KR20160043601A - Device and method for recommendationing digital content - Google Patents

Abstract

A smart terminal for recommending a digital content according to an embodiment of the present invention includes: a memory where a program to recommend the digital content is stored; a database where a subject classification tree including more than one category classified in advance is stored; and a processor to execute the program. At this time, the processor collects the digital content stored in the smart terminal by the execution of the program, and extracts a category corresponding to the digital content collected in the subject classification tree stored in the database, and classifies the collected digital content on the basis of the extracted category, and generates an index of the classified digital content, and recommends the digital content on the basis of the generated index. According to the present invention, it is possible to provide the digital content according to the request of a user more accurately.

Description

TECHNICAL FIELD [0001] The present invention relates to a digital content recommendation terminal and a digital content recommendation terminal,

The present invention relates to a digital content recommendation terminal and method.

A recommendation service is a service that provides information corresponding to a user's interest from information provided in an excessive amount. This service analyzes the rating or preference of the content in the history of the user to generate a user profile and performs recommendation of the content based on the profile.

In particular, recently, as users have been rapidly increasing in various terminals such as computers, smart phones and tablet PCs, a user-customized digital content-based recommendation technology has been developed for various terminals. A customized digital content based recommendation technology for various terminals automatically infer user preferences from the terminals and recommend digital contents to the users based on the interest of the inferred users.

However, unlike a general computing system such as a server or a PC, a smart terminal such as a smart phone supports a low-end computing environment and has a drawback in that it is costly to perform network communication. Therefore, digital contents recommendation service for various terminals should be able to classify and recommend user 's preference information based on minimum hardware resource and minimum network communication.

Conventional technologies for recommending customized digital contents to a user in a smart terminal include a technology for automatically inferring a user's interest in a smart terminal and a technique for recommending digital contents to a user based on the interests of the inferred user . In this conventional technology, in extracting a user's interest and selecting digital contents to be recommended, a category in a subject classification tree generated by processing public directory data is used.

For example, if "soccer" is extracted as a user's interest by analyzing the digital content inside the user's smart device, the digital content deduced in the same category is recommended to the user. However, if the recommended digital content is not correctly classified, then digital content that is not actually associated with "soccer" is recommended, thereby not recommending digital content that meets user interests.

The conventional invention for recommending digital contents is as follows.

Korean Patent Registration No. 2013-0049253 (entitled "Content Recommendation Service Method") discloses a method of collecting history information based on context information when a user using a smart terminal is used and providing customized recommendation contents Lt; / RTI > Specifically, the present invention analyzes user's preference information based on context information such as the time, weather, season, and anniversary of the user when the content is used and services the content.

In addition, Springer's foreign paper ("Efficient Recommendation for Smart TV Contents") analyzes users' preference information in smart TV environment and provides digital TV contents to smart TV users by using preference information Lt; / RTI > Specifically, the method includes generating a user profile using category information of digital content purchased by a user in the past, extracting information of another user similar to the profile of the user based on a collaborative filtering method, Is recommended.

It is an object of the present invention to provide a smart terminal and method for classifying categories of digital contents and recommending digital contents based on the classified categories.

It should be understood, however, that the technical scope of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided a smart terminal for recommending digital contents, comprising: a memory for storing a program for recommending digital contents; A database for storing the tree, and a processor for executing the program. At this time, the processor collects the digital contents stored in the smart terminal, extracts the category corresponding to the digital contents collected in the subject classification tree stored in the database, and stores the collected digital contents based on the extracted category Generates an index of the classified digital contents, and recommends the digital contents based on the generated index.

According to another aspect of the present invention, there is provided a digital content recommendation method for a smart terminal, comprising: collecting digital content stored in a smart terminal; Extracting a category corresponding to digital contents collected in a subject classification tree stored in a database; Classifying the collected digital content based on the extracted category; Generating an index of the classified digital contents; And recommending digital content based on the generated index. At this time, the subject classification tree includes one or more categories classified in advance.

According to any one of the above-mentioned means for solving the problems, the category is extracted based on the digital content used by the user, and the recommendation is performed based on the extracted category, so that the digital content can be more accurately provided to the user's request.

In addition, since the recommendation is performed based on the topic classification tree classified by the category stored in the smart terminal for digital contents recommendation and the word frequency of the digital contents, recommendation can be performed accurately and quickly with less computing resources and network traffic.

1 is a configuration diagram of a smart terminal according to an embodiment of the present invention.
2 is an exemplary diagram of a subject classification tree according to an embodiment of the present invention.
3 is a block diagram of a program for recommending digital contents according to an embodiment of the present invention.
4 is a flowchart illustrating a method of recommending digital contents in a smart terminal according to an exemplary embodiment of the present invention.
5 is a flowchart illustrating a method of extracting categories of digital contents in a smart terminal according to an embodiment of the present invention.
6 is a flowchart illustrating a method of generating indexes of digital contents in a smart terminal according to an exemplary embodiment of the present invention.
7 is a flowchart illustrating a method of recommending digital content in a smart terminal according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly explain the present invention in the drawings, parts not related to the description are omitted.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between . Also, when a part is referred to as "including " an element, it does not exclude other elements unless specifically stated otherwise.

Hereinafter, a smart terminal 100 for recommending digital contents 140 according to an embodiment of the present invention will be described with reference to FIGS. 1 to 3. FIG.

1 is a configuration diagram of a smart terminal 100 according to an embodiment of the present invention.

Referring to FIG. 1, a smart terminal 100 recommends digital content 140 based on digital content 140 stored in a smart terminal 100. At this time, the smart terminal 100 includes a memory 110, a database 120, and a processor 130.

The smart terminal 100 may be implemented as a computer or a portable terminal. At this time, the computer includes, for example, a notebook, a desktop, a laptop, a tablet PC, a slate PC, and the like, each of which is equipped with a WEB Browser. (PDS), a Personal Digital Assistant (PDA), an International Mobile Telecommunication (IMT), and a Personal Digital Assistant (PDS) -2000, Code Division Multiple Access (CDMA) -2000, W-CDMA (W-CDMA), WiBro (Wireless Broadband Internet) terminals, smart phones Of wireless communication devices.

The digital content 140 also includes video, pictures, images, music, social network services (SNS), bookmarks, applications, short message services can do.

The memory 110 stores a program 150 for recommending the digital contents 140. [ Here, the memory 110 is collectively referred to as a non-volatile storage device that keeps stored information even when power is not supplied. For example, the memory 110 may be a compact flash (CF) card, a secure digital (SD) card, a memory stick, a solid-state drive (SSD) A magnetic computer storage device such as a NAND flash memory, a hard disk drive (HDD) and the like, and an optical disc drive such as a CD-ROM, a DVD-ROM, etc. .

The program 150 for recommending the digital contents 140 stored in the memory 110 may be implemented in hardware such as software or an FPGA (Field Programmable Gate Array) or ASIC (Application Specific Integrated Circuit) Roles can be performed.

On the other hand, the database 120 stores a subject classification tree including one or more categories classified in advance. At this time, the database 120 may be an application program installed in the smart terminal 100, or a server program installed in the smart terminal 100. Also, it may be an independent database server connected to the smart terminal 100 via a network.

The database 120 may be a non-volatile memory device such as a cache, a read only memory (ROM), a programmable ROM (PROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM) (Random Access Memory) or a storage medium such as a hard disk drive or a CD-ROM, but is not limited thereto.

The subject classification tree stored in the database 120 classifies the topics of the digital contents 140 into tree form for recommendation of the digital contents 140 as shown in FIG.

2 is an exemplary diagram of a subject classification tree according to an embodiment of the present invention.

Referring to FIG. 2, a subject classification tree includes a hierarchy of categories including various topics in order to extract a topic from text information of the digital content 140, a tree structure (data structure) ).

For example, a subject classification tree can be created based on an open directory project. The Open Directory project is a kind of text database on the web. The Open Directory project classifies texts, such as web documents on the world wide web, into categories using hierarchical ontology and classifies them using a tree data structure.

One embodiment of the present invention utilizes information of a website included in a hierarchical structure of an open directory project as train data of a classifier and generates a subject classification tree according to the characteristics of the smart terminal 100 . For example, when the smart terminal 100 is a smart phone, a subject classification tree suitable for a social network service, a short message service, a moving picture, and a music file that can be mainly used in a smart phone can be created. In addition, if the smart terminal is a personal PC, a subject classification tree can be created and used according to a personal PC usage pattern.

The processor 130 executes a program 150 that recommends the digital content 140 stored in the memory 110. [ That is, the processor 130 collects the digital contents 140 stored in the smart terminal 100, in accordance with the execution of the program 150 recommending the digital contents 140.

The processor 130 may use a back ground service for a predetermined event to collect the digital contents 140 stored in the smart terminal 100. [ For example, the predetermined event may mean pre-setting the time to start the background service, such as daily, hourly, or the like, to start the background service, such as when new digital content 140 is added.

The processor 130 searches for a file stored in the memory 110 of the smart terminal 100 according to the predetermined type of the digital content 140 and stores information of the digital content 140 and the digital content 140 Can be collected. The information of the digital content 140 to be collected may vary according to the type of the digital content 140. Basically, the information of the digital content 140 may include an identifier, a title, and text information of the digital content 140. At this time, the text information includes a description date of the digital content 140, a creation date of the digital content 140 included in the meta-data of the digital content 140, a size of the digital content 140, And location information where the digital content 140 is generated.

Then, the processor 130 extracts a category corresponding to the digital contents 140 collected in the subject classification tree stored in the database 120. For example, in order to extract a category corresponding to the collected digital contents 140, the text information of the digital content 140 and the vector space mode ) And the similarity (similarity).

The processor 130 also classifies the collected digital content 140 based on the extracted category and generates an index of the classified digital content 140 and generates digital content 140 based on the generated index ) Is recommended. In this case, the digital content 140 to be recommended may be the digital content 140 included in the smart terminal 100, and may be a digital contents provider added to the smart terminal 100 via the network Information of the digital content 140, but is not limited thereto.

The recommended digital content 160 may be provided in the form of a list of digital contents 160 recommended by the smart terminal 100 according to the type of the recommended digital contents 160 or may be provided through an application included in the smart terminal 100 Lt; / RTI >

Meanwhile, the processor 130 may extract each word included in the title and text information of the digital content 140. For example, a morphological analysis may be used to extract each word contained in the title and text information of the digital content 140. The processor 130 may extract words matched with the title and text information of the digital content 140 using a word set included in the subject classification tree and a predetermined word set, but the present invention is not limited thereto.

The processor 130 may generate a term vector by applying a weight to words extracted from the digital content 140 based on the vector space model. At this time, the weights can be set based on the term frequency of the words included in the text to be weighted, the total number of learning documents passed through the classifier, and the document frequency of the learning document containing each word have.

For example, the weights can be calculated based on the term frequency-inverse document frequency (TF-IDF) method. Word frequency - Inverse document frequency weighting is a representative word weighting method of the vector space model that is mainly used in the field of information retrieval. For example, it can be calculated as shown in Equation (1).

는 t의 가중치이며,

는 가중치를 구하고자 하는 텍스트에 포함된 t의 빈도, N은 전체 학습 문서의 개수,

는 전체 학습 문서 중 t를 포함하는 문서의 빈도이다. At this time, in Equation (1), t is a word for obtaining a weight,

Is the weight of t ,

Is the frequency of t included in the text to be weighted, N is the number of all learning documents,

Is the frequency of the document containing t among all learning documents.

The processor 130 calculates the degree of similarity by passing it through a classifier for classifying the category of the generated word vector and calculates one or more categories corresponding to a predetermined number according to the degree of similarity calculated in the subject classification tree stored in the database 120 Can be extracted.

For example, the similarity measure may be a cosine similarity as in Equation (2).

는 유사도를 구하고자 하는 단어 벡터,

는 카테고리

의 데이터베이스(120)에 저장된 주제 분류 트리에 포함된 학습 데이터로부터 학습 된 중심 벡터를 의미하며, V는 전체 주제 분류 트리에 포함된 학습 데이터에 사용된 전체 단어의 집합을 의미한다. 또한, 수학식 2에서

는

의 단어 t의 가중치,

는

의 단어 t의 가중치를 의미할 수 있다. In Equation 2,

Is a word vector for which the degree of similarity is sought,

Categories

Means a central vector learned from learning data contained in a subject classification tree stored in the database 120 of the subject classification tree and V denotes a set of all words used in the learning data included in the subject classification tree. In Equation 2,

The

The weight of word t ,

The

May be a weight of the word t .

After calculating the similarity based on the similarity measurement scale, the processor 130 can select k categories with high similarity according to the predetermined number. The selected categories k may be a word vector associated with the high top k pieces (k top-) category.

The processor 130 may divide each of the selected one or more categories into words and select a category including the word having the highest frequency among the divided words.

For example, the processor 130 may select the upper three categories when the predetermined k is 3, and may divide each of the selected upper three categories into words. If the top three categories are "Shopping → Sports", "Sports → Football" and "Sports → Football → UEFA → England", then each category is divided into words and "Shopping → Sports" , "Sports → Football" can be "Sports, Football", "Sports → Football → UEFA → England" can be "Sports, Soccer, UEFA, England". Based on the word thus divided, the processor 130 may calculate the frequency and select a category including the most frequently used word. In the above example, "SPORTS" is selected as the mode, and therefore the processor 130 can select the "shopping to sports" category that includes "SPORTS " as the category corresponding to the digital content 140.

At this time, if the processor 130 divides the category name into words, the information of the four digital contents 140 is (448744, "Shopping, Sports", 1, 0.1), 448744 , 2, 0.2), (492228, "Sports, Soccer", 3, 0.3) and (493203, "Sports, Soccer, UEFA, England", 4, 0.5).

When the frequency of words extracted from the names of the categories divided by words is calculated, there are four "sports", two "shopping", two "football", one "UEFA" and one "England" One. Here, since the word having the mode is "sports ", information of the four digital contents 140 including" sports "in the category name can be selected.

The processor 130 then re-selects the four selected digital content 140 based on the similarity again (493203, "Sports, Soccer, UEFA, England", 4, 0.5), 492228 ), (448744, "Shopping, Sports", 2, 0.2), and (448744, "Shopping, Sports", 1, 0.1). If the number of recommended digital contents 160 is one, the highest degree of similarity is selected (493203, "Sports, Soccer, UEFA, England", 4, 0.5) Can be recommended.

Next, referring to FIG. 3, a program 150 for recommending the digital content 140 according to an embodiment of the present invention will be described.

3 is a configuration diagram of a program 150 recommending the digital content 140 according to an embodiment of the present invention.

The program 150 recommending the digital content 140 may include a digital content acquisition module 151, a digital content classification module 152, a digital content index module 153, and a digital content recommendation module 154.

The digital content acquisition module 151 can automatically collect digital contents 140 such as photos, music, Internet visit records, and bookmark apps included in the smart terminal 100. For example, the digital content acquisition module 151 can be executed according to a predetermined event through the background service. When a predetermined event is executed, searches for a file stored in the memory 110 of the smart terminal 100 according to the file type of the digital content 140, and collects information of the digital content 140 and the digital content 140 have. At this time, information of the digital contents 140 to be collected depends on the type of the digital contents 140. Basically, the information of the digital content 140 may include an identifier, a title, and text information of the digital content 140. [

The digital content classification module 152 may extract the subject of the digital content 140 collected by the digital content collection module 151 based on the subject classification tree. At this time, the digital content classification module 152 extracts words from the title and text information of the digital content 140 to extract a subject of the digital content 140, and applies a weight to the extracted words to generate a word vector have. For example, weights can be word frequency - inverse document frequency method.

The digital content classification module 152 calculates a similarity degree for each category of the subject classification tree stored in the database 120 by passing through a classifier in order to classify categories of word vectors, One or more categories may be extracted. For example, the degree of similarity may use the degree of cosine similarity.

Then, the digital content 140 can be indexed based on the digital content index module 153 after extracting the category of the digital content 140. The digital content index module 153 may generate an index of the digital content 140 including the digital content information table and the category reverse index table. In this case, the index of the digital content 140 may be utilized for quickly accessing the desired digital content 140 by utilizing the information of the digital content 140 or the category information when the recommendation of the digital content 140 is performed.

The digital content recommendation module 154 may perform a recommendation for the digital content 140 in the smart terminal 100 based on the category of the digital content 140 and the digital content index. The digital content recommendation module 154 analyzes the information of the indexed digital content 140 and builds ranking information for the digital content 140 based on the information and includes the recommended digital content 160 according to the ranking information Can be generated.

The recommended digital content 160 may be provided in the form of a list of digital contents 160 recommended by the smart terminal 100 according to the type of the recommended digital contents 160 or may be provided through an application included in the smart terminal 100 Lt; / RTI >

Hereinafter, a method of recommending digital content 140 in the smart terminal 100 according to an embodiment of the present invention will be described with reference to FIGS.

4 is a flowchart illustrating a method of recommending digital content 140 in a smart terminal 100 according to an exemplary embodiment of the present invention.

The digital content recommendation method in the smart terminal 100 according to an exemplary embodiment of the present invention collects the digital contents 140 stored in the smart terminal 100 (S200). At this time, the digital content recommendation method may use a background service for a predetermined event in order to collect the digital contents 140 stored in the smart terminal 100. The predetermined event means presetting the time to start the background service such as daily, hourly, or the like to start the background service, such as when the new digital content 140 is added.

The digital content recommendation method extracts categories corresponding to the digital contents 140 collected in the subject classification tree stored in the database 120 (S210). In this case, the subject classification tree represents a hierarchical structure of categories including various topics in a tree-shaped data structure in order to extract a subject from the text information included in the digital contents 140. For example, a subject classification tree can be generated by using information of a Web site included in a hierarchical structure of an open directory project as learning data of a classifier based on an open directory project.

In addition, the digital content recommendation method classifies the collected digital contents 140 based on the extracted categories (S220), and generates an index of the classified digital contents 140 (S230). For example, the digital content recommendation method utilizes the text information and the subject classification tree of the digital content 140 collected to classify the digital content 140, and the digital content 140 collected based on the vector space model and the similarity Can be extracted.

After generating the index of the classified digital content 140, the digital content recommendation method recommends the digital content 140 based on the generated index (S240). The recommended digital content 160 may be provided in the form of a list of digital contents 160 recommended by the smart terminal 100 according to the type of the recommended digital contents 160 or may be provided through an application included in the smart terminal 100 Lt; / RTI >

5 is a flowchart illustrating a method of extracting categories of digital contents 140 in the smart terminal 100 according to an embodiment of the present invention.

The information of the digital content 140 may include an identifier of the digital content 140, a title of the digital content 140, and text information of the digital content 140. At this time, the text information of the digital content 140 may be descriptive information of the digital content 140 or metadata included in the digital content 140.

Referring to FIG. 5, the digital content recommendation method extracts each word included in the title and text information of the digital content 140 (S211) to extract a category (S220), applies a weight to each extracted word A word vector can be generated (S212). At this time, the weight can be set based on the frequency of each word, the total number of learning documents passed through the classifier, and the frequency of each word included in the entire learning document. For example, weights can be word frequency - inverse document frequency method.

The digital content recommendation method can calculate the degree of similarity by passing the generated word vector through a classifier for classifying categories (S213). For example, the similarity may be a cosine similarity.

The digital content recommendation method extracts one or more categories corresponding to a preset number according to the similarity calculated in the subject classification tree stored in the database 120 (S214), divides each extracted one or more categories into words, The category including the word having the highest frequency among the divided words can be selected (S215).

6 is a flowchart illustrating a method of generating an index of the digital content 140 in the smart terminal 100 according to an exemplary embodiment of the present invention.

Referring to FIG. 6, the digital content recommendation method stores information on the collected digital content 140 and the location of the digital content 140 in a digital content information table to generate an index of the digital content 140 (S230) (S231), and the information of the classified digital contents 140 may be stored in the category reverse index table (S232). At this time, the category reverse index table may include an identifier of the extracted category, a name of the extracted category, an identifier of the digital content 140, and a similarity between the extracted category and the digital content 140.

Since the index of the digital content 140 includes the digital content table and the category inverse index table, the digital content 140 can be displayed on the desired digital content 140 using the information of the digital content 140 or the category information It can be accessed quickly.

7 is a flowchart illustrating a method of performing recommendation of digital contents in the smart terminal 100 according to an exemplary embodiment of the present invention.

Referring to FIG. 7, in order to recommend the digital contents 140, the names of categories stored in the category reverse index table are divided into words (S241), and divided words having the highest frequency among the classified digital contents 140 May be selected (S242). Further, the selected digital contents 140 can be recommended based on the degree of similarity (S243).

The smart terminal 100 and method for recommending the digital contents 140 according to an embodiment of the present invention uses a subject classification tree and selects a category having a high relevance to a word vector and then determines the frequency of words included in the category The category of the category type can be accurately extracted by selecting the category including the word having high frequency, and the digital content 140 can be provided to the user's request by utilizing the category type.

In addition, since the recommendation is performed based on the word frequency of the digital contents 140 classified by the category of the open directory project, it is possible to perform recommendation quickly with a small amount of computing resources and network traffic in an environment like a smart terminal.

One embodiment of the present invention may also be embodied in the form of a recording medium including instructions executable by a computer, such as program modules, being executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. In addition, the computer-readable medium can include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Communication media typically includes any information delivery media, including computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave, or other transport mechanism.

While the methods and systems of the present invention have been described in connection with specific embodiments, some or all of those elements or operations may be implemented using a computer system having a general purpose hardware architecture.

It will be understood by those skilled in the art that the foregoing description of the present invention is for illustrative purposes only and that those of ordinary skill in the art can readily understand that various changes and modifications may be made without departing from the spirit or essential characteristics of the present invention. will be. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

100 smart terminal 110: memory
120: Database 130: Processor

Claims (11)

A smart terminal for recommending digital contents,
A memory in which a program for recommending digital contents is stored,
A database storing a subject classification tree including one or more categories classified in advance; and
And a processor for executing the program,
The processor collects digital contents stored in the smart terminal according to the execution of the program, extracts a category corresponding to the collected digital contents from a subject classification tree stored in the database, A smart terminal for classifying the collected digital contents, generating an index of the classified digital contents, and recommending the digital contents based on the generated index. The method according to claim 1,
Wherein the information of the digital content includes an identifier of the digital content, a title of the digital content, and text information of the digital content,
The processor extracts each word included in the title of the digital content and the text information of the digital content, applies a weight to the extracted word to generate a word vector, classifies the generated word vector into the category Extracting one or more categories corresponding to a predetermined number according to the calculated similarity in a subject classification tree stored in the database, dividing each of the one or more categories into words, A smart terminal that selects a category that contains the word with the highest frequency among the divided words. 3. The method of claim 2,
Wherein the weight is set based on a frequency of each word, a total number of learning documents passed through the classifier, and a frequency of a learning document in which each word is included. The method according to claim 1,
The index of the digital content includes a digital content information table including information of the digital content and information of the digital content,
A category reverse index table including an identifier of the extracted category, a name of the category, an identifier of the digital content, and a similarity,
Storing the collected digital content information and the location of the digital content in a digital content information table, and storing information of the classified digital content in a category reverse index table. 5. The method of claim 4,
Wherein the processor divides the names of the categories stored in the category reverse index table into words and selects one or more digital contents including divided words having the highest frequency among the classified digital contents, Based on the degree of similarity. The method according to claim 1,
Wherein the subject classification tree is generated based on an open directory project. A method for recommending digital content in a smart terminal,
Collecting digital contents stored in the smart terminal;
Extracting a category corresponding to the collected digital contents in a subject classification tree stored in a database;
Classifying the collected digital contents based on the extracted categories;
Generating an index of the classified digital contents; And
And recommending the digital content based on the generated index,
Wherein the subject classification tree comprises one or more categories pre-classified. 8. The method of claim 7,
Wherein the information of the digital content includes an identifier, a title, and text information of the digital content,
The step of extracting the category includes:
Extracting each word included in the title and text information of the digital content;
Generating a word vector by applying a weight to each of the extracted words;
Calculating the similarity by passing the generated word vector through a classifier for classifying the category;
Extracting one or more categories corresponding to a predetermined number according to the calculated similarity in a subject classification tree stored in the database; And
Dividing each of the one or more categories into words, and selecting a category including a word having the highest frequency among the divided words. 9. The method of claim 8,
Wherein the weight is set based on a frequency of each word, a total number of learning documents passed through the classifier, and a frequency of a learning document including the words. 8. The method of claim 7,
The generating of the index may comprise:
Storing the collected digital content information and the location of the digital content in a digital content information table; And
Storing information of the classified digital contents in a category reverse index table,
Wherein the category reverse index table includes an identifier of the extracted category, a name of the category, an identifier of the digital content, and the similarity. 11. The method of claim 10,
Wherein the recommending step comprises:
Dividing the names of the categories stored in the category reverse index table into words;
Selecting one or more digital contents including the divided word having the highest frequency among the classified digital contents; And
And recommending the selected digital content based on the similarity.

Images