method

The present invention relates to an apparatus and method for recommending goods or contents.

Recently, due to the development and industrialization of information technology, the importance of reading information about products or contents that consumers are looking for is emerging. In addition, due to the development of internet technology, internet shopping malls are being created exponentially, allowing consumers to shop on the web. When a consumer purchases a product or content through the shopping mall, a lot of time is required to search. A variety of search methods to find the exact is being studied.

In addition, due to the growth of multimedia contents and the development of ubiquitous technology, mobile terminals are connected to a network to exchange information, and the purchase of goods or contents through mobile terminals is increasing in a situation where a service suitable for a user's characteristics is provided. The importance of searching for is becoming even more important.

Existing methods for searching for desired products or contents include distance scale method and specification erase method. The distance metric method inputs a value desired by a user for each specification, and then obtains a distance from the specification value for each product or content and recommends a product or content having a minimum value. However, there is a disadvantage in that it is not easy for a customer to input a desired specification value as one value. In the case of specification-specific elimination method, if a range value of each specification is presented, a product which is out of the range is eliminated. Products are excluded from the target according to a specific condition. Therefore, a product or content included in the intersection of all conditions is searched. Increasingly, it is not easy to find products and contents that match the consumer's search, and there is a hassle of having to repeat the search from the beginning if you want to search by suggesting other conditions later.

An embodiment of the present invention enables to search for a product or content that meets a plurality of search conditions, and provides a product or content recommendation apparatus and method that can quickly search for a product or content desired by the consumer.

As an embodiment of the present invention, the product or content recommendation apparatus may include an input unit for receiving information about a specification for selecting a desired one, an operation unit for determining a utility value for determining a recommendation rank based on the specification information, and the recommendation order. It includes an output unit for outputting.

In an embodiment, the information about the specification may include n (n is a natural number of two or more) options, a selection order, reference specifications for n (n is a natural number of two or more) options, and the operation unit selects the selection. Based on the specification rank and the standard specification, the utility value that is the basis for determining the recommendation rank for m (m is a natural number of two or more) products or contents is calculated.

In an embodiment, the calculation unit is a weight calculation module that calculates a weight of n (n is a natural number of 2 or more) options according to the option specification order, and the m (m is a natural number of 2 or more) number based on the n reference specifications. Individual specification value calculation module for calculating individual specification value to goods or contents, normalization module for normalizing the individual specification value to calculate normal specification value, and utility value calculation for calculating utility value by calculating the normal specification value and the weight Contains modules

(여기서,

는 i번째 선택사양의 가중치, i=1, 2, …n,

, n은 선택사양의 개수)에 의해서 n(n은 2 이상의 자연수)개의 선택 사양에 가중치를 산정하며, 상기 정규화모듈은 상기 정규사양값이 커질수록 상기 기준사양과 가까워지는 경우에는

(여기서

는 i번째 선택사양에 j번째 상품 또는 콘텐츠의 정규사양값,

는 i번째 선택사양에 j번째 상품 또는 콘텐츠의 개별사양값,

는 i번째 선택사양의 최소 개별사양값,

는 i번째 선택사양의 최대 개별사양값)에 의해서 정규사양값을 산출하고, 상기 정규사양값이 작을수록 상기 기준사양과 가까워지는 경우에는

에 의해서 정규사양값을 산출하고, 상기 효용값 산정모듈은

(여기서

는 j번째 상품 또는 콘텐츠의 효용값,

는 i번째 선택사양의 가중치,

는 i번째 선택사양에 j번째 상품 또는 콘텐츠의 정규사양값)에 의해서 j번째 상품 또는 콘텐츠의 효용값을 연산한다. As an embodiment, the weight calculation module may be represented by an equation;

(here,

Is the weight of the i option, i = 1, 2,... n,

, n is the number of options, and n (n is a natural number of 2 or more). The weight is calculated by the number of options, and the normalization module is closer to the reference specification as the value of the normal specification increases.

(here

Is the canonical specification of the j th product or content,

Is the i-specific value of the j-th product or content,

Is the minimum individual value of the i option,

Is the maximum individual specification value of the i-th option), and when the smaller the normal specification value, the closer to the reference specification,

The normal specification value is calculated by the above, and the utility value calculation module

(here

Is the utility value of the j th product or content,

Is the weight of the i option,

Calculates the utility value of the j-th product or content by the i-th option with the j-th product or content regular specification value).

In an embodiment, when the normal specification value increases as the normal specification value approaches, the output unit determines and outputs a product or content order in which the utility value decreases as a recommendation order, and the smaller the regular specification value as the recommendation order. In the case of approaching the standard specification, the order of goods or contents in which the utility value becomes larger is output as a recommendation order.

In one embodiment of the present invention, the product or content recommendation method receives a selection specification, a selection specification ranking, and reference specification information from an input unit, and calculates a utility value that is a reference for determining a recommendation ranking for the product or content. And receiving the utility value to determine and output the recommendation order.

In an exemplary embodiment, the calculating of the utility value may include calculating a weight of n (n is a natural number of 2 or more) options according to the option rank, and m (m is 2 based on the n reference specifications. Setting individual specification values to the number of products or contents, normalizing the individual specification values to calculate a normal specification value, and calculating a utility value by calculating the normal specification value and the weight value. .

According to an embodiment of the present invention, by weighting a plurality of search conditions to determine a recommendation rank, the consumer may search for and recommend the one closest to a product or content that they want to purchase.

1 is a block diagram of an apparatus for recommending goods or contents according to an embodiment of the present invention.
2 is a graph showing normal specification values when a range of a prior reference specification is not entered.
3 is a graph showing a normal specification value when a range of a preliminary reference specification is input.
Figure 4 is a graph showing the actual experimental results according to the prior art and one embodiment of the present invention.
5 is a flow chart of a product or content recommendation method according to an embodiment of the present invention.

The embodiments may be embodied in many different forms and should not be construed as limited to the aspects set forth herein. Rather, the above aspects make the embodiments more thorough and complete, and fully convey the scope of the embodiments to those skilled in the art.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

An apparatus for recommending goods or contents according to an embodiment of the present invention is an invention related to an apparatus and a method for receiving information on an optional specification and determining a recommendation rank using the optional specification information.

1 is a block diagram of a product or content recommendation apparatus 1000 according to an embodiment of the present invention.

As illustrated in FIG. 1, the product or content recommendation apparatus 1000 includes an input unit 100, a calculation unit 200, and an output unit 300.

By using the information transmitted from the input unit 100, the calculation unit 200 determines a recommendation order of a product or content that a consumer wants to purchase, and outputs it from the output unit 300.

Hereinafter, an embodiment of the present invention will be described in more detail.

The input unit 100 receives an option specification, an option specification ranking, and a reference specification. More specifically, the input unit 100 receives n (n is a natural number of two or more) options, n (n is a natural number of two or more) options, and n (n is a natural number of two or more). Enter the standard specifications for the dog's options.

When a consumer selects a product or content, he considers the specifications that are the criteria. For example, when choosing a product, the price range, manufacturer, release date, and product features are typical specifications that consumers consider when purchasing the product. In the product selection specification, when a category of a product selected by a consumer is input, information stored in an external device (not shown) is transmitted to the input unit. As another example, when selecting contents related to job search, salary, working place, working hours, etc. are typical specifications, and information about the job is transmitted to the input unit.

Option priority refers to priority information about a specification that consumers consider when selecting a product or content. According to the above example of product selection, the price range of a manufacturer, a consumer that takes precedence over a release date, and a consumer that takes priority over a function of a product are determined and input by a consumer who wants to purchase a product. In addition, in order to select the desired content for job search, the job seeker selects an annual salary prior to the work place, and then ranks and inputs the work place and the working hours in order.

The reference specification value for the option of the product may be information about the specification of the product that the consumer wants to purchase with respect to the specification considered when the consumer selects the product. In other words, the customer selects 200,000 ~ 300,000 won from the optional price range, and the release date is based on the customer's reference value that he wants to search for new products released from January 2011 to March 2011. The standard specification can be entered in a range such as a price range, a release date, etc., and a desired specification such as a manufacturer and a product function can be specified.

As in the previous example, the standard specification value for the option of the content is to enter a specific range that the job seeker wants, such as a desired salary of 20 million won to 30 million won, working place Seoul, Gyeonggi, 8 hours to 10 hours.

As described above, the input unit 100 receives an option specification, an option specification rank, and a reference specification.

The calculation unit 200 calculates a utility value which is a reference for determining a recommendation ranking for a product or content based on the optional specification ranking and the reference specification.

Hereinafter, the operation unit 200 will be described in more detail.

The calculation unit 200 is a weight calculation module 210 for calculating a weight of a plurality of options according to the option specification order, an individual specification value for calculating individual specification values of various goods or contents by a plurality of reference specifications transmitted from the input unit. Calculation module 220, the normalization module 230 to normalize the individual specification values to calculate the normal specification value, the utility value calculation module 240 to calculate the utility value by calculating the normal specification value and the weight calculated in the weighting module previously ).

라 하면 소비자가 입력한 선택사양순위는

≥

≥…≥

≥…≥

이 된다. 여기서 소비자가 선택하는 순위는 가중치값의 선후를 결정하는 것이고, 개별 선택사양의 가중치 값에 대해서는 수학식 1에 의해서 계산되어진다. First, the weight calculation module 210 will be described. For example, the weight calculation module 210 has n (n is a natural number of two or more) options based on the information on the option specification ranking input by the consumer in the input unit 100. To be weighted. Where the weighted value of the i th option out of n options

So that the option rank entered by the consumer is

≥

≥… ≥

≥… ≥

. In this case, the rank that the consumer selects is to determine the weight of the individual value, and the weight value of the individual option is calculated by Equation 1.

, n은 선택사양의 개수)(Where i = 1, 2,… n,

, n is the number of options)

)은 상기 수학식 1에 의하면

=

이 된다. The weight value of the first option that consumers consider the most important of the n options (

) According to Equation 1

=

.

=

이 된다. Also

=

.

The weight calculation module 210 calculates a weight value for each option that the consumer has ranked according to Equation (1).

Individual specification value calculation module 220 sets the individual specification value for each product or content in the reference specification.

We will look at how to calculate individual specification values through specific examples.

The embodiment described below is a case where a consumer selects a product, and is not necessarily limited to recommendation ranking for selecting a product.

model name manufacturer price type Memory A1 L company 228,660 slim 1 GB-E A2 Company S 320,000 Dual Folder 120 MB A3 Company S 258,720 slim 120 MB A4 L company 222,420 slide 60 MB A5 K company 320,460 slide 60 MB A6 Company S 294,000 slide 60 MB A7 L company 294,000 slim 1 GB-N A8 Company S 299,000 slide 120 MB

A consumer who wants to purchase a mobile phone enters the standard specifications for each option for the four options: manufacturer, price, type, and memory.

If the standard specifications for the manufacturer entered by the consumer are S or L, the price is 200,000 ~ 300,000 won, the type is slim or slide, and the memory is more than 60MB. Numeric options, such as specifications and prices, will coexist.

If the option shown as a category value is set to a high value close to the reference specification, the manufacturer can be set to 3 points for S company, 2 points for L company, 1 point for K company. Even if it is an option type or a memory, it is set to have a high value when it is close to the standard specification.

As a result, if the products have the individual specification value for each specification through the individual specification value calculation module 220 can be represented as shown in Table 2.

model name manufacturer price type Memory A1 L company (2) 228,660 Slim (2) 1 GB-E (5) A2 S company (3) 320,000 Dual Folder (1) 120 MB (2) A3 S company (3) 258,720 Slim (2) 120 MB (3) A4 L company (2) 222,420 Slide (3) 60 MB (1) A5 K company (1) 320,460 Slide (3) 60 MB (1) A6 S company (3) 294,000 Slide (3) 60 MB (1) A7 L company (2) 294,000 Slim (2) 1 GB-N (4) A8 S company (3) 299,000 Slide (3) 120 MB (3)

Of course, the price of the options is shown as a number, so the price value itself is an individual specification.

After individual specification values are determined for each model, differences between individual specification values exist due to differences in each option. That is, in case of optional manufacturer, individual specification values are distributed in 1 ~ 3, and in case of optional price, individual specification values are distributed in 222,420 ~ 320,000, so it is necessary to unify individual differences among these values.

The normalization module 230 normalizes individual specification values to calculate a normal specification value.

The normalization module 230 normalizes according to Equation 2 below when the normal specification value becomes larger and closer to the reference specification.

는 i번째 선택사양에 j번째 상품의 정규사양값,

는 i번째 선택사양에 j번째 상품의 개별사양값,

는 i번째 선택사양의 최소 개별사양값,

는 i번째 선택사양의 최대 개별사양값)(here

Is the regular specification value of the j item in the i option,

Is the value of the individual specification of the j item to the i option,

Is the minimum individual value of the i option,

Is the maximum individual value of the i option)

However, among the options, the higher the value, for example, the number of repairs, recalls, etc. of the product, there is a kind of option that is far from the standard specification desired by the consumer.

As the normal specification value becomes smaller as described above, the closer to the reference specification, the normalization is performed using Equation 3 below.

는 i번째 선택사양에 j번째 상품의 정규사양값,

는 i번째 선택사양에 j번째 상품의 개별사양값,

는 i번째 선택사양의 최소 개별사양값,

는 i번째 선택사양의 최대 개별사양값)(here

Is the regular specification value of the j item in the i option,

Is the value of the individual specification of the j item to the i option,

Is the minimum individual value of the i option,

Is the maximum individual value of the i option)

As in the above-described embodiment of the present invention, a case in which the reference specification is specified in a range desired by the consumer is compared with the case where it is not.

FIG. 2 is a graph showing normal specification values when a range of a prior reference specification is not input. FIG. 3 is a graph showing a normal specification value when a range of a prior reference specification is entered.

As shown in FIG. 2, when the range of the standard specification is not entered, the consumer searches the product without limitation, and the max (maximum point) displayed on the horizontal axis is the normal specification of the product desired by the actual consumer. It is not a value. That is, distortion may occur in values that are normalized to calculate the utility value.

In contrast, when a range of reference specifications is input as shown in FIG. 3, a consumer searches for a product.

은 최대값을 갖게 되므로 정규사양값이 선택사양내에서 소비자가 원하는 원하는 사양에 대한 의사를 반영할 수 있게 된다. The consumer enters a predetermined range of values for the standard specifications and is displayed on the horizontal axis.

Has a maximum value so that the normal specification value can reflect the intent of the desired specification within the option.

Calculate the normal specification value using the example shown above.

The standard specification selected by the consumer was S or L for the manufacturer entered by the consumer, the price was 200,000 to 300,000 won, the type was slim or slide, and the memory was more than 60MB. Except for A2 and A5, which do not satisfy the price range, as shown in Table 3, the selected models are A1, A3, A4, A6, A7, and A8. For each, the value of the normal specification calculated by the normalization module is shown in parentheses in Table 3.

model name manufacturer price type Memory A1 L company (0) 228,660 (0.92) Slim (0) 1 GB-E (1) A3 S company (1) 258,720 (0.53) Slim (0) 120 MB (0.5) A4 L company (0) 222,420 (1) Slide (1) 60 MB (0) A6 S company (1) 294,000 (0.07) Slide (1) 60 MB (0) A7 L company (0) 294,000 (0.07) Slim (0) 1 GB-N (0.75) A8 S company (1) 299,000 (0) Slide (1) 120 MB (0.5)

Let's take a look at the utility value calculation module.

The utility value calculation module is represented by Equation 4

는 j번째 상품의 효용값,

는 i번째 선택사양의 가중치,

는 i번째 선택사양에 j번째 상품의 정규사양값)에 의해서 j번째 상품의 효용값을 연산한다.(here

Is the utility value of the j th product,

Is the weight of the i option,

Calculates the utility value of the j th product by the i th option).

Table 4 expresses both the weight value for each option and the normal specification value of each product by Equation 1.

The weight value for each option is indicated in parentheses of the option, and the normal value is indicated in parentheses for the option for each product, as shown in Table 3 above.

model name

=0.52)manufacturer
(

= 0.52) price
(

= 0.27) type
(

= 0.15) Memory
(

= 0.06) A1 L company (0) 228,660 (0.92) Slim (0) 1 GB-E (1) A3 S company (1) 258,720 (0.53) Slim (0) 120 MB (0.5) A4 L company (0) 222,420 (1) Slide (1) 60 MB (0) A6 S company (1) 294,000 (0.07) Slide (1) 60 MB (0) A7 L company (0) 294,000 (0.07) Slim (0) 1 GB-N (0.75) A8 S company (1) 299,000 (0) Slide (1) 120 MB (0.5)

Calculating the utility value for each model by the equation (4)

=(0.52)(0)+(0.27)(0.92)+(0.15)(0)+(0.06)(1)=0.31Utility value of A1

= (0.52) (0) + (0.27) (0.92) + (0.15) (0) + (0.06) (1) = 0.31

=(0.52)(1)+(0.27)(0.53)+(0.15)(0)+(0.06)(0.5)=0.693Utility value of A3

= (0.52) (1) + (0.27) (0.53) + (0.15) (0) + (0.06) (0.5) = 0.693

=(0.52)(0)+(0.27)(1)+(0.15)(1)+(0.06)(0)=0.42Utility value of A4

= (0.52) (0) + (0.27) (1) + (0.15) (1) + (0.06) (0) = 0.42

=(0.52)(1)+(0.27)(0.07)+(0.15)(1)+(0.06)(0)=0.6889Utility value of A6

= (0.52) (1) + (0.27) (0.07) + (0.15) (1) + (0.06) (0) = 0.6889

=(0.52)(0)+(0.27)(0.07)+(0.15)(0)+(0.06)(0.75)=0.06Utility value of A7

= (0.52) (0) + (0.27) (0.07) + (0.15) (0) + (0.06) (0.75) = 0.06

=(0.52)(1)+(0.27)(0)+(0.15)(1)+(0.06)(0.5)=0.7Utility value of A8

= (0.52) (1) + (0.27) (0) + (0.15) (1) + (0.06) (0.5) = 0.7

.

As a result, the utility value is in the order of A8> A3> A6> A4> A1> A7 according to the standard specifications suggested by the consumer.

The output unit 300 determines and outputs the recommendation rank according to the magnitude of the utility value determined by the weight calculation module, the individual specification value calculation module, the normalization module, and the utility value calculation module of the calculation unit.

Figure 4 is a graph showing the actual experimental results according to the prior art and one embodiment of the present invention.

In order to compare the method of determining the recommendation order (TD) by the conventional distance method and the method to which the present invention is applied, the reference specification is determined for 60 test subjects, and the recommendation according to the reference specification is the prior art distance method (TD). And by the present invention. The horizontal axis represents the number of products to be recommended, and the vertical axis represents accuracy of expressing whether the top recommendation product of the determined recommendation rank and the 60 people who participated in the experiment are the same as the actual desired product.

As a result, according to the conventional distance method (TD), the horizontal axis indicating the number of participating products tends to coincide with the desired products and recommended products by 60 test participants only when 7 products are to be recommended. Even though five to five products were recommended, the products that 60 test participants wanted and the recommended products matched.

5 is a flow chart of a product or content recommendation method according to an embodiment of the present invention.

As shown in FIG. 5, an input unit receives a type of a product or content (S10), and an input unit receives n (n is a natural number of 2 or more) options, a ranking of options, and a reference specification of the options. In operation S20, the operation unit receives the option, the option specification, and the reference specification information from the input unit, and calculates a utility value that is a reference for determining the recommendation order for m (m is a natural number of 2 or more) products or contents. . The calculating of the utility value may include calculating weights of n (n is a natural number of 2 or more) options according to the option ranking (S30), and m (m is a natural number of 2 or more based on the input n reference specifications). Step S40 of setting individual specification values for the goods or contents (S40), normalizing individual specification values to calculate a normal specification value (S50), and calculating a utility value by calculating a normal specification value and a weight (S60). The output unit receives the utility value from the operation unit to determine a recommendation order (S70) and outputs (S80).

The consumer inputs the type of the product or the content that he / she wants to purchase. (S10) For example, in the case of the product, various items that can be purchased through the Internet, such as mobile phones and MP3s, as well as sneakers and hats, are included in the method of the present invention. Application is possible. Once the type of product that the consumer wants to purchase is entered, the specifications that are the standard for product selection by the selected product type are listed. Information on the options can be provided through a storage medium such as a database. Provide information.

The input unit receives inputs of options, option rankings, and reference specifications. (S20) A consumer may select a desired option from among a plurality of options, and among the selected options, which is important for the consumer to select a product or content. The order of options is entered in order of specification, and the standard specification, that is, the desired specification of each option is specified. Reference specifications may be specified as ranges.

,

,

…

을 계산하는 단계를 의미한다. 즉 각 선택사양에 대한 순위를 통해서 가중치 값을 계산한다. 여기서 n은 선택사양이 n개인 경우를 표시한 것이다. Calculating the weight on the option (S30) through the above-described Equation 1

,

,

...

It means the step of calculating. In other words, the weight value is calculated by rank for each option. Where n represents the case of n options.

The step of setting individual specification values of goods or contents (S40) refers to a reference specification and a specification and difference value of each product and content.

In the step of determining the normal specification value (S50), the individual specification value set in the previous step S40 is a constant value expressed as a number, and the relationship with the reference specification can be expressed uniformly through normalization with values of a plurality of options. Will be. The calculation of the normal specification value is made through Equations 2 and 3 described above.

When the weight value and the normal specification value are calculated, the utility value is determined so that a plurality of actual products or contents can be arranged in the order close to the reference specification. (S60) The utility value is calculated by Equation 4.

After that, the output unit determines recommendation ranks by ordering utility values of a plurality of products (S70), and outputs recommendation ranks of products or contents (S80).

The scope of the present invention is not limited to the above-described embodiments but may be implemented in various forms of embodiments within the scope of the appended claims. Without departing from the gist of the invention claimed in the claims, it is intended that any person skilled in the art to which the present invention pertains falls within the scope of the claims described in the present invention to various extents which can be modified.

100 Input Unit 200 Operation Unit
210 Weighting module 220 Individual specification value module
230 Normalization module 240 Utility value calculation module
300 outputs

Claims (9)

An input unit for receiving a selection specification as a comparison criterion for selection of a product or content, a selection specification ranking indicating a preference order for the selection specification, and a reference specification indicating a criterion of the selection specification;
Calculating a weight value inversely proportional to the option specification for the option, calculating an individual specification value for the option based on the reference specification, and based on the normal specification value and the weight normalizing the individual specification value An operation unit determining a utility value for determining a recommendation order of a product or content; and
Product or content recommendation apparatus comprising an output unit for outputting the recommendation ranking.
delete delete The method of claim 1,
The calculation unit further includes a weight calculation module for calculating the weight,
The weight calculation module is an equation;

(here,

Is the weight of the i option, i = 1, 2,... n,

n is the number of options) product or content recommendation apparatus characterized in that the weighted to the optional specification.
The method of claim 1,
The calculation unit further includes a normalization module for calculating the normal specification value,
The normalization module is closer to the reference specification as the normal specification value increases.

(here

Is the canonical specification of the j th product or content,

Is the i-specific value of the j-th product or content,

Is the minimum individual value of the i option,

Is the maximum individual specification value of the i-th option), and when the smaller the normal specification value, the closer to the reference specification,

A product or content recommendation device, characterized in that for calculating a normal specification value.
The method of claim 1,
The calculation unit further includes a utility value calculation module for calculating the utility value,
The utility value calculation module

(here

Is the utility value of the j th product or content,

Is the weight of the i option,

Is a regular value of the j-th product or content in the i-th option), the utility value of the j-th product or content, characterized in that for calculating the utility value.
The method of claim 1,
The output unit determines and outputs a product or content order in which the utility value decreases as the recommendation order when the normal specification value increases and approaches the reference specification. As the normal specification value decreases, the output specification approaches the reference specification. The product or content recommendation device, characterized in that for outputting the product or content order in which the utility value increases in the order of recommendation.
Receiving, by the operation unit, a selection specification that is a comparison criterion for selection of a product or a content, an selection specification ranking indicating a preference order for the selection specification, and a reference specification indicating a criterion of the selection specification;
The calculation unit calculates a weight that is inversely proportional to the option specification for the option, calculates an individual specification value for the option based on the reference specification, and normal specification value and the weight that normalize the individual specification value. Calculating a utility value for determining a recommendation order of the product or the content based on the
And an output unit receiving the utility value from the operation unit and outputting a recommendation order. delete

Images