KR102223161B1 - System and methods of predicting timeliness products based on weather data - Google Patents

Abstract

The present invention relates to a system for predicting timeliness products based on weather data which can predict the timeliness products required in accordance with the weather data and provide the same to a user. More specifically, according to one aspect of the present invention, a server of the system for predicting the timeliness products based on the weather data comprises: a first period setting unit setting a first period for collecting weather data for the first period; a second period setting unit setting a second period for collecting the weather data for the second period after the first period; a first weather data collection unit which collects first weather data for the first period; a second weather data collection unit which collects second weather data for the second period; a search data collection unit collecting search data for the first period; and a timeliness product output unit outputting a timeliness product corresponding to the second weather data based on the first weather data and the search data.

Description

A system and method for predicting timing products based on weather data {SYSTEM AND METHODS OF PREDICTING TIMELINESS PRODUCTS BASED ON WEATHER DATA}

The present invention relates to a system and method for predicting a timing product based on meteorological data, and more particularly, to a system and method for predicting a timing product based on meteorological data for predicting a timing product using meteorological data and search data. will be.

The sale of goods through vending machines is developing in various forms. Initially, beverage products such as canned beverages and coffee were mainly sold, but in recent years, various vending machines that sell products such as instant food, hygiene products, books, and toys are in operation. It is increasing rapidly.

However, vending machines that sell beverage products such as canned drinks and coffee are disappearing. According to the Seoul Metropolitan Government, as of 2016, the number of beverage vending machines in Seoul decreased by 57% from 15,623 in 2008 to 6,658. There are many inconveniences such as poor hygiene, items limited to beverages and coffee, and the vending machine market, which has been rapidly evolving since 2016, is filled with unique vending machines such as flowers, cosmetics, salads, and book vending machines.

Recently, vending machines that sell many types of products are increasing, but there is a problem that there is no vending machine that sells necessary products according to weather information of the day.

Seasonal products, which are products that can surge in demand over time and season, are becoming a necessity for busy modern people. In addition to fine dust masks that can prevent fine dust, modern people are in need of seasonal products according to weather information such as umbrellas, raincoats, hot packs, cool packs, cool towels, portable fans, and fans.

KR 10-1796720 B1 KR 10-2019-0103505 A

The present invention is to solve the above problems, and it is possible to predict and provide the necessary timely products according to the meteorological data based on meteorological data such as temperature, precipitation, cloudiness, relative humidity, and the concentration of fine dust in the atmosphere. The object of the present invention is to provide a system and method for predicting timing products based on existing weather data.

The object of the present invention is not limited to the above-mentioned object, and other objects not mentioned will be clearly understood from the following description.

A server of a timing product prediction system based on meteorological data according to an aspect of the present invention for achieving the above object includes a first period setting unit for setting a first period for collecting meteorological data for a first period, A second period setting unit for setting a second period for collecting weather data for a second period after the first period, a first weather data collection unit for collecting first weather data for the first period, and A second weather data collection unit that collects second weather data for a period of 2, a search data collection unit that collects search data for a first period, and second weather data based on the first weather data and search data. It includes a timing product output unit for outputting the corresponding timing product.

According to another aspect of the present invention, a timing product prediction system based on meteorological data collects a plurality of vending machines, meteorological data and search data of a first period, and weather data of a second period after the first period, and the first Based on the weather data and search data of the period, a server that outputs the seasonal product corresponding to the weather data of the second period, and receives information of the timing product from the server, and generates coordinate information corresponding to the location of the vending machine, or And a vending machine management module including a coordinate information unit to be managed and an identification code setting unit for setting an identification code for identifying the vending machine.

By analyzing the weather data and recommending appropriate timely products suitable for the weather information on the day, there is an effect that the user can purchase the desired product.

By setting the priorities of vending machines according to the number of sales of seasonal products per hour, it is possible to know which vending machines and regions sell best seasonal products, and it is effective to increase the efficiency of vending machine management.

In addition, there is an effect of reducing the inconvenience of running out of inventory and wasting inventory by setting the sales volume according to the vending machine and region, which sells the most timely products.

The effects according to the present invention are not limited by the contents exemplified above, and more various effects are included in the present specification.

1 is a block diagram showing a system for predicting a timing product based on weather data according to an embodiment of the present invention.
2 is a block diagram illustrating a server and a vending machine management module of a system for predicting a timing product based on weather data according to an embodiment of the present invention.
3 is a flowchart illustrating a method of predicting a timing product based on weather data according to another embodiment of the present invention.

Advantages and features of the present invention, and a method of achieving them will become apparent with reference to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but will be implemented in a variety of different situations, and only these embodiments make the disclosure of the present invention complete, and those skilled in the art to which the present invention pertains. As provided to fully inform the scope of the invention to the person, the invention is only defined by the description of the claims. On the other hand, terms used in the present specification are for describing exemplary embodiments, and are not intended to limit the present invention. In this specification, the singular form also includes the plural form unless specifically stated in the phrase.

Hereinafter, a system and method for predicting a timing product based on weather data according to a preferred embodiment of the present invention will be described with reference to the drawings.

In addition, in the embodiments of the present invention, functionally identical components are described by matching reference numerals, and repeated descriptions are omitted.

Referring to FIG. 1, a system for predicting timing products based on weather data according to an embodiment of the present invention includes a server 100 and a vending machine management module 200. It includes a plurality of vending machines (300).

The plurality of vending machines 300 may be provided with various sensors, such as a sensor for measuring inventory and a GPS for acquiring coordinate information.

2, the server 100 includes a first period setting unit 110, a second period setting unit 115, a weather data item generation unit 120, a first weather data collection unit 125, and a second Weather data collection unit 130, item generation unit 135, keyword generation unit 140, search data collection unit 145, normalization unit 150, search data maximum value extraction unit 155, binary conversion unit ( 160), a model unit 165, a learning unit 170, a timing product output unit 175, and a communication unit 180.

The first period setting unit 110 sets a first period (i) for collecting weather data during the first period.

The second period setting unit 115 sets a second period for collecting weather data for a second period after the first period.

The meteorological data item generating unit 120 generates a meteorological item including at least one of temperature, precipitation, wind speed, humidity, cloudiness, and fine dust concentration.

The meteorological data item generation unit 120 includes temperature, minimum temperature, maximum temperature, precipitation, average humidity, average wind speed, maximum wind speed, sunshine time, average total cloud volume, sulfur dioxide (SO2), ozone (O3), nitrogen dioxide (NO2). , Carbon monoxide (CO), fine dust (PM10), and ultrafine dust (PM2.5) may be generating at least one of the gas phase items.

The first meteorological data collection unit 125 collects first meteorological data for a first period.

The first weather data may represent weather data including values corresponding to each weather item for each date during the first period.

The first meteorological data collection unit 125 may be collecting meteorological data including a value corresponding to the meteorological item generated by the meteorological data item generator 120 for each date during the first period.

The first meteorological data collection unit 125 may collect meteorological data from an external meteorological information providing server that provides weather-related information.

The second meteorological data collection unit 130 collects second meteorological data for a second period.

The second weather data may represent weather data including values corresponding to each weather item during the second period.

The second meteorological data collection unit 130 may collect meteorological data including a value corresponding to the meteorological item generated by the meteorological data item generating unit 120 for each date during the second period.

The item generation unit 135 generates at least one item from among a plurality of temporal products for collecting search data.

Seasonal products refer to products in which demand may surge in season and hourly periods.

The item generation unit 135 includes'sun cream','fan', portable fan','cool toshi','cool mat','cool pack','ice cream','hot pack', and'cosmetic cotton', which correspond to seasonal products. ','Nuts','Vitamin','Hangover Relief','Fine Dust Mask','Mask','Umbrella' and'Beverage' items are created.

The keyword generation unit 140 generates a keyword corresponding to the timing product of the item.

The keyword generator 140 may generate a plurality of keywords, respectively, according to items of a preset timing product.

In addition, the keyword generation unit 140 may be configured to generate a plurality of keywords by combining the first period or the date included in the first period according to the item of the preset timing product.

The keyword generator 140 may generate a search keyword of search data to be collected based on the item generated by the item generator 135.

For example, the keyword generation unit 140 generates a plurality of search keywords by combining a date corresponding to one day of the first period and an item'sunscreen', or a date and item corresponding to another day. It may be to generate a plurality of search keywords by combining the'mask'.

The search data collection unit 145 collects search data for a first period.

The search data may be data including a value of the number of times of search for timing products searched on the website for each date during the first period.

The search data collection unit 145 may collect search data according to the number of searches corresponding to the keyword generated by the keyword generation unit 140 for each date during the first period.

That is, the search data collection unit 145 determines the number of searches corresponding to the keyword generated by the keyword generation unit 140 for each date during the first period from an external search server capable of outputting a result according to a query using the keyword. It may be to collect search data according to.

...

...

15.1 12.6 13.3 ... 9.5 14.6 12.2 11.5 ... 8.3

6.9 8.7 6.8 ... 8.3 5.9 8.1 7.1 ... 7.2

21.9 17.3 19.6 ... 14.2 20.2 18.2 19.5 ... 15.4 .... ... ... ... ... ... ... ... ... ... ...

6.3 2.6 1.3 ... 3.1 6.5 3.2 1.8 ... 2.9

45 32 29 ... 336

264 321 412 ... 85

121 129 251 ... 145 .... ... ... ... ... ...

64 85 32 ... 19

Table 1 is a first matrix showing the collection of the first weather data, search data, and second weather data during the second period (a) in a matrix form during the first period (i).

Referring to Table 1, the first matrix has values of the first weather data during the first period, the value of search data, and the values of the second weather data during the second period, but the row is the weather item (j) row. It is divided into a set and an item (k) row set (, and the columns are divided into a first period (i) column set and a second period (a) column set.

으로 마지막 번째 열을

로 하고, 제2 기간(a)에 속하는 열들은 첫번째 열을

으로 하고 마지막 번째 열을

으로 하며, 기상 항목(j)에 속하는 행들은 첫번째 행을

으로 마지막 번째 행을

로 하고, 품목(k)에 속하는 행들은 첫번째 행을

으로 마지막 번째 행을

로 하여 설명하도록 한다.Hereinafter, for convenience of description, the columns belonging to the first period (i) in Table 1 are referred to as the first column.

The last column with

And, for the columns belonging to the second period (a), the first column is

And the last column

And the first line is the line belonging to the weather item (j).

The last row with

And the rows belonging to item (k) are the first row

The last row with

Let's explain it.

Referring to Table 1, the first weather data corresponds to an area where the weather item (j) row set and the first period (i) column set overlap, and the second weather data is the weather item (j) row set and the second The period (a) corresponds to an area where the column set overlaps, and the search data may correspond to an area where the item (k) row set and the first period (i) column set overlap.

The normalization unit 150 normalizes a value corresponding to a weather item included in the first weather data and the second weather data, and a value of the number of searches included in the search data to a value in a preset range.

Since the difference in units and values between the first weather data, the second weather data, and the search data is different, the normalization unit 150 converts values and search data corresponding to the weather items included in the first weather data and the second weather data. It may be normalized to a value in a preset range.

The normalization unit 150 may normalize based on Equation 1 below.

Here, x is a value subject to normalization, min is a value corresponding to each weather item in the first period and a minimum value of a value corresponding to each item, and Max is a value corresponding to each weather item in the first period And the maximum value of the value corresponding to each item.

' 기간의 '

' 기상 항목에 해당하는 값을 정규화할 시, 제1 기간(i) 동안

의 최대값이 40, 최소값이 1 인 경우 정규화부(150)는 수학식 1에 기초하여 '0.362'로 정규화할 수 있다.For example, referring to Table 1, '

'Of the period'

'' When normalizing the value corresponding to the weather item, during the first period (i)

When the maximum value of is 40 and the minimum value is 1, the normalization unit 150 may normalize to '0.362' based on Equation 1.

'의 '

' 품목에 해당하는 값을 정규화할 시, 제1 기간(i) 동안

의 최대값이 521, 최소값이 32 인 경우 정규화부(150)는 수학식 1에 기초하여 '0.777'로 정규화할 수 있다.In addition, in Table 1,

'Of'

'When normalizing the value corresponding to the item, during the first period (i)

When the maximum value of is 521 and the minimum value is 32, the normalization unit 150 may normalize it to '0.777' based on Equation (1).

...

...

0.362 0.314 0.321 ... 0.273 0.352 0.318 0.328 ... 0.265

0.214 0.215 0.196 ... 0.208 0.201 0.210 0.204 ... 0.205

0.312 0.309 0.308 ... 0.309 0.301 0.318 0.307 ... 0.313 .... ... ... ... ... ... ... ... ... ... ...

0.621 0.235 0.134 ... 0.352 0.625 0.244 0.148 ... 0.388

0.214 0.198 0.189 ... 0.854

0.741 0.768 0.777 ... 0.325

0.622 0.677 0.754 ... 0.588 .... ... ... ... ... ...

0.325 0.387 0.214 ... 0.118

Table 2 shows a second matrix in which the normalizer 150 normalizes values corresponding to the first matrix.

The search data maximum value extracting unit 155 extracts a maximum value for each date from among the values of the normalized search data for the first period.

)의 검색데이터의 값 중에서 최대값을 추출하는 것일 수 있다.The search data maximum value extracting unit 155 includes each period (

) May be extracting the maximum value from the value of search data.

' 기간에서는 최대값인 '0.741'를 추출하고, '

' 기간에서는 최대값인 '0.777'를 추출한다.For example, the search data maximum value extracting unit 155 includes'

'In the period, the maximum value of '0.741' is extracted, and'

'In the period, the maximum value of '0.777' is extracted.

The binary conversion unit 160 converts the maximum value extracted by the search data maximum value extraction unit 155 into a first binary data value, and converts a value of search data other than the maximum value into a second binary data value.

The first binary data value may be '1' or '0', and the second binary data value may be '0' or '1'.

...

...

0.362 0.314 0.321 ... 0.273 0.352 0.318 0.328 ... 0.265

0.214 0.215 0.196 ... 0.208 0.201 0.210 0.204 ... 0.205

0.312 0.309 0.308 ... 0.309 0.301 0.318 0.307 ... 0.313 .... ... ... ... ... ... ... ... ... ... ...

0.621 0.235 0.134 ... 0.352 0.625 0.244 0.148 ... 0.388

0 0 0 ... One

One One One ... 0

0 0 0 ... 0 .... ... ... ... ... ...

0 0 0 ... 0

Table 3 shows a third matrix in which the search data binary converter 160 converts the search data values of Table 2 into first binary data values or second binary data values.

)과, 제1 기간(i)의 검색데이터에 대한 랜덤 포레스트 회귀 모델에 해당하는 제2 모델값(

)을 산출하여 랜덤 포레스트 모델을 구축한다.The model unit 165 includes a first model value corresponding to a random forest regression model for the first meteorological data of the first period (i) in order to predict a timing product based on the random forest.

) And a second model value corresponding to the random forest regression model for the search data of the first period (i) (

) To build a random forest model.

Random Forest is an algorithm that creates a decision tree by randomly sampling data exclusively for learning, and collects the results of the decision trees to derive a final result with a majority vote. It is a kind of learning method.

Here, the decision tree is a model that predicts the value of a target variable according to several input variables, and is a tree used to make a decision. It has high accuracy in generating a large number of decision trees and learning it in a group to derive the result by majority vote.

)과 제2 모델값(

)을 각각 수식으로 나타낸 것이다.Equations 2 and 3 below are the first model values (

) And the second model value (

) Is represented by an equation.

는 반응변수를 나타내고,

는 설명변수를 나타낸다.here,

Represents the response variable,

Represents an explanatory variable.

는 반응변수를 나타내고,

는 설명변수를 나타낸다.here,

Represents the response variable,

Represents an explanatory variable.

)과 제2 모델값(

)에 기초하여 제1 기상데이터, 검색데이터를 학습 전용 데이터를 랜덤 샘플링하여 랜덤 포레스트 학습을 수행한다.The learning unit 170 is the first model value (

) And the second model value (

), the first weather data and search data are randomly sampled for learning-only data to perform random forest learning.

) 내에 검색데이터의 품목(k) 중에서 제1 이진 데이터 값을 가지는 상품을 해당 기간의 기상데이터에 대응되는 시기성상품으로 예측하는 랜덤 포레스트 학습을 수행하는 것일 수 있다.The learning unit 170 includes each period of the first period (i) (

), a product having a first binary data value among items (k) of search data in) may be a random forest learning that predicts a product having a first binary data value as a timely product corresponding to the weather data for a corresponding period.

기간의 기상데이터에는

기간의 검색데이터의 품목(k) 중에서 제1 이진데이터 값에 대응되는 상품을 시기성상품으로 예측하고,

기간의 기상데이터에는

기간의 검색데이터의 품목(k) 중에서 제1 이진데이터 값에 대응되는 상품을 시기성상품으로 예측하며,

기간의 기상데이터에는

기간의 검색데이터의 품목(k) 중에서 제1 이진데이터 값에 대응되는 상품을 시기성상품으로 예측하는 랜덤 포레스트 학습을 수행하는 것일 수 있다.The learning unit 170

In the weather data of the period

The product corresponding to the first binary data value among the items (k) of the search data of the period is predicted as a timing product,

In the weather data of the period

The product corresponding to the first binary data value among the items (k) of the search data of the period is predicted as a timing product,

In the weather data of the period

It may be to perform random forest learning for predicting a product corresponding to the first binary data value as a timed product among items (k) of the search data of the period.

The learning unit 170 may perform learning using techniques such as big data, deep-learning, and data-mining, as well as random forest learning.

The timing product output unit 175 outputs a timing product corresponding to the second weather data of the second period (a).

The timing product output unit 175 may be to output a timing product corresponding to the second weather data based on the data of the random forest learning performed by the learning unit 170.

When the second weather data is input, the timing product output unit 175 outputs the corresponding timing product according to the value of each weather data of the weather item according to the data of the random forest learning performed by the learning unit 170. Can be.

The output of the timing product in the timing product output unit 175 may be the name of the timing product or outputting information about the timing product (for example, specifications, specifications, serial numbers, etc. of the timing product). have.

The communication unit 180 transmits the timing product output from the timing product output unit 175 to the vending machine management module 200.

2, the vending machine management module 200 includes a communication unit 210, a coordinate information unit 220, an identification code setting unit 230, an inventory information unit 240, a sales information unit 250, a priority setting unit 260 ), and a sales amount setting unit 270.

The communication unit 210 receives the timing product output from the server 100.

The communication unit 210 may receive a name of a timing product or information on a timing product from the server 100.

The coordinate information unit 220 generates or manages coordinate information corresponding to the positions of a plurality of vending machines 300 respectively installed in different regions.

The coordinate information unit 220 may generate coordinate information through a sensor such as GPS provided in the vending machine 300.

The identification code setting unit 230 sets an identification code for identifying a plurality of vending machines 300.

The inventory information unit 240 outputs inventory information for timing products for each vending machine.

The sales information unit 250 outputs the stock amount of the timing products of each vending machine 300 and the number of sales by time according to coordinate information and identification codes of the plurality of vending machines 300.

Coordinate information Identification code Elapsed time 'Mask'
Number of sales Remaining inventory Vending machine 1 Chungnyeol-daero 137beon-gil A0001 6h 186 114 Vending machine 2 Uncheoncheon-ro 285beon-gil A0011 6h 240 60 Vending machine 3 Minam-ro 132beon-gil A0020 6h 198 102 Vending machine 4 Jungang-daero 1367beon-gil A0032 6h 126 174

Table 4 shows the inventory amount of timing products and the number of sales by time in a plurality of vending machines 300 output by the sales information unit 250 according to a preset elapsed time when the timing product received from the server 100 is a'mask'. Is shown.

The priority setting unit 260 sets the priority according to the priority given according to the number of sales of timing products.

The priority setting unit 260 sets the priority according to the number of sales per hour according to the number of sales of the timing products and the elapsed time for each vending machine based on the timing products received from the server 100.

For example, if the timing product received from the server 100 is'mask', the number of'mask' sales per hour of'vending machine 1'is '31', and the number of'mask' sales per hour of'vending machine 2'is' Since the number of'masks' sold per hour for '40' and'Vending Machine 3'is '33', and the number of'masks' sold per hour for'Vending Machine 4'is '21', the priority setting unit 260 is called'Vending Machine 2'. The priority may be set with '1st priority,'vending machine 3'as 2nd priority,'vending machine 1'as 3rd priority, and'vending machine 4'as 4th priority.

The priority setting unit 260 may be configured to set the priority by calculating the number of sales per hour, as well as the number of sales of timing products for each time slot and for each season.

The sales quantity setting unit 270 sets the sales quantity to be replenished based on the stock quantity and priority of the timing products of the plurality of vending machines 300.

The sales amount setting unit 270 sets the sales amount to be replenished for each priority based on the current inventory amount of the plurality of vending machines 300 and the number of sales per hour.

For example, since the vending machine is operated 24 hours, the sales volume setting unit 270 sets the number of sales per hour of'vending machine 2'multiplied by '24' as the number of sales per day, and the number of sales per day is the current inventory amount. It may be to set the number of subtracted by as the amount of sales to be supplemented.

3 is a flowchart illustrating a method of predicting a timing product based on weather data according to another embodiment of the present invention. A method for predicting a timely product based on weather data according to another embodiment of the present invention may be performed by a system for predicting a timely product based on weather data according to an embodiment of the present invention. Hereinafter, with reference to FIG. 3, a method for predicting a timing product based on weather data according to another embodiment of the present invention will be described, and weather data according to an embodiment of the present invention described with reference to FIGS. 1 to 2 Contents and configurations overlapping with the timing product prediction system based on are matched with reference numerals, and detailed descriptions will be omitted for convenience of description.

First, the first period setting unit 110 sets a first period (i) for collecting weather data during the first period (S110).

The second period setting unit 115 sets a second period for collecting weather data for a second period after the first period (S115).

The meteorological data item generation unit 120 generates a meteorological item including at least one of temperature, precipitation, wind speed, humidity, cloud cover, and fine dust concentration (S120).

The meteorological data item generation unit 120 includes temperature, minimum temperature, maximum temperature, precipitation, average humidity, average wind speed, maximum wind speed, sunshine time, average total cloud volume, sulfur dioxide (SO2), ozone (O3), nitrogen dioxide (NO2). , Carbon monoxide (CO), fine dust (PM10), and ultrafine dust (PM2.5) may be generating at least one of the gas phase items.

The first meteorological data collection unit 125 collects first meteorological data for a first period (S125).

The first weather data may represent weather data including values corresponding to each weather item for each date during the first period.

The first meteorological data collection unit 125 may be collecting meteorological data including a value corresponding to the meteorological item generated by the meteorological data item generator 120 for each date during the first period.

The first meteorological data collection unit 125 may collect meteorological data from an external meteorological information providing server that provides weather-related information.

The second meteorological data collection unit 130 collects second meteorological data for a second period (S130).

The second weather data may represent weather data including values corresponding to each weather item during the second period.

The second meteorological data collection unit 130 may collect meteorological data including a value corresponding to the meteorological item generated by the meteorological data item generating unit 120 for each date during the second period.

The item generation unit 135 generates at least one item from among a plurality of timing products for collecting search data (S135).

Seasonal products refer to products in which demand may surge in season and hourly periods.

The item generation unit 135 includes'sun cream','fan', portable fan','cool toshi','cool mat','cool pack','ice cream','hot pack', and'cosmetic cotton', which correspond to seasonal products. ','Nuts','Vitamin','Hangover Relief','Fine Dust Mask','Mask','Umbrella' and'Beverage' items are created.

The keyword generation unit 140 generates a keyword corresponding to the timing product of the item (S140).

The keyword generator 140 may generate a plurality of keywords, respectively, according to items of a preset timing product.

In addition, the keyword generation unit 140 may be configured to generate a plurality of keywords by combining the first period or the date included in the first period according to the item of the preset timing product.

The keyword generator 140 may generate a search keyword of search data to be collected based on the item generated by the item generator 135.

For example, the keyword generation unit 140 generates a plurality of search keywords by combining a date corresponding to one day of the first period and an item'sunscreen', or a date and item corresponding to another day. It may be to generate a plurality of search keywords by combining the'mask'.

The search data collection unit 145 collects search data for the first period (S145).

The search data may be data including a value of the number of times of search for timing products searched on the website for each date during the first period.

The search data collection unit 145 may collect search data according to the number of searches corresponding to the keyword generated by the keyword generation unit 140 for each date during the first period.

That is, the search data collection unit 145 determines the number of searches corresponding to the keyword generated by the keyword generation unit 140 for each date during the first period from an external search server capable of outputting a result according to a query using the keyword. It may be to collect search data according to.

The normalization unit 150 normalizes a value corresponding to a weather item included in the first weather data and the second weather data, and a value of the number of searches included in the search data to a value within a preset range (S150).

Since the difference in units and values between the first weather data, the second weather data, and the search data is different, the normalization unit 150 includes values and search data corresponding to the weather items included in the first weather data and the second weather data. May be normalizing to a value in a preset range.

)의 검색데이터의 값 중에서 최대값을 추출하는 것일 수 있다.The search data maximum value extracting unit 155 includes each period (

) May be extracting the maximum value from the value of search data.

' 기간에서는 최대값인 '0.741'를 추출하고, '

' 기간에서는 최대값인 '0.777'를 추출한 다.For example, the search data maximum value extracting unit 155 includes'

'In the period, the maximum value of '0.741' is extracted, and'

In the'period,' the maximum value of '0.777' is extracted.

The binary conversion unit 160 converts the maximum value extracted by the search data maximum value extraction unit 155 into a first binary data value, and converts the search data value other than the maximum value into a second binary data value (S160).

The first binary data value may be '1' or '0', and the second binary data value may be '0' or '1'.

)과, 제1 기간(i)의 검색데이터에 대한 랜덤 포레스트 회귀 모델에 해당하는 제2 모델값(

)을 산출하여 랜덤 포레스트 모델을 구축한다(S165).The model unit 165 includes a first model value corresponding to the random forest regression model for the first meteorological data of the first period (i) in order to predict the timing product based on the random forest.

) And a second model value corresponding to the random forest regression model for the search data of the first period (i) (

) To build a random forest model (S165).

)과 제2 모델값(

)에 기초하여 제1 기상데이터, 검색데이터를 학습 전용 데이터를 랜덤 샘플링하여 랜덤 포레스트 학습을 수행한다(S170).The learning unit 170 is the first model value (

) And the second model value (

), random forest learning is performed by randomly sampling first weather data and search data for learning-only data (S170).

) 내에 검색데이터의 품목(k) 중에서 제1 이진 데이터 값을 가지는 상품을 해당 기간의 기상데이터에 대응되는 시기성상품으로 예측하는 랜덤 포레스트 학습을 수행하는 것일 수 있다.The learning unit 170 includes each period of the first period (i) (

), a product having a first binary data value among items (k) of search data in) may be a random forest learning that predicts a product having a first binary data value as a timely product corresponding to the weather data for a corresponding period.

기간의 기상데이터에는

기간의 검색데이터의 품목(k) 중에서 제1 이진데이터 값에 대응되는 상품을 시기성상품으로 예측하고,

기간의 기상데이터에는

기간의 검색데이터의 품목(k) 중에서 제1 이진데이터 값에 대응되는 상품을 시기성상품으로 예측하며,

기간의 기상데이터에는

기간의 검색데이터의 품목(k) 중에서 제1 이진데이터 값에 대응되는 상품을 시기성상품으로 예측하는 랜덤 포레스트 학습을 수행하는 것일 수 있다.The learning unit 170

In the weather data of the period

The product corresponding to the first binary data value among the items (k) of the search data of the period is predicted as a timing product,

In the weather data of the period

The product corresponding to the first binary data value among the items (k) of the search data of the period is predicted as a timing product,

In the weather data of the period

It may be to perform random forest learning for predicting a product corresponding to the first binary data value as a timed product among items (k) of the search data of the period.

The learning unit 170 may perform learning using techniques such as big data, deep-learning, and data-mining, as well as random forest learning.

The timing product output unit 175 outputs a timing product corresponding to the second weather data of the second period (a) (S175).

The timing product output unit 175 may be to output a timing product corresponding to the second weather data based on the data of the random forest learning performed by the learning unit 170.

When the second weather data is input, the timing product output unit 175 outputs the corresponding timing product according to the value of each weather data of the weather item according to the data of the random forest learning performed by the learning unit 170. Can be.

Outputting a timing product in the timing product output unit 175 may be the name of a timing product or outputting information about the timing product (for example, specifications, specifications, serial numbers, etc. of the timing product). have.

The communication unit 180 transmits the timing product output from the timing product output unit 175 to the vending machine management server 200 (S180).

The communication unit 210 receives the timing product output from the server 100 (S210).

The communication unit 210 may receive a name of a timing product or information on a timing product from the server 100.

The coordinate information unit 220 generates or manages coordinate information corresponding to the positions of the plurality of vending machines 300 respectively installed in different regions (S220).

The identification code setting unit 230 sets an identification code for identifying a plurality of vending machines 300 (S230).

The inventory information unit 240 outputs inventory information for timing products for each vending machine (S240).

The sales information unit 250 outputs the stock quantity of the timing products of each vending machine 300 and the number of sales by time according to coordinate information and identification codes of the plurality of vending machines 300 (S250).

The priority setting unit 260 sets the priority according to the priority given according to the number of sales of timing products (S260).

The priority setting unit 260 sets a priority according to the number of sales per hour according to the number of sales and elapsed time of timing products for each vending machine based on the timing products received from the server 100.

The priority setting unit 260 may be configured to set the priority by calculating the number of sales per hour, as well as the number of sales of timing products for each time slot and for each season.

The sales amount setting unit 270 sets the sales amount to be replenished based on the stock amount and priority of the timing products of the plurality of vending machines 300 (S270).

The sales amount setting unit 270 sets the sales amount to be replenished for each priority based on the current inventory amount of the plurality of vending machines 300 and the number of sales per hour.

As described above, the present invention has been described in detail through preferred embodiments, but the present invention is not limited thereto and may be variously implemented within the scope of the claims. Since the above description somewhat broadly describes the features and technical strengths of the present invention in order to better understand the claims of the invention to be described later, the concept and specific embodiments of the present invention described above are similar to the present invention. It should be recognized by those skilled in the art that it can be used immediately as a basis for designing or changing a shape.

It can be understood that the above-described embodiment is only one embodiment according to the present invention, and can be implemented in various modifications and altered forms within the scope of the technical idea of the present invention by those of ordinary skill in the relevant technical field. There will be. Accordingly, the disclosed embodiments should be considered from a descriptive point of view rather than a limiting point of view, and such various modifications and changes are also shown in the claims of the present invention to be described later as belonging to the scope of the technical idea of the present invention, and the equivalent scope thereof. All differences within will be construed as being included in the present invention.

100: server,
110: first period setting unit, 115: second period setting unit,
120: meteorological data item generation unit, 125: first meteorological data collection unit,
130: second meteorological data collection unit, 135: item generation unit,
140: keyword generation unit, 145: search data collection unit,
150: normalization unit, 155: search data maximum value extraction unit,
160: binary transformation unit, 165: model unit, 170: learning unit,
175: timing product output unit, 180: communication unit,
200: vending machine management module,
210: communication unit, 220: coordinate information unit,
230: identification code setting unit, 240: inventory information unit,
250: sales information unit, 260: priority setting unit,
270: sales volume setting unit,
300: vending machine.

Claims (10)

A first period setting unit that sets a first period for collecting weather data during the first period;
A second period setting unit for setting a second period for collecting weather data for a second period after the first period;
A first meteorological data collection unit for collecting first meteorological data during the first period;
A second meteorological data collection unit for collecting second meteorological data during the second period;
A search data collection unit for collecting search data including a number of searches for each of a plurality of preset periodic products during the first period;
Normalization of normalizing a value corresponding to a weather item included in the first and second weather data and the number of times of searching for each of the plurality of seasonal products to a value in a preset range for each date belonging to each period part;
A search data maximum value extracting unit for extracting a maximum value for each date from values of search data normalized by the normalizing unit for a value of the number of search times for each of the plurality of temporal products;
A binary converter configured to convert the maximum value for each date as a first binary data value among the values of the normalized search data and convert the value of the search data other than the maximum value for each date into a second binary data value;
A first corresponding to a random forest regression model based on a plurality of values normalized by the normalization unit for each date belonging to the first period with respect to the first weather data of the first period in order to predict a timing product based on a random forest. A random forest model by calculating a model value and a second model value corresponding to a random forest regression model based on the binary data values converted by the binary converter for each date belonging to the first period for the search data of the first period. A model unit to build it;
A learning unit that performs random forest learning based on the first model value and the second model value, using the first weather data and the search data as training-only data; And
A server of a timing product prediction system based on weather data comprising a; timing product output unit for outputting a timing product corresponding to the second weather data based on the random forest learning. The method of claim 1,
A meteorological data item generating unit for generating a meteorological item including at least one of temperature, precipitation, wind speed, humidity, cloud cover, and fine dust concentration; further comprising,
The first meteorological data collection unit and the second meteorological data collection unit,
Collecting weather data values corresponding to the weather items
Server of timing product prediction system based on human weather data. The method of claim 1,
An item generator for generating at least one item from among a plurality of timing products for collecting the search data; And
Further comprising; a keyword generating unit for generating a keyword corresponding to the timing product of the item,
The search data collection unit,
Collecting search data according to the number of times the keyword is searched based on the keyword corresponding to the item
Server of timing product prediction system based on human weather data. delete delete delete A first period setting unit for setting a first period for collecting meteorological data during a first period, and a second period setting for setting a second period for collecting weather data for a second period after the first period Part, a first meteorological data collection unit for collecting the first meteorological data for the first period, a second meteorological data collection unit for collecting the second meteorological data for the second period, a preset during the first period A search data collection unit that collects search data including the number of searches for each of a plurality of seasonal products, a value corresponding to a weather item included in the first and second weather data, and the plurality of seasonal products A normalization unit that normalizes the value of the number of searches for each to a value in a preset range for each date belonging to each period, and values of search data normalized by the normalization unit for the value of the number of searches for each of the plurality of timely products. A search data maximum value extracting unit that extracts the maximum value for each date from, the maximum value for each date among the values of the normalized search data as a first binary data value, and a value of the search data other than the maximum value for each date as a second value. A binary conversion unit that converts binary data into binary data values, a plurality of values normalized by the normalization unit for each date belonging to the first period with respect to the first meteorological data of the first period in order to predict a timing product based on a random forest Corresponds to a first model value corresponding to a random forest regression model based on and a random forest regression model based on binary data values converted by the binary converter for each date belonging to the first period for the search data of the first period. A model unit that constructs a random forest model by calculating a second model value, and performs random forest learning based on the first model value and the second model value using the first weather data and the search data as training-only data. A server including a learning unit to perform and a timing product output unit for outputting a timing product corresponding to the second weather data based on the random forest learning;
A plurality of vending machines; And
A vending machine management module including a coordinate information unit for receiving information on timing products from the server, generating or managing coordinate information corresponding to the location of the vending machine, and an identification code setting unit for setting an identification code for identifying the vending machine A timing product prediction system based on meteorological data including; The method of claim 7,
The vending machine management module,
An inventory information unit for outputting inventory information on timing products for each vending machine;
A sales information unit for outputting the number of sales of timing products by time of the vending machine according to the coordinate information and the identification code; And
Further comprising a sales amount setting unit for setting a sales amount based on an inventory amount and the number of sales for each timing product of the vending machine.
Timely product prediction system based on human weather data. The method of claim 7,
The vending machine management module,
Further comprising a priority setting unit for setting the priority according to the priority given according to the number of sales of the timing products of the vending machine.
Timely product prediction system based on human weather data. The method of claim 9,
The priority setting unit,
Setting the priority according to the number of sales per hour according to the number of sales and elapsed time of timing products for each vending machine based on the timing products received from the server.
Timely product prediction system based on human weather data.

Images