KR101963817B1 - Apparatus and method for generating prediction information based on a keyword search volume - Google Patents

Abstract

An apparatus and a method for generating keyword search amount-based prediction information are disclosed. A method of generating a keyword search amount based prediction information according to an exemplary embodiment of the present invention includes: inputting internal data and classifying internal data values by at least one type; Generating search information for each type in which a keyword search amount is collected for each type using keyword search information collected by an Internet portal site; and generating predicted data values at a prediction time from the internal data values using the type- And generating predictive information by calculating the predictive information.

Description

[0001] APPARATUS AND METHOD FOR GENERATING PREDICTION INFORMATION BASED ON A KEYWORD SEARCH VOLUME [0002]

The present invention relates to a data analysis and prediction technique, and more particularly, to a data analysis and prediction technique based on a keyword search amount of an Internet portal site.

Due to recent rapid changes in the global economy, frequent changes in consumption trends, and the emergence of alternative technologies, the cycle of product demand life is shortening, demand fluctuations become uncertain, demand forecasting for stable order quantity forecasting and production planning It is becoming more and more important. Especially, from the viewpoint of corporate management, demand forecasting is also a problem in demand forecasting, but too little demand decision leads to very dangerous result, so accurate demand forecasting is needed.

In this situation, big data analysis technology that attracts consumer interest for forecast future demand with big data analysis technology is getting attention. In this big data analysis technology, text mining technology that catches consumer interest by analyzing posts such as Social Network Service (SNS) is getting attention. However, it is necessary to build a server that acquires, stores, and manages data, and there is a difficulty in analyzing SNS posts that are mostly natural language.

As a result, companies use the keyword search information of Google Trends and Naver Trend, which are open to Internet portal sites such as Google and Naver, to set up marketing strategies based on keyword search volume and search volume change trends. These keyword queries are keyword-centric, which means that they have fewer errors than SNS Big Data, which is atypical, and they have the advantage of being able to exclude corporate advertisements posted on SNS posts.

Korean Patent Laid-Open No. 10-2014-0037384, entitled " Apparatus Demand Forecasting Apparatus and Method Based on Social Network Service Data ", discloses a system for predicting a demand for a commodity based on social network service (SNS) Apparatus and method.

However, Korean Patent Laid-Open Publication No. 10-2014-0037384, however, has to build an SNS server to acquire, store and manage data, and since the SNS post is made up of natural language, valid data must be extracted therefrom, , There is a limit to the accuracy and reliability of the data.

An object of the present invention is to predict data of next period from previous period data simply and quickly based on a keyword search amount of an Internet portal site.

It is another object of the present invention to quickly catch a trend in which keyword search volume is rapidly increasing, and accurately predict next period data by reflecting a trend type to previous period data.

It is another object of the present invention to determine a keyword corresponding to a type of previous period data existing as data of various types, and to predict the next period data by type.

The present invention also aims to apply the present invention to fields requiring various data predictions such as forecasting sales volume, forecasting demand, estimating production volume, forecasting trends, and estimating order quantity, which require forecasting of future data.

According to another aspect of the present invention, there is provided a method of generating a keyword search volume-based prediction information by using a keyword search volume-based prediction information generation apparatus, Classifying the values; Generating search information for each type in which a keyword search amount is collected for each type using keyword search information collected by an Internet portal site; and generating predicted data values at a prediction time from the internal data values using the type- And generating predictive information by calculating the predictive information.

At this time, the internal data may include at least one of an item, a type, an order quantity, a sales quantity, and a selling rate as internal data values.

In this case, the classifying step may include classifying the internal data values by one or more items, performing a second classification for each type corresponding to the detailed classification of the items, And the selling rate, respectively.

In this case, in the generating the search information by type, the keyword search amount may be collected for two or more periods of each keyword of the similar keyword group corresponding to the type.

In this case, the generating the search information by type may generate the search information for each type by adding the keyword search amounts collected for two or more periods of each keyword of the similar keyword group for each type.

The step of generating the search information by type may include calculating a search amount increase / decrease rate using a keyword search amount for two or more periods of the type, and using the search amount increase / decrease rate to determine a trend type at the prediction time It can be judged.

In this case, the step of generating the search information by type may perform a trend line analysis on a keyword search amount for any one period of the type to determine a trend type at the prediction time.

In this case, the step of generating the search information for each type may perform a trend line analysis for a keyword search amount for two or more periods of the type by the two or more periods.

In this case, the generating of the type-specific search information may determine a trend type at the prediction time using the result of performing trend line analysis for each of the two or more periods and the search amount increase / decrease rate.

In this case, the generating of the type-specific prediction information may correct the internal data values for each type using the variation of the internal data values in the first period.

In this case, generating the type-specific prediction information may predict a second variation of the internal data values of the second period using the first variation of the internal data values of the first period.

In this case, the step of generating prediction information by type may calculate the prediction data values at the prediction time point corresponding to the end time of the second period.

In this case, in the generating of the prediction information by type, when the predicted data value of any one type is '0', the predicted data value of the type having the predicted data value of '0' The first search ratio between the different types of keyword search amounts corresponding to the search amount increase rate of the type of " 0 ".

In this case, if the predictive data value of any one of the types is '0', the step of generating the predictive data by using the predictive data value of the other type and the predictive data value of the type &Quot; 0 ". < / RTI >

According to another aspect of the present invention, there is provided a keyword search amount-based prediction information generating apparatus, comprising: a data classifying unit that receives internal data and classifies internal data values by at least one type; A search information generation unit for generating search information for each type of keyword collection amounts collected by the type using the keyword search information collected by the Internet portal site and a search information generation unit for searching the internal data values from the internal data values at the prediction time And a prediction information generation unit that generates prediction information by calculating prediction data values.

The present invention can easily and quickly predict the next period data from the previous period data based on the keyword search amount of the Internet portal site.

In addition, the present invention can quickly catch trends in which the keyword search volume is rapidly increasing, and accurately forecast the next period data by reflecting the trend type to the previous period data.

In addition, the present invention can predict the next period data by type by determining a keyword corresponding to the type with respect to the previous period data existing in various types of data.

In addition, the present invention can be applied to fields requiring various data prediction such as sales volume forecast, demand survey, production forecast, trend investigation, and order quantity forecast.

1 is a block diagram showing a keyword search amount-based prediction information generating apparatus according to an embodiment of the present invention.
FIG. 2 is a flowchart illustrating a keyword search amount-based prediction information generating method according to an embodiment of the present invention.
FIG. 3 is an operation flow chart showing an example of the search information generating step shown in FIG. 2 in detail.
FIG. 4 and FIG. 5 are operation flow charts detailing an example of the prediction information generating step shown in FIG.
6 is a block diagram illustrating a computer system in accordance with an embodiment of the present invention.

The present invention will now be described in detail with reference to the accompanying drawings. Hereinafter, a repeated description, a known function that may obscure the gist of the present invention, and a detailed description of the configuration will be omitted. Embodiments of the present invention are provided to more fully describe the present invention to those skilled in the art. Accordingly, the shapes and sizes of the elements in the drawings and the like can be exaggerated for clarity.

Throughout the specification, when an element is referred to as " comprising ", it means that it can include other elements as well, without excluding other elements unless specifically stated otherwise.

The present invention can perform data analysis and prediction based on keyword search volume of an Internet portal site. That is, the present invention can perform data prediction in various fields such as sales volume forecast, demand survey, production forecast, trend investigation, and order quantity forecast.

Therefore, in the present invention, the design of the order quantity using the in-house sales data of the apparel product is described as an embodiment of the present invention, and the invention is not limited thereto.

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

1 is a block diagram showing a keyword search amount-based prediction information generating apparatus according to an embodiment of the present invention.

Referring to FIG. 1, a keyword search amount based prediction information generating apparatus according to an embodiment of the present invention includes a data classifying unit 110, a search information generating unit 120, and a prediction information generating unit 130.

The data classifier 110 receives the internal data and classifies the internal data values into one or more types.

The search information generation unit 120 may generate search information for each type of keyword search volume by using the keyword search information collected by the Internet portal site.

The prediction information generation unit 130 may generate prediction information by calculating prediction data values at a prediction time point from the internal data values using the type-specific search information.

Details of the data classification unit 110, the search information generation unit 120, and the prediction information generation unit 130 will be described in detail with reference to FIG. 2 to FIG.

FIG. 2 is a flowchart illustrating a keyword search amount-based prediction information generating method according to an embodiment of the present invention. FIG. 3 is an operation flow chart showing an example of the search information generating step shown in FIG. 2 in detail. FIG. 4 and FIG. 5 are operation flow charts detailing an example of the prediction information generating step shown in FIG.

Referring to FIG. 2, in the keyword search amount based prediction information generation method according to an embodiment of the present invention, the data classification unit 110 performs the internal data classification step S210 using the keyword search amount based prediction information generation apparatus The search information generating unit 120 may perform the search information generating step S220 and the prediction information generating unit 130 may perform the prediction information generating step S240.

First, in step S210, the received internal data can be classified.

That is, the step S210 may receive the internal data and classify the internal data values by one or more types.

For example, in the order quantity design using the in-house sales data of the apparel product according to the embodiment of the present invention, the apparel product style can be defined as shown in Table 1

item gender type Product group Style code shirt Woman Flannel shirt KI G01 shirt Woman Flannel shirt B1 G02 shirt Woman Cuff shirt B2 G03 shirt Woman Cuff shirt DE G04

The item is defined as the top division unit of the apparel product. For example, it can correspond to a kind of clothing such as a coat, a T-shirt, a shirt, and a one-to-one man.

The gender may correspond to the gender of the consumer targeting the apparel product.

The type is defined as the detailed division unit of the apparel item for the item. For example, the type of coat may be defined as a single tailored coat, a handmade coat, and the like.

The style code defines the code for one garment item whose design is unique. One type can be defined for multiple styles, and one style can have multiple colors and sizes.

The store rating is defined by the size of the store and the store classification according to the sales scale. In the order of the size and the sales scale, it can be given from A grade to D grade.

Accordingly, the product group can be defined as K1, B1, B2, and DE depending on the store grade.

The KEYBASIC can correspond to the product group (style) distributed to all stores. Basic 1 may correspond to product groups that distribute to stores of A, B, C store class. BASIC 2 may correspond to product groups that distribute to stores in A and B store class. Designs may correspond to product groups that distribute to stores of Class A store ratings.

style Orders sales rate Sales rate One 250 200 80% 2 200 150 75% 3 250 50 20% 4 250 40 16% 5 200 50 25% ... ... ... ...

Table 2 is an example of in-house data, which is a collection of sales information for each style.

In Table 2, the style may correspond to one garment item whose design is unique. The order quantity may correspond to the clothes order quantity of the previous season. Sales volume may correspond to sales volume of previous season apparel merchandise. The selling rate may correspond to the ratio of the sales volume to the order volume.

For example, in the order quantity design using in-house sales data of apparel goods according to an embodiment of the present invention, in-house data can be classified into items and types as shown in Table 3. [

style Orders sales rate Sales rate type item One 250 200 80% A I 2 200 150 75% A I 3 250 50 20% B I 4 250 40 16% B I 5 200 50 25% C I ... ... ... ... ... ...

That is, the internal data according to an embodiment of the present invention may include at least one of an item, a type, an order quantity, a sales quantity, and a selling rate as internal data values.

At this time, the step S210 may perform the first sorting of the internal data values by one or more items, and may perform the second sorting by the type corresponding to the detailed sorting of the items.

For example, in step S210, the internal data classified by items and types can be added for each type as shown in Table 4. [

item type Orders sales rate Sales rate I A 550 350 64% I B 500 90 18% I C 200 50 25%

In addition, the keyword search amount based prediction information generation method according to an embodiment of the present invention can generate search information by type (S220).

That is, the step S220 can generate search information for each type of keyword search amount by the type using the keyword search information collected by the Internet portal site.

For example, Internet portal sites such as Google and Naver provide keyword search information through Google Trends and Naver Trend. Since the keyword search information can be confirmed by anyone, the present invention can collect the keyword search amount of the keyword corresponding to the type using the keyword search information provided by the Internet portal sites such as Google and Naver.

At this time, the keyword search information may include at least one or more information such as a keyword, a search volume, sex, age, region, country, and education level of searchers.

Most consumers browse and compare relevant information through Internet portal site before purchasing products, so analyzing keyword search information on Internet portal site can predict consumer trends.

Furthermore, in the keyword search information of the Internet portal site, it is possible to grasp that factors that are considered to be important when purchasing products of a consumer are changed with time.

Referring to FIG. 3, in operation S220, the type-specific keyword may be first banked (S221).

That is, in step S221, the keyword search amount may be collected for two or more periods for each keyword of the similar keyword group corresponding to the type.

For example, in the order quantity design using the in-house sales data of the apparel product according to the embodiment of the present invention, as shown in Table 5, the similar keyword groups for each type can be listed up.

type Search Keywords One 2 3 4 5 Cotton pants Cotton pants Cotton pants Chino pants Chino pants Cotton Pants A A- A " A * -A AA B B BB B- B B * C C C " * C -C C

That is, since consumers can search for a product type of cotton pants by cotton pants, cotton pants, chino pants, chino pants, or cotton pants on the Internet portal site, the step S221 of the present invention can be applied to a similar keyword You can create a group.

In addition, the step S220 may collect the keyword search amount by type (S222).

That is, the step S222 may generate the search information for each type by summing the keyword search amounts collected for two or more periods of each keyword of the similar keyword group for each type.

For example, in the order quantity design using in-house sales data of the apparel product according to an embodiment of the present invention, keyword search amounts collected for each keyword listed in the similar keyword group as shown in Table 6 can be collected.

type keyword 2015 2016 Increase / decrease rate Cotton pants Cotton pants 300 200 -33% Cotton pants Cotton pants 5000 3000 -40% Cotton pants Chino pants 3000 -33% Cotton pants Chino pants 4000 -13% Cotton pants Cotton Pants 500 -90% A A- 10000 20000 100% A A " 100 200 100% A A * 50 100 100% A -A 5000 5500 10% A AA 600 650 8%

At this time, in step S222, the keyword search amount data of Table 6 can be summed by type as shown in Table 7. [

type Search volume Increase / decrease rate 2015 2016 Cotton pants 12800 8750 -32% A 15750 26450 68% B 30000 10000 -67% C 10000 20000 100%

In addition, the step S220 can calculate the search amount increase / decrease rate by type and the trend type (S233).

That is, in step S223, the search amount increase / decrease rate may be calculated using the keyword search amount for two or more periods for each type, and the trend type at the prediction time point may be determined using the search amount increase / decrease rate.

At this time, the search amount increase / decrease rate can be corrected based on the absolute amount of the keyword search amount in proportion to the increase / decrease rate.

For example, in the order quantity design using in-house sales data of the apparel product according to an embodiment of the present invention, the trend of the next season can be predicted based on the search amount increase / decrease rate by keyword type search amount data as shown in Table 8. [

type Increase / decrease rate Trend type 2015 2016 2015 2016 Cotton pants 17% -32% - Declining trend A 157% 68% Rising Trend Megatrend B 9% -67% Megatrend Declining trend C -24% 100% - Rising Trend D 401% Rising Trend

That is, as shown in Table 8, in step S223, it is possible to determine the trend type using the search magnification ratio of two or more previous periods.

At this time, in step S223, if the search amount increase / decrease rate decreases, the trend type can be determined as the downward trend, and if the search amount increase / decrease rate increases, the trend type can be determined as the upward trend.

In this case, the rising trend may correspond to a case where the search amount is seen to be maintained, and the trend amount is increasing beyond a certain range of the search amount increase / decrease rate.

In this case, the downward trend may correspond to a case in which the search volume is out of a certain range of the search volume increase rate, which is regarded as being maintained, and shows a decrease trend.

At this time, the range can be arbitrarily set by the user according to the brand characteristic. Even with small changes in trends, brands that respond sensitively to customers can set a smaller range.

For example, referring to Table 8, in step S223, if it is determined that the decrease is 30% or more, it is determined to be a 'falling trend', and if it is decreased to 30% or less, We can judge the trend of 2016 of cotton pants as 'declining trend'.

At this time, if the increase rate of search volume suddenly increases, such as the type of C and D, or the rate of change of search volume is not large, it may be judged as a rising trend.

For example, referring to Table 8, if the search volume increases by 100% or more and judges that it is a 'rising trend', step (S223) increases the A trend of 2015 and the 2016 trend of C and D Trend 'can be judged.

In addition, the step S223 can be judged as a megatrend when the absolute amount of the retrieval amount becomes a certain value or more.

For example, referring to Table 7 and Table 8, if the type with a search volume of 25,000 or more is determined to be a megatrend type, step (S223) is a step of converting the 2016 trend of type A and the 2015 trend of type B into megatrend .

At this time, the step S223 may perform a trend line analysis on the keyword search amount for any one period of the type to determine the trend type at the predicted time.

In this case, trend line analysis can use all the trend line analysis techniques used in general graph data.

For example, in step S223, only the search volume for 2015 and 2016 is described in Table 6. However, the more the data on the search volume collection period is added, the more accurate the result of the trend line analysis can be.

At this time, in step S223, it is possible to determine the trend by using at least one information such as the searcher's gender, age, region, country, and education level as well as the search amount in the keyword search information.

Accordingly, in step S223, if the trend line is a rising line according to the trend line analysis, the trend type can be determined as a rising trend, and if the trend line is a falling line, it can be determined as a falling trend.

At this time, in step S223, the rising trend described above can be determined as a rising trend when the rising angle of the trend line is not less than a predetermined angle.

In addition, the step S223 may perform the trend line analysis for the keyword search amount for two or more periods of the type by the two or more periods.

For example, the step S223 may analyze the first trend line from 2012 to 2014 and the second trend line from 2014 to 2016, respectively.

At this time, the step S223 can determine the trend type at the prediction time using the result of performing the trend line analysis for the two or more periods and the search amount increase / decrease rate.

In addition, the step S220 can confirm whether all types of trend types are judged (S224).

In this case, in step S224, if all types of trend types for the keyword search amount data are determined for each period, the search information may be generated (S225).

However, if all types of trend types for the keyword retrieval amount data are not determined for each period, step S224 returns to step S221 to determine the trend type of the keyword retrieval amount data for which the trend type is not determined have.

In addition, the keyword search amount based prediction information generating method according to an embodiment of the present invention can generate the type-specific prediction information (S230).

Referring to FIG. 4, in operation S230, a predicted data value may be calculated in operation S231.

That is, in step S231, prediction data values at a prediction time point may be calculated from the internal data values using the type-specific search information.

In this case, in step S231, the internal data values may be corrected for each type using the change amount of the internal data values in the first period.

At this time, the step S231 may predict the second variation of the internal data values of the second period using the first variation of the internal data values of the first period.

At this time, the step S231 may calculate the prediction data values at the prediction time point corresponding to the end time of the second period.

For example, in the designing of the order quantity of a product that utilizes the intra-company sales data and the portal search quantity according to an embodiment of the present invention, the final sales quantity of the season that has not yet closed can be predicted based on the trend line.

According to the embodiment of the present invention, when the input internal data is still the sales amount during the season and is not the final determined internal data, the total sales amount at the end of the season is calculated through the sales amount graph (PLC) Can be predicted.

First, in step S231, it is possible to confirm whether there is a sales volume graph for each parking in the previous year sales graph (type PLC) for each type in the internal data.

At this time, in step S231, if there is a type PLC, the sales amount of the remaining period according to the last year's weekly sales ratio of the type PLC can be predicted.

If there is no type PLC, step S231 can be determined as a new type, and it can be confirmed whether the PEAK of the item exists within 4 weeks.

At this time, in step S231, if the PEAK exists within 4 weeks, it can be predicted by the item PLC. If there is a PEAK of the item, the gradient changes in the graph without going to the trend. If the present point is PEAK, the upward trend from the past to the PEAK point may not lead to a further increase in sales, but rather a decrease.

Accordingly, if PEAK is present within four weeks, step S231 can multiply the last parking sales volume in which the data exists and the weekly differential sales volume ratio to predict the weekly sales volume at the season closing time.

If the PEAK does not exist within 4 weeks, the final sales amount and the sales rate at the end of the season can be predicted by the trend line of the item PLC.

In this case, if the PEAK does not exist within 4 weeks, the step S231 may reflect the trend for the predetermined period and reflect the trend for the last 6 weeks.

Lastly, in step S231, the final sales amount and the sales ratio at the end of the season can be predicted using the estimated sales ratio of the new store.

At this time, Table 9 shows the calculation of the sales amount at the end of the season in the internal data.

Item Quantity Design (2016) item Reflect new store
Estimated revenue Orders sales rate Sales rate Deadline prediction
reflection
sales rate Deadline prediction
reflection
Sales rate shirt 106% 148,400 78,652 53% 78,899 53%

In addition, in step S230, it is confirmed whether or not all predicted data values are calculated (S234).

At this time, in step S234, if all predicted data values are calculated, it can be checked whether a predicted data value has a value of '0' (S235).

In addition, in step S234, if all prediction data values have not been calculated, the process returns to step S231 to predict the internal data value variation.

In the keyword search amount based prediction information generation method according to an embodiment of the present invention, it is possible to check whether the predicted data value is '0' (S235).

Referring to FIG. 5, in step S235, when the predicted data value is not '0', the predictive data value based on the search increment / decrement rate may be calculated (S236).

In this case, the step S236 may correct the prediction data values for each type using the search magnification ratio.

At this time, the step S236 may increase or decrease the predicted data value in proportion to the search amount increase / decrease rate.

For example, in the order quantity design using in-house sales data of the apparel product according to an embodiment of the present invention, as shown in Table 10, the sales volume can be calculated based on the corrected sales volume corrected for the sales volume through the season- The order quantity can be calculated using the correction increase / decrease rate corrected for the increase / decrease rate.

type In-house sales data
(Internal data) Keyword search volume
data Order quantity
(Prediction data) sales rate Corrected sales volume Increase / decrease rate Correction increase / decrease rate A 570 ① 68% ㉠ ① × (㉠ + 1) B 190 ② -67% ㉡ ② × (㉡ + 1) C 230 ③ 100% ㉢ ③ × (㉢ + 1)

At this time, since the correction increase / decrease rate can not be treated equally as 100% increase in the search amount of 10 million and 100% increase in 100,000, the increase / decrease rate can be corrected based on the search absolute amount.

In this case, the order quantity can be adjusted by further reflecting the store open plan.

If the predicted data value of any one type is '0', the keyword retrieval amount of the type in which the predicted data value is '0' and the retrieval amount of the type in which the predicted data value is '0' A first search ratio between other types of keyword search amounts corresponding to the increase / decrease rate may be calculated (S237).

At this time, if the predicted data value is '0' in step S237, there is no internal data.

For example, in an order quantity design using in-house sales data of a garment product according to an embodiment of the present invention, it may correspond to a new product of a type in which a sales quantity does not exist.

At this time, in step S237, if the prediction data value of any one of the types is '0', the predictive data value of the predictive data value of '0' A prediction data value of a type having a value '0' can be calculated.

At this time, when the predicted data value is '0', it may correspond to the absence of the internal data.

For example, in the order quantity design using in-house sales data of apparel goods according to an embodiment of the present invention, as shown in Table 11, when in-house sales data does not exist, search for a similar or the same type of search quantity increase / The order quantity can be calculated by using the sales quantity increase per quantity.

type In-house sales data Keyword search volume data Order quantity
(Prediction data) sales rate correction
sales rate 2015 2016 Increase / decrease rate correction
Increase / decrease rate A 570 ① 15750 26450 68% ㉠ ① × (㉠ + 1) B 190 ② 30000 10000 -67% ㉡ ② × (㉡ + 1) C 230 ③ 10000 20000 100% ㉢ ③ × (㉢ + 1) D - - 20000 40000 100% ㉣ x

That is, since the in-house sales data does not exist in the D type in Table 11, the order quantity should be predicted based on the keyword search quantity data.

The increase / decrease rate of type D is 100%, and the type with the same increase / decrease rate is the type C.

At this time, since the first search ratio described above is a ratio between the search amount of the C type and the search amount of the D type, it is found that the first search ratio is 100%.

At this time, if the C type sales amount is equal to the sales amount, the order quantity of the C type which is the predicted data value of the other type described above may be 460 which is 100% increased from the sales quantity.

Therefore, the order quantity of type D can be 920, which increases the first search ratio 100% from the order quantity 460 of the C type.

In addition, in step S230, it is confirmed whether or not all predicted data values are calculated (S239).

That is, in step S239, if all the prediction data values are calculated, the prediction information can be generated (S240). If there is a type in which the prediction data value is not calculated, the process returns to step S235, . ≪ / RTI >

As described above, according to the embodiment of the present invention, if the change trend can be grasped through the keyword search information of the Internet portal site, it is possible to solve the over-ordering problem such as the case where the sales rate of the product is good and the quantity is further increased, . This problem can be solved by failing to design the quantity of the type which is the current mega trend but gradually declining trend as it is based on the sales volume of previous season and deciding the order quantity next season.

6 is a block diagram illustrating a computer system in accordance with an embodiment of the present invention.

Referring to FIG. 6, embodiments of the present invention may be implemented in a computer system 1100, such as a computer readable recording medium. 6, the computer system 1100 includes one or more processors 1110, a memory 1130, a user interface input device 1140, a user interface output device 1150, And storage 1160. In addition, the computer system 1100 may further include a network interface 1170 connected to the network 1180. The processor 1110 may be a central processing unit or a semiconductor device that executes the processing instructions stored in the memory 1130 or the storage 1160. Memory 1130 and storage 1160 can be various types of volatile or non-volatile storage media. For example, the memory may include ROM 1131 or RAM 1132.

That is, the keyword search volume-based prediction information generation method of the present invention can be stored in the storage 1160 in the form of software and implemented through the computer system 1100.

As described above, the apparatus and method for generating keyword-based search amount-based prediction information according to the present invention are not limited to the configurations and methods of the embodiments described above, but the embodiments can be modified in various ways All or some of the embodiments may be selectively combined.

110: data classifier 120: search information generator
130: prediction information generating unit
1100: Computer system 1110: Processor
1120: bus 1130: memory
1131: ROM 1132: RAM
1140: User interface input device
1150: User interface output device
1160: Storage 1170: Network Interface
1180: Network

Claims (16)

A method of using a keyword search amount-based prediction information generating apparatus,
Receiving internal data and classifying internal data values by one or more types;
Generating search information for each type in which the keyword search amount is collected for each type using the keyword search information collected by the Internet portal site; And
Generating prediction information by calculating prediction data values at a prediction time point from the internal data values using the type-specific search information;
Lt; / RTI >
The step of generating the prediction information
Calculating a predictive data value for each type by increasing or decreasing an internal data value for each type by a search amount increase / decrease rate of a keyword search amount for the first period collected for each type,
A search amount increase / decrease rate corresponding to a search amount increase / decrease rate of the keyword search amount during the first period of the first type in which at least one of the internal data values of the internal data does not exist, Searching for the second type in which the predicted data value is calculated,
A search ratio that is a rate of change between the keyword search amount of the first type and the keyword search amount of the second type during the first period is calculated and the search ratio is calculated using the search ratio and the second type of predicted data value, One type of prediction data value is calculated,
The step of generating the prediction information
Determining a second type as a search type having a search amount increase / decrease rate most similar to a search amount increase / decrease rate of a keyword search amount for the first period of the first type,
Calculating a ratio corresponding to a difference in the keyword search amount of the second type with respect to the keyword search amount of the first type, determining a ratio corresponding to the difference as the search ratio,
And a value obtained by increasing or decrementing the prediction data value of the second type by the search ratio is calculated and determined as the prediction data value of the first type. The method according to claim 1,
The internal data
Wherein the at least one of the item, the type, the order quantity, the sales quantity, and the sales rate is included as internal data values. The method of claim 2,
The classifying step
Performing a first classification on the internal data values by one or more items, performing a second classification on the basis of the types corresponding to the detailed classification on the items, summing the order quantity, the sales quantity, And generating a keyword search amount based prediction information. The method according to claim 1,
The step of generating the search information by type
Wherein the keyword search volume is collected for at least two periods of each keyword of the similar keyword group corresponding to the type. The method of claim 4,
The step of generating the search information by type
Wherein the search information for each type is generated by summing the keyword search amounts collected for two or more periods of each keyword of the similar keyword group for each type. The method of claim 5,
The step of generating the search information by type
Calculating a search amount increase / decrease rate using a keyword search amount for two or more periods for each type, and determining a trend type at the prediction time point using the search amount increasing / decreasing rate Way. The method of claim 6,
The step of generating the search information by type
Wherein the trend type analysis unit determines trend types at the predicted point by performing a trend line analysis on a keyword search amount for any one period of the type. The method of claim 7,
The step of generating the search information by type
Wherein a trend line analysis for a keyword search amount for two or more periods of the type is performed for each of the two or more periods. The method of claim 8,
The step of generating the search information by type
Wherein the trend type at the predicted time point is determined using the result of performing the trend line analysis for each of the two or more periods and the search amount increase / decrease rate. The method of claim 9,
The step of generating the type-specific prediction information includes:
Wherein the internal data values are corrected for each type by using a change amount of the internal data values in the first period. The method of claim 10,
The step of generating the type-specific prediction information includes:
And estimating a second variation of the internal data values of the second period using the first variation of the internal data values of the first period. The method of claim 11,
The step of generating the type-specific prediction information includes:
And calculating the predicted data values at the predicted time corresponding to the end of the second period. The method of claim 12,
The step of generating the type-specific prediction information includes:
And the predictive data values are corrected for each type using the search magnification change rate. delete delete A data classifier for receiving internal data and classifying internal data values by at least one type;
A search information generating unit for generating search information for each type of keyword collection amount collected by the type using the keyword search information collected by the Internet portal site; And
A prediction information generation unit for generating prediction information by calculating prediction data values at a prediction time point from the internal data values using the type-specific search information;
Lt; / RTI >
The prediction information generation unit
Calculating a predictive data value for each type by increasing or decreasing an internal data value for each type by a search amount increase / decrease rate of a keyword search amount for the first period collected for each type,
A search amount increase / decrease rate corresponding to a search amount increase / decrease rate of the keyword search amount during the first period of the first type in which at least one of the internal data values of the internal data does not exist, Searching for the second type in which the predicted data value is calculated,
A search ratio that is a rate of change between the keyword search amount of the first type and the keyword search amount of the second type during the first period is calculated and the search ratio is calculated using the search ratio and the second type of predicted data value, One type of prediction data value is calculated,
The prediction information generation unit
Determining a second type as a search type having a search amount increase / decrease rate most similar to a search amount increase / decrease rate of a keyword search amount for the first period of the first type,
Calculating a ratio corresponding to a difference in the keyword search amount of the second type with respect to the keyword search amount of the first type, determining a ratio corresponding to the difference as the search ratio,
And a value obtained by increasing or decrementing the predictive data value of the second type by the search ratio is calculated and determined as the predictive data value of the first type.

Images