KR20200112495A - Server and method for providing commercial analysis services - Google Patents

Abstract

An objective of the present invention is to efficiently improve sales of a business. According to one embodiment of the present invention, a server for providing a business district analysis service comprises: a memory storing a program for providing a business district analysis service; and a processor to execute the program stored in the memory. The processor receives sales information including a business type, a location, a date, and sales from a business terminal, and calculates predicted sales of the business by using a fuzzy system based on the sales information, as the program is executed. The fuzzy system is modeled based on a cluster center wherein previously stored past sales information including a business type, a location, a date, and sales is clustered by a subtractive clustering algorithm. The subtractive clustering algorithm normalizes the past sales information to calculate the distance between data, and calculates the cluster center based on the size of the distance potential value between the normalized data.

Description

Commercial area analysis service providing server and its method {SERVER AND METHOD FOR PROVIDING COMMERCIAL ANALYSIS SERVICES}

The present invention relates to a commercial area analysis service providing server and method thereof, and provides it to a business operator by accurately grasping commercial area related information such as expected sales based on business information such as location, sales, and date of a business place.

With the recent increase in the minimum wage and the recession of the international economy, the economy of small businesses, which is the basis of the common people's economy, is deteriorating. In particular, in the current domestic economy where the gap between the rich and the poor is widening, the economic gap between regions is also increasing. Therefore, it is necessary to introduce a location-based local small business platform that can revitalize and develop the small business and the local economy.

Also, for small business owners, the risk of sales is one of the important issues. It is possible to reduce the risk of small business owners operating a store by checking the past sales according to the business district such as the type of business and location they intend to do and predicting the future sales of the business site. This allows you to make planned expenditures and analyze the parts that affect sales.

However, at the present time, there is no suitable solution for predicting sales of the business site, and even predicting sales is not accurate, so it does not greatly help in business operation.

Republic of Korea Patent Publication No. 10-2007-0076884 (title of the invention: sales forecasting system through sensory evaluation and sales forecasting method of the forecasting system)

The server for providing a commercial area analysis service and its method according to the present invention is intended to efficiently improve the business of the business operator by providing predicted sales information to the business owner by analyzing the business area and by providing data analyzing various factors affecting the sales.

The server for providing a commercial area analysis service and its method according to the present invention aims to reduce the risk when starting a business by predicting in a time series how much future sales will be expected in the commercial area where small business owners intend to operate.

A server providing a commercial area analysis service according to an embodiment of the present invention, comprising: a memory storing a program providing a commercial area analysis service; And a processor for executing a program stored in the memory, wherein the processor receives sales information including business type, location, date and sales from the business operator terminal according to the execution of the program, and based on the sales information, a fuzzy system system), and the fuzzy system focuses on clusters in which past sales information including pre-stored business type, location, date, and sales are clustered through a subtractive clustering algorithm. It is modeled on the basis, and the subtractive clustering algorithm normalizes the past sales information to calculate the distance between data, and calculates the cluster center based on the size of the distance potential value between the normalized data. Want to provide.

In the commercial area analysis service providing method performed by the commercial area analysis service providing server according to an embodiment of the present invention, the method comprising: receiving business information including a business type, location, date, and sales from a business operator terminal; And calculating the expected sales of the business site using a fuzzy system based on the business information, and the fuzzy system includes the previously stored business information including the business type, location, date and sales of the business site. It is modeled based on the cluster center through a subtractive clustering algorithm, and the subtractive clustering algorithm normalizes past sales information to calculate the distance between data, and calculates the size of the distance potential between normalized data. It is intended to provide a method of providing a commercial area analysis service that calculates the cluster center on the basis.

The server and method for providing a commercial area analysis service according to an embodiment of the present invention provide predicted sales information to the business owner by analyzing the business area, and provide data that analyzes various factors affecting sales to efficiently improve the business of the business operator. I can make it.

The server and method for providing a commercial area analysis service according to an embodiment of the present invention can reduce the risk when small business owners start a business by providing a time-series prediction of how much future expected sales will be in a commercial area where small business owners intend to do business.

1 is a diagram showing the configuration of a commercial area analysis service providing server according to an embodiment of the present invention.
2 is a flowchart illustrating a process of a method for providing a commercial area analysis service according to an embodiment of the present invention.
3 is a flowchart illustrating a process of a subtractive clustering method according to an embodiment of the present invention.
4 is a diagram showing the configuration of a purge system according to an embodiment of the present invention.
5 is a flowchart illustrating an example of a process of calculating expected sales according to an embodiment of the present invention.
6 is a diagram showing a graph of predicted sales statistics of a business operator terminal program according to an embodiment of the present invention.
7 is a diagram illustrating a prediction error according to the number of fuzzy rules in a fuzzy system according to an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art may easily implement the present invention. However, the present invention may be implemented in various different forms and is not limited to the embodiments described herein. Further, in the drawings, parts not related to the description are omitted in order to clearly describe the present invention, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" with another part, this includes not only "directly connected" but also "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, it means that other components may be further included rather than excluding other components unless specifically stated to the contrary.

Hereinafter, a commercial area analysis service providing server and a method according to the present invention will be described with reference to the accompanying drawings.

1 is a diagram showing the configuration of a commercial area analysis service providing server according to an embodiment of the present invention.

Referring to FIG. 1, a commercial area analysis service providing server 100 according to an embodiment of the present invention is a server for providing a commercial area analysis result to a business area based on sales information provided from the business operator terminal in communication with a business operator terminal. It is about.

The commercial area analysis service providing server 100 may include a communication module 110, a memory 120, a processor 130, and a database (DB: 140).

The communication module 110 provides a communication interface to the commercial area analysis service providing server 100 by interworking with a communication network, and may perform a role of transmitting and receiving data to and from a business operator terminal. Here, the communication module 110 may be a device including hardware and software necessary for transmitting and receiving a signal such as a control signal or a data signal through a wired or wireless connection with another network device.

In the memory 120, a program for providing a commercial area analysis service may be recorded. In addition, the memory 120 may perform a function of temporarily or permanently storing data processed by the processor 130. Here, the memory 120 may include a volatile storage medium or a non-volatile storage medium, but the scope of the present invention is not limited thereto.

The processor 130 may control an entire process performed by a program for providing a commercial area analysis service in the commercial area analysis service providing server 100. Each step of the process performed by the processor 130 will be described later with reference to FIGS. 2 to 7.

Here, the processor 130 may include all types of devices capable of processing data, such as a processor. Here, the'processor' may refer to a data processing device embedded in hardware having a circuit physically structured to perform a function represented by a code or instruction included in a program. As an example of a data processing device built into the hardware as described above, a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, and an application-specific integrated (ASIC) circuit) and processing devices such as field programmable gate arrays (FPGAs), but the scope of the present invention is not limited thereto.

In the database 140, information for providing a commercial area analysis service may be stored. For example, sales information including a location of a business place, type of business, advertisement cost, sales, date, etc. to be described later may be stored.

Hereinafter, a method of providing the commercial area analysis service providing server 100 according to an embodiment of the present invention will be described in detail.

Referring to FIG. 2, first, the commercial area analysis service providing server 100 may perform a step S200 of receiving sales information including a business type, location, date, and sales from a business operator terminal. Here, the sales information is information generated while a business operator operates a business site, and may further include advertisement cost, advertisement efficiency (sales/advertisement cost * 100%), and the like.

Step S200 Next, the commercial area analysis service providing server 100 may perform a step S210 of calculating the expected sales of the business site using a fuzzy system based on the business information.

Here, the fuzzy system may be modeled based on a cluster center in which past sales information including previously stored past business types, locations, dates, and sales are clustered through a subtractive clustering algorithm. The past business information is information generated as a result of the operation of a number of business sites in the corresponding commercial area, and may further include advertising costs, advertising efficiency (sales/advertising costs * 100%), and the like.

The subtractive clustering algorithm normalizes past sales information, calculates a distance between data, and calculates a cluster center based on the size of a distance potential value between normalized data. Preferably, the subtractive clustering algorithm finds the center of the cluster with the largest size of the distance potential between normalized data as the center of the first cluster, removes the influence of the center of the first cluster, and then the size of the distance potential between the data is the largest or , Even if the size of the potential value is small, a process of obtaining a cluster center having a large distance between cluster centers as a second cluster center may be repeated to obtain a plurality of cluster centers.

3 is a flowchart illustrating a process of a subtractive clustering method according to an embodiment of the present invention.

Hereinafter, a subtractive clustering algorithm will be described in detail with reference to FIG. 3.

When input/output data is given to a fuzzy system, the data clustering algorithm, which is one of the methods of modeling a fuzzy system, is a technique that optimizes the fuzzy system to minimize the objective function, using the K-means algorithm or FCM (Fuzzy It is based on the C-Means) algorithm. The K-means algorithm is based on minimization of a performance index expressed as the sum of the squared distances from all data points in the cluster range to the cluster center, and the FCM algorithm Based on the distance to and based on the similarity measurement, an optimization method of the objective function is used.

Another clustering algorithm is the mountain clustering method. The acid clustering method divides a data area into grid lines and calculates a potential value between a data value and an intersection point of the grid line. The potential value is given as a function of distance, and the intersection of the grid line having the largest value among the calculated potential values becomes the first cluster center. The second cluster center is selected by removing the influence of the first cluster center to prevent the second cluster center from being determined near the first cluster center, and then obtaining the second cluster center. This process is repeated until the potential value reaches a constant value. However, this method has a problem that the amount of computation exponentially increases as the dimension increases.

Accordingly, in the present invention, in order to compensate for the above problem, a subtractive clustering algorithm, which is a modified algorithm that obtains the cluster center from a given data, is used without dividing the data area with a grid line.

In the subtractive clustering algorithm, the point with the maximum potential value given as a function of the distance between data is selected as the center of the first cluster, and the data with the maximum potential value becomes the center of the next cluster with the influence of the center of the first cluster removed. This process is repeated until the potential value reaches a certain criterion.

가 주어졌을 때, 섭트랙티브 클러스터링 알고리즘은 다음과 같은 순서로 이루어진다. 첫째, 주어진 데이터를 [0,1]로 정규화 한다(S300). 둘째, 데이터 간의 거리를 구한다. 즉, 각 데이터 간의 거리는 하기 [수학식1]과 같이 주어지는 함수 P로 표현되며 P_i를 포텐셜 값이라 한다. 그리고, 데이터 간의 거리 포텐셜 값이 가장 큰 k(1≤k≤N)번째 클러스터 중심을 산출한다(S310).N data in the (n+1)-dimensional input/output space where the input is n-dimensional and the output is a 1-dimensional space

When is given, the subtractive clustering algorithm is performed in the following order. First, the given data is normalized to [0,1] (S300). Second, find the distance between data. That is, the distance between each data is expressed by a function P given as in [Equation 1] below, and P _i is called a potential value. Then, the k (1≦k≦N)-th cluster center having the largest distance potential between data is calculated (S310).

[Equation 1]

P _i = distance potential value between data

N = number of input data

X _i , X _j = data value

로 주어지며, r_a는 양의 상수로서 r_a밖의 데이터는 포텐셜 값에는 영향을 거의 주지 못하게 된다. 실험을 통해 r_a가 0.3일 때 최적화된 결과를 보여주었다. 즉, r_a가 0.3 일 때, 이를 기초로 생성된 퍼지 규칙을 통해 알고리즘을 수행한 결과 오차율이 최소로 출력 되었다.Where α is

It is given as, and r _a is a positive constant, and data outside of r _a hardly affects the potential value. Through the experiment, the optimized result was shown when r _a was 0.3. That is, when r _a is 0.3, the error rate is output to the minimum as a result of performing the algorithm through the fuzzy rule generated based on this.

Specifically, after normalizing the data, the center of the first cluster having the largest distance potential value between the data is obtained. The largest of the N potential values is P ₁ ^* , and the data at this time becomes the center of the first cluster X ₁ ^* . Next, the influence of the center of the first cluster is removed. Since there is a lot of data near the center of the first cluster, if the influence is not removed, the center of the second cluster is likely to occur near the center of the first cluster. The removed potential value P _i 'is calculated.

[Equation 2]

로 주어지며, r_b는 양의 상수로서 r_a보다 큰 값을 취해 클러스터 중심 근처에 다음 클러스터 중심이 나타나지 않도록 한다. 실험을 통해 r_b값의 설정은 1.5r_a 로 하였을 때 좋은 결과를 보여주었다. 즉, r_b가 1.5r_a일 때, 이를 기초로 생성된 퍼지 규칙을 통해 알고리즘을 수행한 결과 오차율이 최소로 출력 되었다.Here, P _i is a value calculated by Equation [Equation 1]. Also, β is

Is given by, and r _b is a positive constant and takes a value greater than r _a so that the next cluster center does not appear near the cluster center. Through the experiment, the setting of the r _b value showed good results when set to 1.5 r _a . That is, when r _b is 1.5 r _a , the error rate is output to the minimum as a result of performing the algorithm through the fuzzy rule generated based on this.

Thereby, removal of the influence of the first cluster center N of potential value P _i 'the highest place la P ₂ ^* in the data of this time is a second cluster center X ₂ ^* can obtain the second cluster center .

When the k-th cluster center X _k ^* is obtained by repeating this process, the potential value from which the influence of the k-th cluster center has been removed can be expressed by the following [Equation 3], and the (k+1)-th cluster center can be obtained. Yes (S320).

[Equation 3]

P _i ^' = potential value with the effect of the k-th cluster center removed

P _i = distance potential value between data

P _k ^* = potential value of the k-th cluster center

X _i = data

, (r_b=양의 상수)β =

, (r _b = positive constant)

을 클러스터 중심들

중에서 상호간의 가장 짧은 거리로 정의하였을 때,

이면 클러스터 중심으로 채택하고 위의 과정을 반복한다.

이고

이면 클러스터 중심으로 채택하고,

이고

이면 그 때의

를 0으로 하고 다음으로 높은 포텐셜 값을 선정한 후,

이면 새로운 클러스터 중심으로 채택한다.

이면 위의 과정의 반복을 중단한다. Next,

Cluster centers

Among them, when defined as the shortest distance between each other,

In this case, it is adopted as the cluster center and the above process is repeated.

ego

Adopt the back cluster center,

ego

Back then

Is set to 0 and the next highest potential value is selected,

If so, it is adopted as the center of the new cluster.

If so, the repetition of the above process is stopped.

와

은 클러스터 중심 선정의 상하 기준이 되는 상수로서 0부터 1사이의 값이 된다. 즉, P_k ^*를 P₁ ^* 으로 나누어서 첫 번째 포텐셜 값의 크기를 1로 만들었으며, 또한 클러스터 중심 선정의 상하 기준(

,

)을 둠으로써 P_k ^*/P₁ ^* 이

보다 크면 클러스터 중심으로 선정하고

보다 작으면 클러스터 중심으로 선정하지 않으며 그 사이일 경우에는 조건부로 클러스터 중심으로 선정하게 된다(S330). 이러한 조건부를 선정 기준에 넣음으로써 포텐셜 값이 작더라도 클러스터 중심 간의 거리(

)가 크면 클러스터 중심으로 선정되게 된다.At this time,

Wow

Is a constant that is the upper and lower standard of cluster center selection and is a value between 0 and 1. That is, P _k ^* is divided by P ₁ ^* to make the size of the first potential value 1, and the upper and lower criteria for cluster center selection (

,

), P _k ^* /P ₁ ^*

If it is larger, it is selected as the cluster center

If it is smaller than, the cluster center is not selected, and in the case of the interval, the cluster center is conditionally selected (S330). By putting these conditionals into the selection criterion, the distance between cluster centers (

If) is large, it is selected as the cluster center.

을 결정할 수 있으며, 결정된 클러스터 중심을 기초로 하여 퍼지 시스템을 모델링할 수 있는 것이다.When input/output data is given to the fuzzy system through the above process, the input/output space is centered on M clusters with (n+1) dimensions.

Can be determined, and a fuzzy system can be modeled based on the determined cluster center.

4 is a diagram showing the configuration of a purge system according to an embodiment of the present invention.

Hereinafter, referring to FIG. 4, a fuzzy system having M rules is formed from the process of determining the centers of M clusters having (n+1) dimensions by clustering N data given in n-dimensional input and 1-dimensional output. Let's look at the modeling process.

Fuzzy system modeling may be performed based on a Momdani fuzzy model or a Takagi-Sugeno-Kang (T-S-K) fuzzy model. In the case of modeling with the mamdani fuzzy algorithm, the modeling process is divided into a process of dividing the cluster center into input/output spaces and a process of mapping with fuzzy rules. In the case of modeling with the T-S-K fuzzy algorithm, the process of optimizing the parameters of the post-construction rule is added to the process of mapping with the fuzzy rule.

A key feature of a fuzzy system is the concept of fuzzy partitioning of given information, which operates based on fuzzy sets instead of a crisis number. Fuzzy sets contain more information than simple numbers. Is done. Therefore, generalization of information based on inaccuracy can be achieved by using fuzzy sets.

The design process of a fuzzy system begins by first determining input/output variables, dividing the space of each variable, and allocating an appropriate number of language labels. When language labels for input/output variables are determined, fuzzy rules for expressing expert knowledge are defined using them, and membership functions are mapped to each language label to form a fuzzy set. In addition, it is necessary to determine the method of fuzzification, defuzzification, and reasoning.

Referring to FIG. 4, it can be seen that the fuzzy system is composed of fuzzification, rule base, fuzzy inference, and defuzzification.

Fuzzification plays an important role in dealing with ambiguity and uncertain information in natural language. In particular, since the input/output data of the autonomous mobile robot is always numerical data, the numerical data must be converted into a fuzzy set to be used in a fuzzy system. These methods include converting the measured value to a fuzzy singleton when the measured numerical value is considered reliable, and the probabilistic characteristics of the measured value when the measured numerical value is considered to be uncertain due to influences such as disturbance. Considering this, there is a method of converting to a fuzzy set in the form of Gaussian, isosceles triangle, and trapezoid.

The rule base includes situations that are the basis of the creation and application of fuzzy rules and fuzzy inference from the experience and knowledge of experts. Fuzzy inference has a logical execution process to apply fuzzy rules described in linguistic form to a fuzzy system. Reasoning refers to the process of leaking out a new fact or relationship from a given fact or relationship, and obtaining an output value through the synthesis of rules and inferences based on expert knowledge. Inference methods include max-min inference and max-product inference.

Defuzzification performs the function of converting the fuzzy values generated in the fuzzy system into numerical values for actual use. For the defuzzification method, a center of gravity method and a mean of maximum method can be used.

Fuzzy models can be classified into two broad categories, which are classified by the form of information representation. The first is the mamdani fuzzy model, which performs fuzzy inference based on IF-THEN language rules described by ambiguous predicates. This type of fuzzy model uses natural language as it is based on linguistic models that describe system behavior. In other words, both the front and back parts of the rule are expressed in a fuzzy set to have flexibility. As an example, a fuzzy controller can be considered, and the fuzzy control rules are expressed as a language representing a control strategy based on common sense.

The second is the T-S-K fuzzy model, which has a mixture of fuzzy and non-fuzzy models. The rule form has a pre-determined part composed of a fuzzy set and a post-determined part provided in a function form, and has the ability to express numerical information. Therefore, if information is given in the form of data, system identification can be performed more easily than a fuzzy model using a learning algorithm. Unlike the Mamdani fuzzy model, the T-S-K fuzzy model is a mixed fuzzy reasoning method that expresses the posterior part membership function as an input/output linear relational expression or a constant value. This inference method has the advantage of being able to accurately express input/output relationships with nonlinear characteristics even with fewer rules than the conventional linear approximation method by fuzzy partitioning the input space and allocating linear input/output relationships for each fuzzy subspace.

과 1개의 출력 변수 y_i를 갖는 T-S-K 퍼지모델의 퍼지규칙을 보여주고 있다. [Equation 4] below shows n input variables

It shows the fuzzy rule of the TSK fuzzy model with and one output variable y _i .

[Equation 4]

는 퍼지 입력 멤버십함수이며, 가우시안 형태일 경우 하기 [수학식5] 로 표현될 수 있다. y_i는 i번째 규칙의 출력을 나타내며,

는 입력 변수들과 상수들의 선형 함수로서 하기 [수학식6]으로 표현될 수 있다.From here

Is a fuzzy input membership function, and in the case of a Gaussian form, it can be expressed by the following [Equation 5]. y _i represents the output of the ith rule,

Is a linear function of input variables and constants, and can be expressed as Equation 6 below.

[Equation 5]

[Equation 6]

는 데이터가 주어지면 결정되어야 하는 상수이다. Here, c _ji is the center of the membership function, and w _ji is the variable representing the width of the membership function.

Is a constant that must be determined given data.

번째 전건부 규칙의 정규화한 적합도 μ_i는 식 하기 [수학식7]과 같이 표현되며 퍼지 시스템의 최종 추론값 y^*는 하기 [수학식8]과 같이 표현될 수 있다. When using the max-product inference method with M rules,

The normalized goodness-of-fit μ _i of the first pre-determined rule is expressed as Equation 7 below, and the final inference y ^* of the fuzzy system can be expressed as Equation 8 below.

[Equation 7]

[Equation 8]

The T-S-K fuzzy model has two advantages over the Momdani fuzzy model in that the posterior part variable is expressed as a function of the input variable. First, rule extraction from the operator's knowledge or experience results in a task of estimating the coefficient of the linear equation from the given input/output data, so that fuzzy rules can be easily generated. Second, it is possible to include a lot of information with a small number of control rules by using a linear equation in the posterior part.

On the other hand, when modeling with the mamdani fuzzy model, the modeling process can be divided into a process of dividing the cluster center into an input/output space and a process of mapping with a fuzzy rule.

은 입출력 공간으로 분할될 수 있다. 즉, i번째 클러스터 중심 X_i ^* 는 입력 공간이 n차원인 x_i ^*로, 출력 공간이 1차원인 z_i ^*로 분할될 수 있다. First, looking at the process of dividing the cluster center into input/output spaces, the (n+1) dimensional M cluster centers

Can be divided into input/output space. That is, the i-th cluster center X _i ^* may be divided into x _i ^* having an n-dimensional input space and z _i ^* having a 1-dimensional output space.

Next, the divided input/output space is mapped with a fuzzy rule. That is, the divided input space x _i ^* determines the position of the pre-construction unit membership function of the fuzzy rule, and the divided output space z _i ^* determines the position of the output membership function. In the study, a Gaussian-type membership function as shown in [Equation 5] was used, and the width w _ji was determined through [Equation 9] below based on experience.

[Equation 9]

는 클러스터 중심의 입력 공간에서 벡터간의 최대거리를 나타낸다.Where M is the number of cluster centers, r _a is the same value as used in [Equation 1], and

Represents the maximum distance between vectors in the cluster-centered input space.

In the Mamdani fuzzy model in which the input variable is n-dimensional x and the output variable is 1-dimensional z, the fuzzy rule is given by [Equation 10] in the form of [Equation 4], and the max-product inference method When is used, the fitness μ _i of the i-th fuzzy rule can be obtained through the following [Equation 11].

[Equation 10]

[Equation 11]

Here, α is the same value as the value used in [Equation 1], and the input membership function is in the Gaussian form as in [Equation 5].

The final output y ^* can be expressed by the following [Equation 12] when the posterior member membership function of the rule is given as a singleton z _i ^* and a simplified reasoning method is used.

[Equation 12]

Here, the divided output space z _i ^* becomes a linear function of constants and input variables as shown in [Equation 13] below.

[Equation 13]

When linear least-squares estimation is applied to determine the parameters G _i and h _i , the normalized goodness of fit ρ _i of the i-th fuzzy rule is expressed as Equation 14, and the output y is It is expressed as Equation 15].

[Equation 14]

[Equation 15]

및 출력 데이터

가 주어지면 하기 [수학식16]을 통해 계산할 수 있다. Determination of parameters G _i and h _i is N input data

And output data

Given is, it can be calculated through the following [Equation 16].

[Equation 16]

Where i=1,2,... ,M and j=1,2,... When ,N, ρ _ji represents the normalized fitness of the i-th rule in the j-th data.

와

를 구할 수 있다. 단, 입출력 데이터의 수는 결정하고자 하는 파라미터보다 많아야 한다. [Equation 16] is a process of determining a parameter to minimize the squared error between the output of the model and the output of the given data, and can be simply expressed in the form of [Equation 17] and recursive least-squares estimation) or pseudo inverse

Wow

Can be obtained. However, the number of input/output data must be greater than the parameter to be determined.

[Equation 17]

를 최소화하는 방법인 의사 역행렬 방법의 하기 [수학식18]을 사용하여 하기 [수학식19]로 해를 구하였다. To determine G _i and h _i

The solution was obtained by the following [Equation 19] using the following [Equation 18] of the pseudo-inverse matrix method, which is a method of minimizing.

[Equation 18]

은 많은 계산 시간을 필요로 하며,

의 역행렬이 존재하지 않을 수 있기 때문에 재귀형 최소자승법을 사용하여 해를 구하기도 한다. But,

Takes a lot of computation time,

Since the inverse matrix of may not exist, a recursive least-squares method is used to find the solution.

When the posterior construction parameters G _i and h _i of the mamdani fuzzy rule are determined, the final output value y ^* can be obtained by the following [Equation 20].

[Equation 20]

It can be seen that the result of [Equation 20] is consistent with the final output in the T-S-K fuzzy model. In the T-S-K fuzzy system model used in this study, the posterior construction part is expressed in the form of a function of the previous construction part input variables. The T-S-K fuzzy model is a mixed fuzzy inference method that expresses the posterior part membership function as an input/output linear relational expression or a constant value. This inference method has the advantage of being able to express input/output relationships having nonlinear characteristics with fewer rules than the conventional linear approximation method by fuzzy partitioning the input space and allocating linear input/output relationships to each fuzzy subspace.

In summary, the method of providing a commercial area analysis service according to an embodiment of the present invention sets sales information as input/output data, and based on the cluster center clustered through a subtractive clustering algorithm, the Momdani fuzzy model or the TSK fuzzy model Through the fuzzy system modeled through, it is possible to calculate the expected sales of the business site.

5 is a flowchart illustrating an example of a process of calculating expected sales according to an embodiment of the present invention.

Referring to FIG. 5, for example, a business operator may sequentially input business information of a location of a business place, a predicted month, a business type, and an advertisement cost in a program of a business operator's terminal. The commercial area analysis service providing server may receive the sales information, input it into a designed fuzzy system, and output expected future sales as an analysis result. The output predicted sales information may be transmitted to the business operator's terminal and provided in a program of the business operator's terminal.

6 is a diagram showing a graph of predicted sales statistics of a business operator terminal program according to an embodiment of the present invention.

Referring to FIG. 6, as a final step, the estimated sales output through the fuzzy system may be provided as time-series statistical graph data in a program of the business operator's terminal (S220).

7 is a diagram illustrating a prediction error according to the number of fuzzy rules in a fuzzy system according to an embodiment of the present invention.

Referring to FIG. 7, it can be seen that the prediction error rate decreases according to the number of fuzzy rules. For example, if the input variables are the industry, region, advertising cost, and sales of a foliar farm, if 5 fuzzy rules are generated, there is some prediction error (e ^k _min) , but when 10 fuzzy rules are generated It can be seen that the prediction error (e ^k _min) is relatively small.

The server and method for providing a commercial area analysis service according to an embodiment of the present invention described above provide predicted sales information to the business owner by analyzing the business area, and provide data obtained by analyzing various factors affecting sales to manage the business of the business area. It can be improved efficiently.

The server and method for providing a commercial area analysis service according to an embodiment of the present invention can reduce the risk when small business owners start a business by providing a time-series prediction of how much future expected sales will be in a commercial area where small business owners intend to do business.

Meanwhile, the method for providing a commercial area analysis service according to an embodiment of the present invention may be implemented in the form of a recording medium including instructions executable by a computer such as a program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer, and includes both volatile and nonvolatile media, removable and non-removable media. Further, the computer-readable medium may include a computer storage medium. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Although the methods and systems of the present invention have been described in connection with specific embodiments, some or all of their components or operations may be implemented using a computer system having a general-purpose hardware architecture.

The above description is merely illustrative of the technical idea of the present invention, and those of ordinary skill in the art to which the present invention pertains will be able to make various modifications and variations without departing from the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention, but to explain the technical idea, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be interpreted by the claims below, and all technical ideas within the scope equivalent thereto should be interpreted as being included in the scope of the present invention.

100: commercial area analysis service providing server
110: communication module
120: memory
130: processor
140: database

Claims (17)

In the commercial district analysis service providing method performed by the commercial district analysis service providing server,
Receiving business information including business type, location, date and sales from the business operator terminal; And
Comprising the step of calculating the expected sales of the business site using a fuzzy system based on the business information,
The fuzzy system is modeled based on a cluster center in which past sales information including pre-stored business type, location, date, and sales are clustered through a subtractive clustering algorithm,
The subtractive clustering algorithm normalizes the past sales information to calculate a distance between data, and calculates a cluster center based on a size of a distance potential value between the normalized data. The method of claim 1,
The subtractive clustering algorithm,
A method for providing a commercial area analysis service for calculating a distance between normalized data based on the following [Equation 1] and calculating a k (1≦k≦N)-th cluster center having the largest distance potential value between the data.
[Equation 1]

P _i = distance potential value between data
N = number of input data
X _i , X _j = data value
α =

, (r _a = positive constant) The method of claim 2,
Comprising the step of calculating the k+1-th cluster center (X _k+1 ) from which the calculated k-th cluster center (X _k ^* ) is removed based on the following [Equation 3] .
[Equation 3]

P _i ^' = potential value with the effect of the k-th cluster center removed
P _i = distance potential value between data
P _k ^* = potential value of the k-th cluster center
X _i = data
β =

, (r _b = positive constant) The method of claim 3,
For the calculated cluster centers, the upper limit criteria for cluster center selection (

, 1>

) And lower limit standard (

,

>

> 0),

this

If it is larger, it is selected as the cluster center,

this

If it is smaller than that, it is not selected as the cluster center, and if it is in between, the method of providing commercial district analysis service is conditionally selected as the cluster center. The method of claim 2,
r _a = 0.3 people, how to provide commercial analysis service. The method of claim 3,
r _b = 1.5r _a , how to provide commercial analysis services. The method of claim 1,
The fuzzy system is a method for providing a commercial area analysis service based on a Momdani fuzzy algorithm. The method of claim 1,
The fuzzy system is a method for providing a commercial area analysis service based on a TSK (Takagi-Sugeno-Kang) fuzzy algorithm. In the server that provides a commercial area analysis service,
A memory in which a program providing a commercial area analysis service is stored; And
Comprising a processor for executing a program stored in the memory,
According to the execution of the program, the processor,
Receives sales information including business type, location, date and sales from the business operator terminal, and calculates the expected sales of the business site using a fuzzy system based on the business information,
The fuzzy system is modeled based on a cluster center in which past sales information including pre-stored business type, location, date, and sales are clustered through a subtractive clustering algorithm,
The subtractive clustering algorithm normalizes the past sales information to calculate a distance between data, and calculates a cluster center based on a size of a distance potential value between the normalized data. The method of claim 9,
The subtractive clustering algorithm,
A commercial area analysis service providing server for calculating a distance between normalized data based on the following [Equation 1] and calculating a k (1≦k≦N)-th cluster center having the largest distance potential value between the data.
[Equation 1]

P _i = distance potential value between data
N = number of input data
X _i , X _j = data
α =

, (r _a = positive constant) The method of claim 10,
Comprising the step of obtaining the k+1-th cluster center (X _k+1 ) from which the calculated k-th cluster center (X _k ^* ) is removed based on the following [Equation 3].
[Equation 3]

Pi' = potential value with the effect of the k-th cluster center removed
P _i = distance potential value between data
P _k ^* = potential value of the k-th cluster center
X _i = data
β =

, (rb=positive constant) The method of claim 11,
The method of claim 3,
For the calculated cluster centers, the upper limit criteria for cluster center selection (

, 1>

) And lower limit standard (

,

>

> 0),

this

If it is larger, it is selected as the cluster center,

this

If it is smaller than, it is not selected as the cluster center, and if it is in between, it is conditionally selected as the cluster center. The method of claim 10,
ra = 0.3 person, server providing commercial analysis service. The method of claim 11,
rb = 1.5ra, commercial area analysis service providing server. The method of claim 9,
The fuzzy system is based on a Momdani fuzzy algorithm, a commercial area analysis service providing server. The method of claim 9,
The fuzzy system is based on a TSK (Takagi-Sugeno-Kang) fuzzy algorithm, a commercial area analysis service providing server. A computer-readable recording medium recording a program for performing the method according to any one of claims 1 to 8 on a computer.

Images