KR102653483B1 - Method of predicting price of artwork based on artificial intelligence - Google Patents

Abstract

According to an embodiment of the present invention, a method for predicting the price of a work of art includes, for a first work of art sold through an auction, the first artist corresponding to the first work of art, the size of the first work of art, and the size of the first artist. Obtaining sales data including survival status, auction date, and successful bid price, calculating a media index corresponding to the first author and the auction date, a plurality of predetermined groups of the first author on the auction date Calculating an interest index in an online community - the plurality of groups include a first group, a second group, and a third group, and each of the first group, the second group, and the third group includes one or more online and a community, wherein the interest index of the first writer in the first group is the number of times the first writer was mentioned in the online community included in the first group during a first predetermined period before the auction date. It is a value divided by the total number of postings published in the online community included in the first group during the first period -, size of the first artwork, survival of the first artist, the media index, and the first artist's Associating an association factor including an interest index in a group, an interest index in the second group of the first author, and an interest index in the third group of the first author with the successful bid and storing it in a learning database. A step of generating a first artificial intelligence model that predicts the price of a work of art based on the learning database, and inputting an association factor for a second work of art into the first artificial intelligence model to determine the expected price of the second work of art. It may include an acquisition step.

Description

{METHOD OF PREDICTING PRICE OF ARTWORK BASED ON ARTIFICIAL INTELLIGENCE}

The present invention relates to a method for predicting the price of artwork based on artificial intelligence.

Artificial Intelligence refers to a computer system with the functions of human intelligence, and refers to the artificial implementation of human intelligence in machines, etc. Machine learning, or machine learning, is a field of artificial intelligence and refers to a technology in which a computer learns to create a model that can produce appropriate output to solve problems. Deep Learning refers to a set of machine learning algorithms that attempt a high level of abstraction (the task of extracting only key content from a large amount of complex data) through a combination of several non-linear transformation techniques.

Various attempts have been made to predict the price of art works by applying artificial intelligence, but the reality is that they are not adequately reflecting the demand of the diversifying art market.

Traditionally, the vast majority of consumers in the art market were art experts, including professional collectors. On the other hand, in recent years, with the increase in national income, consumers who want to purchase art as an extension of appreciating art as a hobby, and with the rise of art tech, consumers who aim to make profit by selling art rather than appreciating art, existing art experts and The proportion of consumers participating in the purchase of art from different perspectives is increasing.

There has been no research yet on the extent to which the interest of these new types of art consumers leads to actual purchases and affects the prices of art works. Therefore, the interest of new types of art consumers is not taken into consideration when predicting the price of art works using artificial intelligence. There are limits to the accuracy of predictions.

The purpose of the present invention is to provide an artificial intelligence-based art price prediction method that can increase accuracy by considering the interests of various types of art consumers.

According to an embodiment of the present invention, a method for predicting the price of a work of art includes, for a first work of art sold through an auction, the first artist corresponding to the first work of art, the size of the first work of art, and the size of the first artist. Obtaining sales data including survival status, auction date, and successful bid price, calculating a media index corresponding to the first author and the auction date, a plurality of predetermined groups of the first author on the auction date Calculating an interest index in an online community - the plurality of groups include a first group, a second group, and a third group, and each of the first group, the second group, and the third group includes one or more online and a community, wherein the interest index of the first writer in the first group is the number of times the first writer was mentioned in the online community included in the first group during a first predetermined period before the auction date. It is a value divided by the total number of postings published in the online community included in the first group during the first period -, size of the first artwork, survival of the first artist, the media index, and the first artist's Associating an association factor including an interest index in a group, an interest index in the second group of the first author, and an interest index in the third group of the first author with the successful bid and storing it in a learning database. A step of generating a first artificial intelligence model that predicts the price of a work of art based on the learning database, and inputting an association factor for a second work of art into the first artificial intelligence model to determine the expected price of the second work of art. It may include an acquisition step.

According to an embodiment of the present invention, the method for predicting the price of an artwork is to classify Internet communities with common tendencies of art consumers into one group, calculate the level of interest in the artist in various groups, and reflect this in the price prediction. By doing so, it can contribute to increasing the accuracy of prediction.

In addition, the method according to an embodiment of the present invention contributes to conveniently adding new Internet communities to the group in line with rapidly changing Internet community trends by providing recommended groups for new Internet communities where the group is not defined. You can.

1 illustrates an art price prediction system according to various embodiments of the present invention.
Figure 2 shows components of a server included in an art price prediction system according to various embodiments of the present invention.
Figure 3 illustrates a method performed on a server included in an art price prediction system according to various embodiments of the present invention.
Figure 4 shows an example of groups into which online communities are classified in a method according to various embodiments of the present invention.
Figure 5 shows data input to an artificial intelligence model and data output from the artificial intelligence model in methods according to various embodiments of the present invention.

As used herein, “comprises” and/or “comprising” does not exclude the presence or addition of one or more other elements in addition to the mentioned elements. The various embodiments of this specification and the terms used herein are not intended to limit the technical features described in this specification to specific embodiments, but should be understood to include various changes, equivalents, or replacements of the relevant embodiments. In connection with the description of the drawings, similar reference numbers may be used for similar or related components. The singular form of a noun corresponding to an item may include one or more of the above items, unless the relevant context clearly indicates otherwise. As used herein, “A or B”, “at least one of A and B”, “at least one of A or B”, “A, B or C”, “at least one of A, B and C”, and “A Each of phrases such as “at least one of , B, or C” may include any one of the items listed together in the corresponding phrase, or any possible combination thereof. Terms such as "first", "second", or "first" or "second" may be used simply to distinguish one element from another, and may be used to distinguish such elements in other respects, such as importance or order) is not limited. When one (e.g. first) component is said to be "connected" to another (e.g. second) component, with or without the terms "functionally" or "communicatively", it means This means that the certain component can be connected to the other component directly (eg, wired), wirelessly, or through a third component.

Various embodiments of the present specification may be implemented as software including one or more instructions stored in a storage medium (eg, internal memory or external memory) that can be read by a machine. For example, the processor of the device may call at least one instruction among one or more instructions stored from a storage medium and execute it. This allows the device to be operated to perform at least one function according to the at least one instruction called. The one or more instructions may include code generated by a compiler or code that can be executed by an interpreter. A storage medium that can be read by a device may be provided in the form of a non-transitory storage medium. Here, 'non-transitory' only means that the storage medium is a tangible device and does not contain signals (e.g. electromagnetic waves), and this term refers to cases where data is semi-permanently stored in the storage medium. There is no distinction between temporary storage cases.

According to one embodiment, methods according to various embodiments disclosed in this specification may be provided and included in a computer program product. Computer program products are commodities and can be traded between sellers and buyers. The computer program product may be distributed in the form of a machine-readable storage medium (e.g. compact disc read only memory (CD-ROM)) or via an application store (e.g. Play Store ^TM ) or on two user devices (e.g. It can be distributed (e.g. downloaded or uploaded) directly between smart phones) or online. In the case of online distribution, at least a portion of the computer program product may be at least temporarily stored or temporarily created in a machine-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

According to various embodiments, each component (e.g., module or program) of the above-described components may include a single or plural entity, and some of the plurality of entities may be separately placed in other components. there is. According to various embodiments, one or more of the components or steps described above may be omitted, or one or more other components or steps may be added. Alternatively or additionally, multiple components (eg, modules or programs) may be integrated into a single component. In this case, the integrated component may perform one or more functions of each component of the plurality of components in the same or similar manner as those performed by the corresponding component of the plurality of components prior to the integration. . According to various embodiments, steps performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically, or one or more of the steps may be executed in a different order, or omitted. Alternatively, one or more other steps may be added.

1 illustrates an art price prediction system 100 according to various embodiments of the present invention. Referring to FIG. 1, the art price prediction system 100 may include a server 110 and a user device 120.

The server 110 may receive sales data from the user device 120 regarding a number of works of art that have previously been sold through auction. The server 110 may derive association factors related to artworks based on sales data, associate the derived association factors with the successful bid price, and store them in a learning database. The server 110 may create a first artificial intelligence model that predicts the price of an artwork based on a learning database. Thereafter, the server 110 obtains information about the second work of art for which price prediction is required from the user device 120, derives an association factor about the second artwork based on the obtained information, and uses the derived association factor. The expected price of the second work of art can be obtained by inputting it into the first artificial intelligence model, and the expected price of the second work of art can be transmitted to the user device 120.

According to various embodiments, the user device 120 may be an electronic device used by an administrator of the server 110. According to various embodiments, the user device 120 may be a consumer's electronic device that uses an art price prediction service provided by the manager of the server 110.

Figure 2 shows components of a server included in an art price prediction system according to various embodiments of the present invention. Referring to FIG. 2 , the server 110 may include a communication circuit 210, a processor 220, and a memory 230. The communication circuit 210 can transmit information to or receive information from another electronic device, and the type of communication supported by the communication circuit 210 is not limited.

The processor 220 performs an operation based on data received through the communication circuit 210 and/or data stored in the memory 230, and sends at least a portion of the result of the operation to another electronic device through the communication circuit 210. It can be transmitted to or stored in the memory 230.

The processor 220 may include a data learning unit 221 and a data recognition unit 222. The data learning unit 221 may create an artificial intelligence model for predicting the price of an artwork. The data recognition unit 222 may preprocess sales data and provide the preprocessed data to the data learning unit 221 for learning.

At least one of the data learning unit 221 and the data recognition unit 222 is implemented as a dedicated hardware chip for artificial intelligence, or as part of an existing general-purpose processor (e.g., AP or CPU) or a graphics-specific processor. It could be.

According to various embodiments, unlike the data learning unit 221 and the data recognition unit 222 shown in FIG. 2 as being included in the server 110, the data learning unit 221 and the data recognition unit 222 are Each may be mounted on a separate electronic device.

In this case, the data learning unit 221 and the data recognition unit 222 are connected to each other by wire or wirelessly, and the model information generated by the data learning unit 221 is provided to the data recognition unit 222, or the data recognition unit 222 The data input at 222 may be provided to the data learning unit 221 as additional learning data.

At least one of the data learning unit 221 and the data recognition unit 222 may be implemented as a software module. In this case, the software module may be stored in a non-transitory computer-readable recording medium. At least some of the software modules may be provided by an operating system (OS) or may be provided by a predetermined application.

Figure 3 illustrates a method performed on a server included in an art price prediction system according to various embodiments of the present invention.

In step 310, the server 110 may obtain sales data about the first artwork sold through auction from the user device 120. Sales data for the first work of art may include the first artist corresponding to the first work of art, the size of the first work of art, whether the first artist is alive, the auction date, and the winning bid. According to various embodiments, sales data for the first artwork may further include exhibition history information of the first artist. The exhibition history information of the first artist may include a list of the period during which the exhibition was held and the name of the art museum where the exhibition was held.

In step 320, the server 110 may calculate a media index corresponding to the artist and auction date of the first artwork and an interest index in the online community of a plurality of predetermined groups.

According to various embodiments, the media index may be calculated in proportion to the number of articles written before the auction date among a plurality of articles included in the search result for the first author. The server 110 may query an external search engine for the first author and obtain a plurality of articles and the creation date of each article as a search result.

In this specification, an online community can be defined as any space where people gather in the Internet space. For example, the content of a specific channel on YouTube and a set of comments on the content can be defined as one online community. For example, an entire specific site may be defined as one online community, or a specific bulletin board on a specific site may be defined as one online community.

The plurality of predetermined groups include a first group, a second group, and a third group, and each of the first group, the second group, and the third group includes one or more online communities. An example of an online community of a plurality of predetermined groups is shown in Figure 4. Referring to Figure 4, the first group includes online communities such as online community A1, online community A2, and online community A3, and the second group includes online communities such as online community B1, online community B2, and online community B3. And the third group includes online communities such as online community C1, online community C2, and online community C3. Which group to classify various online communities into may be predefined according to an administrator's input through the user device 120. According to one embodiment, the manager classifies a community where consumers who are interested in art tech gather as a first group, a community where consumers who appreciate art as a hobby gather as a second group, and professional collectors, art critics, etc. The community where art experts gather can be classified into the third group. Initially, when the administrator designates an online community included in each group, the server 110 may determine a recommended group and provide it to the user device 120 for a new online community for which no group is specified, using a method described later. there is.

The index of interest in the first group of the first author is the number of times the first author is mentioned in the online community included in the first group during a predetermined first period before the auction date. It can be defined as the value divided by the total number of posts posted in the online community. For example, the first period may be defined as 6 months. For example, the interest index in the first group of the first author may be expressed in a format such as once per 100 postings.

In step 330, the server 110 may store the association factor for the first artwork in the learning database by associating it with the successful bid price. The association factors for the first work of art are the size of the first work of art, whether the first artist is alive or not, the media index, the index of interest in the first group of the first artist, the index of interest in the second group of the first artist, and the index of interest in the second group of the first artist. 1 May include an index of the author's interest in third groups. Among these, the media index, the interest index in the first group of the first author, the interest index in the second group of the first author, and the interest index in the third group of the first author are values calculated in step 320.

According to various embodiments, when sales data regarding a first work of art includes exhibition history information of the first artist, the association factors for the first work of art include the number of exhibitions by the first artist and the museum in which the first artist's exhibition was held. It may further include a first score calculated based on . Here, the first score can be defined as the average of the weight multiplied by the art museum score corresponding to one or more art museums where the first artist's exhibition was held. The art museum score may be calculated by referring to the art museum scores stored in advance in the memory of the server 110 and stored corresponding to a plurality of art museums. The server 110 may use a fixed value for the art museum score by inputting it through the user device 120, or may manage the art museum score by regularly updating it according to the level of interest in the online community. When regularly updating the art museum score according to the level of interest in the online community, the server 110 determines the interest index in the first group, the interest index in the second group, and the first art museum included in the plurality of art museums. The score of the art museum corresponding to the first art museum can be calculated in proportion to the sum of the interest indices in the three groups. The index of interest in the first group for the first art museum is the index of interest in the first group for the first art museum in the online community included in the first group during the time period before the first predetermined period (e.g., 6 months) from the time of calculating the art museum score. It can be defined as the number of times mentioned divided by the total number of postings made in the online community included in the first group during the first period. The weight may be determined based on when the exhibition was held. For example, the server 110 may store in memory a method for determining the weight, such that the more recently the exhibition was held based on the auction date, the higher the weight.

In step 340, the server 110 may create a first artificial intelligence model that predicts the price of an artwork based on the learning database. The learning method for creating the first artificial intelligence model is not limited. For example, the first artificial intelligence model may be created using various machine learning techniques. For example, at least one of a Recurrent Neural Network (RNN), a Convolution Neural Network (CNN), an Artificial Neural Network (ANN), and a Transformer model may be used for learning to create the first artificial intelligence model.

Data input to the artificial intelligence model and data output from the artificial intelligence model are shown in Figure 5. Referring to FIG. 5, correlation factors related to the artwork requiring price prediction are input to the first artificial intelligence model, and the first artificial intelligence model can output the expected price of the artwork. Notwithstanding the example in Figure 5, the correlation factor may further include the artist's interest index in more than three groups, or may be calculated based on the number of artist's exhibitions described above and the art museum where the artist's exhibitions are held. 1 more point can be included.

In step 350, the server 110 may obtain an expected price of the second work of art by inputting an association factor for the second work of art into the first artificial intelligence model.

The server 110 obtains sales data for the second work of art from the user device 120 similarly to step 310 in order to obtain an association factor for the second work of art, and, similar to step 320, determines the media index and a plurality of predetermined pieces of art. The interest index in the group's online community can be calculated. However, since the auction has already occurred in step 320, unlike calculating the media index and interest index based on the auction date, the server 110 can calculate the media index and interest index based on the time of performing step 350. . That is, the media index included as an association factor for the second work of art in step 350 is the number of articles written before the date on which step 350 is performed among a plurality of articles included in the search result for the second author corresponding to the second work of art. It can be calculated in proportion to . Likewise, the index of interest in the first group of the second artist included as an association factor for the second artwork is the It can be defined as the number of times an author is mentioned divided by the total number of postings made in the online community included in the first group during the first period.

According to various embodiments, the server 110 may recommend to the user device 120 which group to classify the first online community that is not classified into one of a plurality of predetermined groups.

The server 110 may obtain the URL of the first online community that is not classified into one of a plurality of predetermined groups. The server 110 may calculate an interest index in the first group, an interest index in the second group, and an interest index in the third group for all authors included in the learning database. At this stage, the interest index of a specific third writer in the first group is determined by the first group during the time period from before the first period, which is predetermined based on the time of calculating the interest index, to the time of calculating the interest index. It may be defined as the number of times a third author is mentioned in the included online community divided by the total number of postings made in the online community included in the first group during the first period. The server 110 may calculate a comprehensive interest index that is the sum of the interest index in the first group, the interest index in the second group, and the interest index in the third group for all writers included in the learning database. . For example, the interest index of a particular third author in the first group is 1 time per 100 postings, the interest index in the second group is 1 time per 200 postings, and the interest index in the 3rd group is 1 time per 200 postings. In the case of 1 time, the overall interest index of the third author is (1/100)+(1/200)+(1/200)=1/50, so it can be 1 time per 50 postings. In calculating the comprehensive interest index, the size of the online community included in each group is not considered.

Thereafter, the server 110 creates a list of at least one writer whose value obtained by dividing the interest index in the first group by the overall interest index is greater than or equal to a predetermined first value. You can check. here inside refers to each author. For example, if the predetermined first value is 0.6, the third writer in the above-mentioned example is added to the list because the interest index in the first group divided by 1/100 by the overall interest index 1/50 is less than the first value. not included. List Writers included are writers who are mentioned relatively more frequently in the online communities of the first group than in the online communities of other groups. Similarly, the server 110 provides a list of at least one writer whose interest index in the second group divided by the overall interest index is greater than or equal to the first value. and a list of at least one writer whose interest index in the third group divided by the overall interest index is greater than or equal to the first value. You can check.

Afterwards, the server 110 operates in the first online community, interest index can be calculated. Writer Eun As an index of interest in the first online community for, the writer in the first online community during the time period from before the first period predetermined based on the time of calculating the interest index to the time of calculating the interest index. It can be defined as the number of times mentioned divided by the total number of postings published in the first online community during the first period. Similarly, the server 110 is a writer in the first online community, interest index , and author interest index can be calculated.

Server 110 is inside The average value of , inside The average value of , and inside The average value of The group corresponding to the largest value may be determined as the recommended group for the first online community. for example, , , and middle If this value is the highest, the server 110 may determine the first group as a recommended group for the first online community. The server 110 may transmit the determined recommendation group to the user device 120. Thereafter, the server 110 may include the first online community in a group designated according to an input from the user device 120 among the plurality of groups.

Above, embodiments of the present invention have been described with reference to the attached drawings, but those skilled in the art will understand that the present invention can be implemented in other specific forms without changing its technical idea or essential features. You will be able to understand it. Therefore, the embodiments described above should be understood in all respects as illustrative and not restrictive.

Claims (5)

In the method for predicting the price of a work of art, performed by the server 110,
For a first artwork sold through an auction, sales data including the first artist corresponding to the first artwork, the size of the first artwork, whether the first artist is alive, the auction date, and the successful bid are stored on the user device 120. ) step of obtaining from,
calculating a media index corresponding to the first artist and the auction date;
Calculating an interest index in an online community of a plurality of predetermined groups of the first artist on the auction date, wherein the plurality of groups include a first group, a second group, and a third group, and the first group , the second group, and the third group each include one or more online communities, and the interest index of the first author in the first group is determined by the first group for a predetermined first period of time before the auction date. The number of times the first author is mentioned in the included online community divided by the total number of postings made in the online community included in the first group during the first period -,
The size of the first artwork, whether the first artist is alive, the media index, the interest index of the first artist in the first group, the interest index of the first artist in the second group, and the first artist. 1. A step of associating an association factor including an artist's index of interest in the third group with the winning bid and storing it in a learning database;
Generating a first artificial intelligence model that predicts the price of an artwork based on the learning database,
Obtaining an expected price of the second work of art by inputting a related factor for the second work of art into the first artificial intelligence model;
A list of at least one writer whose interest index in the first group divided by the overall interest index is greater than or equal to a predetermined first value

Confirming - the comprehensive interest index is the sum of the interest index in the first group, the interest index in the second group, and the interest index in the third group -,
A list of at least one writer whose interest index in the second group divided by the overall interest index is greater than or equal to the first value

Steps to check,
A list of at least one writer whose interest index in the third group divided by the overall interest index is greater than or equal to the first value

Steps to check,
Confirming the URL of a first online community that is not classified into one of the plurality of predetermined groups,
In the first online community above, the writer interest index , author interest index , and author interest index A step of calculating, and
remind inside The average value of , remind inside The average value of , and above inside The average value of determining the group corresponding to the largest value as a recommended group for the first online community.
Method, including. According to paragraph 1,
The method wherein the media index corresponding to the first author and the auction date is proportional to the number of articles written before the auction date among a plurality of articles included in a search result for the first author. delete According to paragraph 1,
The method further includes storing an art museum score corresponding to a plurality of art museums,
The sales data regarding the first artwork further includes exhibition history information of the first artist,
The association factor for the first work of art further includes a first score calculated based on the number of exhibitions of the first artist and the museum in which the exhibition of the first artist is held,
The first score is an average of the scores of the art museum corresponding to one or more art museums where the exhibition of the first artist was held multiplied by a weight,
The method wherein the weight is determined based on when the exhibition was held. According to clause 4,
Corresponds to the first art museum based on the sum of the interest index in the first group, the interest index in the second group, and the interest index in the third group for the first art museum included in the plurality of art museums. A method further comprising the step of calculating an art museum score.