KR20230106579A - Method and Apparatus for Providing Knowledge Compass Service - Google Patents

Abstract

The present invention relates to a method for providing a learning compass service, which comprises the steps of: constructing a keyword dictionary whose service target fields may be divided according to each subject and each detailed course by selecting keywords for each subject and keywords for each detailed course for an unassigned curriculum; collecting learning data on free learning content for the unassigned curriculum by performing data crawling on the free learning platforms based on the keywords for each subject and the keywords for each detailed course; extracting data for analysis from the learning data by performing data preprocessing on the learning data; distributing the free learning content to the unassigned curriculum to generate an assigned curriculum by processing the data for analysis using an artificial intelligence classification model which is learned to perform content classification through machine learning; calculating satisfaction rating for each of the free learning content based on a rating calculation model; and providing at least a part of the assignment curriculum together with the satisfaction rating for a user terminal in response to a user query input with respect to the service target field. Therefore, the convenience and efficiency of a user utilizing free learning content can be improved.

Description

Method and Apparatus for Providing Knowledge Compass Service}

The present invention relates to a method of providing a free learning curation service. More specifically, the present invention collects free learning contents through data crawling, classifies them with an artificial intelligence-based classification model, completes the curriculum, calculates ratings with a satisfaction-based statistical model, and provides services to users. It relates to servers and methods.

Demand for education and learning through free learning contents is increasing. Learning by watching free video lectures is gradually spreading through free learning platforms such as YouTube, edwith, and K-MOOC.

According to a research by a market research company, it was found that the frequency and satisfaction with the learning contents provided on YouTube, etc., of the subjects of the survey were quite high, and that learning time and cost reduction could be experienced by the learning contents. . In particular, as the public's interest in technologies related to the 4th industrial revolution increases in recent years, demand for free learning content is also increasing in areas such as big data, artificial intelligence and software.

As described above, although learning through free learning content provides high satisfaction and the demand for it is increasing, learning through free learning content still causes a lot of inconvenience to learners.

For example, it can take a lot of time in the process of searching for content on various platforms and determining whether the actual lecture content meets the learner's needs, and the content is mixed on multiple platforms without an established system, so that learners can It can be a problem that it can be difficult to use them effectively.

A technical problem to be solved by the present invention is to solve problems related to inconvenience and inefficiency in the process of learning using free learning contents by providing a free learning curation service.

As a means for solving the above technical problem, a server providing a free learning curation service according to an aspect of the present invention includes a memory storing instructions for providing the free learning curation service; and by executing the above commands: constructing a keyword dictionary by selecting keywords for each subject and keywords for each detailed process for an unassigned curriculum that classifies service target fields by subject and detailed process, and constructs a keyword dictionary by selecting the keywords for each subject and the detailed process Data crawling of free learning platforms based on keywords for each course is performed to collect learning data of free learning contents for the unassigned curriculum, and data for analysis is performed from the learning data by performing data pre-processing on the learning data. By extracting and processing the data for analysis using an artificial intelligence classification model learned to classify content through machine learning, the free learning contents are distributed to the unassigned curriculum to create an assignment curriculum, and score calculation model A processor for calculating a satisfaction rating for each of the free learning contents based on, and providing at least a part of the assignment curriculum to a user terminal together with the satisfaction rating in response to a user's query input regarding the service target field. include

A method for providing a free learning curation service by a server providing a free learning curation service according to another aspect of the present invention includes keywords for each subject for an unassigned curriculum that classifies service target fields by subject and detailed course. and constructing a keyword dictionary by selecting keywords for each detailed process; Collecting learning data of free learning contents for the unassigned curriculum by performing data crawling on free learning platforms based on the keywords for each subject and the keywords for each detailed course; extracting data for analysis from the learning data by performing data pre-processing on the learning data; generating an assigned curriculum by distributing the free learning contents to the unassigned curriculum by processing the data for analysis using an artificial intelligence classification model learned to classify contents through machine learning; Calculating a satisfaction rating for each of the free learning contents based on a rating calculation model; and providing at least a part of the assignment curriculum together with the satisfaction rating to a user terminal in response to a user's query input regarding the service target field.

According to the free learning curator service according to the present invention, data crawling for free learning platforms is performed, curriculum assignment for free learning contents is performed by an artificial intelligence classification model, and free learning contents according to a rating calculation model As a result of calculating the satisfaction rating, the assignment curriculum and satisfaction rating can be provided to the user, problems in the process of learning using free learning contents are solved, and the convenience and efficiency of users using free learning contents are improved. can be improved

1 is a diagram for explaining a system for providing a free learning curation service according to some embodiments.
2 is a block diagram illustrating elements constituting a server providing a free learning curation service according to some embodiments.
3 is a diagram for explaining a process of providing a free learning curation service according to some embodiments.
4 is a diagram for explaining a process of setting an unassigned curriculum and constructing a keyword dictionary according to some embodiments.
5 is a diagram for explaining a process of collecting learning data by performing data crawling on free learning platforms according to some embodiments.
6 is a diagram for explaining a process of collecting text data corresponding to voice data by performing voice recognition according to some embodiments.
7 is a diagram for explaining a process of extracting data for analysis and generating an assignment curriculum according to some embodiments.
8 is a diagram for explaining a process of generating a score calculation model to calculate a satisfaction score according to some embodiments.
9 is a diagram for explaining a process of providing an assignment curriculum and satisfaction rating to a user terminal according to some embodiments.
10 is a flowchart illustrating steps constituting a method of providing a free learning curation service according to some embodiments.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the drawings. The following description is only for specifying the embodiments, and is not intended to limit or limit the scope of rights according to the present invention. What a person skilled in the art can easily infer from the detailed description and examples of the present invention should be construed as belonging to the scope of the present invention.

The terms used in the present invention have been described as general terms widely used in the technical field of the present invention, but the meaning of the terms used in the present invention may vary depending on the intention of a person working in the field, a precedent, or the emergence of new technology. there is. Some terms may be arbitrarily selected by the applicant, in which case the meaning of the terms will be explained in detail. The terms used in the present invention should be interpreted as meanings according to the overall context of the specification, not just dictionary meanings.

Terms such as 'consisting of' or 'includes' used in the present invention should not necessarily be construed as including all of the components or steps described in the specification, and when some components or steps are not included, and The inclusion of additional components or steps should also be interpreted as intended from the term.

Terms including ordinal numbers such as 'first' or 'second' used in the present invention may be used to describe various components or steps, but the components or steps should not be limited by ordinal numbers. . Terms containing ordinal numbers should only be construed to distinguish one component or step from other components or steps.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the drawings. A detailed description of matters widely known to those skilled in the art will be omitted.

1 is a diagram for explaining a system for providing a free learning curation service according to some embodiments.

Referring to FIG. 1 , a system 10 providing a free learning curation service may include a server 100 providing a free learning curation service, free learning platforms 200, and a user terminal 300. . However, it is not limited thereto, and other general-purpose components other than the components shown in FIG. 1 may be further included in the system 10 .

In the system 10 , the server 100 may utilize the free learning platforms 200 to provide a free learning curation service to users using the user terminal 300 . For example, the system 10 may be implemented in a web environment and the free learning curation service may be provided to users through a web browser, or the system 10 may be implemented as an application and the free learning curation service may be provided through a mobile device. It may be provided to the user through

The server 100 may be a device for providing a free learning curation service to the user terminal 300 in the system 10 . For example, the server 100 may refer to hardware, software, or a combination thereof for implementing a free learning curation service operating in a web or app environment.

The free learning platforms 200 may refer to content providers that provide free learning materials such as video lectures through websites or applications. For example, the free learning platforms 200 may refer to a video sharing platform such as YouTube, an education platform such as Edwith, and an online open lecture platform such as K-MOOC.

The user terminal 300 may mean a device for a user to access a free learning curation service provided from the server 100 . For example, the user terminal 300 may refer to various electronic devices such as a PC, laptop computer, tablet PC, and smart phone capable of driving a website or application implementing the system 10 .

Interaction between components of system 10 may be performed in a wired or wireless data communication environment. Interactions between the free learning platforms 200 and the server 100 and between the server 100 and the user terminal 300 can be achieved through wired/wireless data communication.

In the system 10, since a free learning curation service can be provided to the user terminal 300 by the server 100 utilizing the free learning platforms 200, the user can learn for free according to the preference condition to learn. Since learning contents and related curriculum can be provided, the user's learning convenience and efficiency can be improved.

2 is a block diagram illustrating elements constituting a server providing a free learning curation service according to some embodiments.

Referring to FIG. 2 , a server 100 providing a free learning curation service may include a memory 110 and a processor 120 . However, it is not limited thereto, and other general-purpose elements other than the elements shown in FIG. 2 may be further included in the server 100.

The server 100 may be a computing device for providing a free learning curation service. The server 100 may include the memory 110 as a means for storing various data, instructions, and at least one program or software, and executes the instructions or at least one program to process various data. As a means for doing so, the processor 120 may be included.

The memory 110 may store instructions for providing a free learning curation service. For example, memory 110 may store instructions constituting software such as a computer program or mobile application. In addition, the memory 110 may store instructions or various data necessary for executing at least one program.

The memory 110 may be implemented with non-volatile memory such as ROM, PROM, EPROM, EEPROM, flash memory, PRAM, MRAM, RRAM, FRAM, etc., or volatile memory such as DRAM, SRAM, SDRAM, PRAM, RRAM, FeRAM, etc. can be implemented as Also, the memory 110 may be implemented as a HDD, SSD, SD, Micro-SD, or the like.

The processor 120 may provide a free learning curation service by executing instructions stored in the memory 110 or at least one program. The processor 120 may perform a series of processes for providing a free learning curation service. The processor 120 may perform overall functions for controlling the server 100 and may process various internal operations of the server 100 .

Processor 120 may be implemented as an array of multiple logic gates or as a general-purpose microprocessor. Processor 120 may consist of a single processor or a plurality of processors. The processor 120 may be integrated with the memory 110, rather than a separate component from the memory 110 storing at least one program. The processor 120 may be at least one of a CPU, a GPU, and an AP provided in the device 100, but this is only an example, and the processor 120 may be implemented in various other forms.

The server 100 may perform steps of a method for providing a free learning curation service. As the server 100 performs each step of the method for providing the free learning curation service, the free learning curation service may be provided to the user.

The processor 120 may build a keyword dictionary by selecting keywords for each subject and keywords for each detailed course for an unassigned curriculum that classifies service target fields by subject and detailed course.

The service target field may mean a specific field that is a target of learning. For example, the service target field may be a 4th industrial revolution technology field including at least one of big data, artificial intelligence, machine learning, and software. In addition, service targets may correspond to various fields such as English, foreign languages, MS Office, financial technology, and fitness.

The unassigned curriculum may mean that a service target field is divided into subjects and detailed courses. For example, the 4th industrial revolution technology field can be divided into big data subjects, artificial intelligence subjects, etc., and detailed processes of learning can be distinguished in each subject. However, the unassigned curriculum may refer to a curriculum in which detailed contents have not yet been assigned to each subject and detailed course.

As an unassigned curriculum, a pre-configured by experts in each service target field can be used. Alternatively, it may be configured by referring to the table of contents of widely used learning textbooks in each service target field.

The keyword dictionary may include keywords for each subject of the unassigned curriculum and keywords for each detailed course. In the process of setting the unassigned curriculum, keywords may be selected for each subject and each detailed course by analyzing which keywords are related to each subject of a specific service target field and each detailed course. On the other hand, keyword selection may be performed by an expert constituting an unassigned curriculum or by referring to learning materials.

The processor 120 may perform data crawling of the free learning platforms 200 based on keywords for each subject and keywords for each detailed course to collect learning data of free learning contents for an unassigned curriculum.

Data crawling may refer to a process of collecting free learning contents provided through the free learning platforms 200 . Data crawling or web crawling may refer to an operation that enables big data analysis by collecting vast amounts of data from web pages, portal sites, etc. through a high-performance processing unit instead of a direct search by a person. Free learning contents may be collected from the free learning platforms 200 according to data crawling.

Learning data collected by data crawling is various data about free learning contents, and may refer to content introductions, titles, subtitles, comments, and the like that can be collected by crawling. Meanwhile, the learning data may include audio data extracted from video lectures and may include text data converted from the corresponding audio data.

Data crawling can be performed based on keywords. Keywords for each subject selected from the unassigned curriculum and keywords for each detailed course may be search words for data crawling. However, search results of keywords are not immediately distributed to subjects and detailed courses of the unassigned curriculum, and free learning contents, which are search results, may be distributed to the unassigned curriculum after going through a separate classification algorithm.

The processor 120 may extract data for analysis from the learning data by performing data preprocessing on the training data.

Since training data is unstructured data such as text and voice collected by data crawling, it may be difficult to apply a classification model to training data directly. Accordingly, data preprocessing may be performed to change the training data into a form in which a classification model can be easily applied.

Data pre-processing may refer to a process of quantifying training data. For example, calculating TF-IDF or applying the word2vec algorithm may correspond to data preprocessing. However, it is not limited thereto, and if the learning data is changed to be more manageable, a processing method for changing data in a form other than numerical values may also correspond to data pre-processing.

The processor 120 may generate an assigned curriculum by distributing free learning contents to an unassigned curriculum by processing data for analysis using an artificial intelligence classification model learned to classify contents through machine learning.

The artificial intelligence classification model can be trained to perform content classification by machine learning. According to the artificial intelligence classification model, free learning contents can be classified according to which subject and which detailed course of the unassigned curriculum are classified by subject and detailed course. Learning of such an artificial intelligence classification model can be performed in various ways such as CNN, SVM, and Bayesian classification.

The artificial intelligence classification model can receive data for analysis and output which subject and detailed course the free learning contents belong to the most appropriate. However, since it is difficult for an artificial intelligence classification model to directly process training data, which is unstructured data, analysis data extracted from training data by data preprocessing can be input to an artificial intelligence classification model.

When the analysis data is processed by the artificial intelligence classification model, the assigned curriculum may be created by distributing the free learning contents to the non-assigned curriculum. Therefore, the free learning contents collected as a result of data crawling based on the keyword dictionary of the unassigned curriculum can be classified into subjects and detailed courses only by classification by the artificial intelligence classification model, and then distributed and assigned to the corresponding subjects and detailed courses. Curriculum can be formed.

Meanwhile, when the artificial intelligence classification model classifies free learning contents, whether the free learning contents are advertising contents may also be classified together. That is, an artificial intelligence classification model may be trained to classify together whether the free learning contents include advertisements and the degree of the advertisements.

The processor 120 may calculate a satisfaction rating for each of the free learning contents based on the rating calculation model.

In the assignment curriculum, free learning contents are distributed according to subjects and detailed courses, and in addition, satisfaction ratings for each of the free learning contents may be provided together. To this end, a score calculation model that calculates a satisfaction score may be developed. The rating calculation model may refer to a model for rating satisfaction with free learning contents using a statistical method such as multiple regression analysis.

The processor 120 may provide the user terminal with at least a part of the assignment curriculum along with a satisfaction rating in response to a user's query input regarding the service target field.

In a state in which the assignment curriculum is created and the satisfaction rating is calculated, the user may transmit his interest field or preferred lecture conditions to the server 100 through the user terminal 300 as a query input, and in response, the processor Step 120 may provide the user terminal 300 with a result corresponding to the user's query input.

For example, when the assignment curriculum relates to the 4th Industrial Revolution technology field and the user's query input is the AI technology field, the processor 120 filters only contents classified as AI subjects from the assignment curriculum classified by subject. This can be provided to the user along with the satisfaction rating.

As described above, the processor 120 classifies the result of performing data crawling into an artificial intelligence classification model, calculates a satisfaction rating for it, and provides it to the user in the form of a curriculum, so that the user can learn for himself free of charge. Since a curriculum for learning can be provided together with a satisfaction rating without the effort of searching the platforms 200 or personally checking what content the free learning contents cover, the convenience and efficiency of learning can be improved.

3 is a diagram for explaining a process of providing a free learning curation service according to some embodiments.

Referring to FIG. 3 , a process of providing a free learning curation service is shown as processes 310 to 350 . Processes 310 to 350 may mean processes for providing a free learning curation service performed by the processor 120 of the server 100 .

In process 310, an unassigned curriculum and keyword dictionary may be created. The server 100 may build a keyword dictionary by selecting keywords for each subject and keywords for each detailed course for an unassigned curriculum that classifies service target fields by subject and detailed course.

In process 320, data crawling may be performed. The server 100 may collect learning data of free learning contents for an unassigned curriculum by performing data crawling on the free learning platforms 200 based on keywords for each subject and keywords for each detailed course.

In process 330, data pre-processing may be performed, and free learning contents may be allocated using an artificial intelligence classification model. The server 100 may extract data for analysis from the learning data by performing data pre-processing on the learning data, and processing the data for analysis by utilizing an artificial intelligence classification model learned to perform content classification through machine learning. An assigned curriculum may be created by distributing free learning contents to an unassigned curriculum.

In process 340, a satisfaction rating may be calculated. The server 100 may calculate a satisfaction rating for each of the free learning contents based on the rating calculation model.

In process 350, the placement curriculum and satisfaction rating may be provided to the user. The server 100 may provide the user terminal with at least a part of the assignment curriculum along with a satisfaction rating in response to a user's query input regarding the service target field.

As shown, the process flow proceeds from process 310 to process 350, and video classification, satisfaction rating, and service operation may be performed by the basic process. On the other hand, since a reverse cycle process exists in the entire process, feedback in the reverse direction may be achieved. For example, synonyms may be added to a keyword dictionary as a result of data pre-processing, and user feedback may be used in generating a rating calculation model or analyzing customer attributes in an artificial intelligence classification model during service operation.

4 is a diagram for explaining a process of setting an unassigned curriculum and constructing a keyword dictionary according to some embodiments.

Referring to FIG. 4, in process 310 of FIG. 3, the server 100 selects keywords for each subject and keywords for each detailed process for an unassigned curriculum that classifies service target fields by subject and detailed process to build a keyword dictionary. In order to explain the process in more detail, an unassigned curriculum 410 and a keyword dictionary 420 in which only a portion thereof is shown are shown.

The unassigned curriculum 410 may classify service target fields by subject and detailed course. In the illustrated example, data management analysis subjects in the field of big data may be classified into beginner, intermediate, and advanced levels, and each level may be further classified into learning topics corresponding to detailed courses.

The keyword dictionary 420 may include keywords selected for each of the subjects and detailed courses included in the unassigned curriculum 410 . The keyword dictionary 420 may also include additional keywords, such as synonyms, for each of the subjects and sub-courses. As described above, the addition of synonyms may be performed in a data pre-processing process for learning data of free learning contents.

5 is a diagram for explaining a process of collecting learning data by performing data crawling on free learning platforms according to some embodiments.

Referring to FIG. 5, in process 320 of FIG. 3, the server 100 performs data crawling of the free learning platforms 200 based on keywords for each subject and keywords for each detailed course, thereby free learning for the unassigned curriculum. A crawling process 510 for explaining a process of collecting learning data of contents in more detail is shown.

In the preliminary preparation step of the crawling process 510, an unassigned curriculum (basic curriculum) may be created using expert knowledge and basic textbooks of related subjects, and a keyword dictionary may be created by selecting keywords for each subject and detailed process. can

Thereafter, data crawling for the free learning platforms 200 (video lecture data sources) may be performed, and learning data of free learning contents may be collected therefrom to build a basic lecture DB. When the processor 120 collects learning data, the title, subject, introduction, subtitle, comment, instructor information, URL, number of views, affiliated channel, number of subscribers, and likes for each of the free learning contents from the free learning platforms 200. At least one of the number, content size, and voice data may be collected.

In addition, when the training data is collected, the processor 120 performs speech recognition (STT, Speech To Text) on the speech data when the training data includes the speech data to collect text data corresponding to the speech data. can That is, since only formal data of free learning contents is collected, and text data corresponding to voice data can be collected, what content the free learning contents actually contain can be collected by data crawling. .

6 is a diagram for explaining a process of collecting text data corresponding to voice data by performing voice recognition on voice data according to some embodiments.

Referring to FIG. 6 , an R-medus 610 corresponding to a previous study on speech recognition is shown. R-medus (610) is a mobile R practice application that performs coding through voice.

The voice recognition technology in R-medus 610 can also be applied to free learning curation services. When the processor 120 performs voice recognition (STT) on voice data and collects text data corresponding to the voice data, voice recognition (STT) may be performed using the R-medus 610.

7 is a diagram for explaining a process of extracting data for analysis and generating an assignment curriculum according to some embodiments.

Referring to FIG. 7 , data pre-processing process 710 for free learning contents (C1, C2, C3) collected as a result of data crawling for free learning platforms 200, content classification is performed through machine learning An artificial intelligence classification model 720 learned to do so and an assignment curriculum 730 generated by assignment of free learning contents C1, C2, and C3 are shown.

The free learning contents C1 , C2 , and C3 may be collected as a result of data crawling of the free learning platforms 200 . In the illustrated example, it is expressed that three contents are collected, but this is only an example, and various numbers of contents may be collected.

The free learning contents C1, C2, and C3 may have learning data. For example, as described above, video titles, subtitles, explanatory phrases, comments, and text corresponding to voices of the free learning content C1 may correspond to the learning data 711, and the remaining free learning content C2 , C3) may also have their own learning data.

In the data pre-processing process 710, data pre-processing is performed on the training data, and data for analysis may be extracted. For example, data 712 for analysis may be extracted by data pre-processing of the learning data 711 of the free learning content C1. For example, the analysis data 712 may be word embedding converted from the learning data 711 corresponding to unstructured data and expressed in a vector form.

The data pre-processing process 710 may refer to a process of converting training data 711 corresponding to text into analysis data 712 expressed numerically such as a vector. To this end, when extracting data 712 for analysis, the processor 120 performs natural language processing (NLP)-based morphological analysis and text mining on the learning data 711 to generate the training data 711. extract word data from; In addition, at least one of TF-IDF (Term Frequency-Inverse Document Frequency) calculation and word2vec algorithm application may be performed for word data.

When morpheme analysis and text mining based on natural language processing are performed on the training data 711, morphemes or words constituting the text of the training data 711 may be extracted as word data. Statistical values representing the importance of the word data may be calculated for such word data, such as TF-IDF, or the word data may be numerically expressed in the form of a vector by applying the word2vec algorithm.

Meanwhile, when extracting the analysis data 712, the processor 120 may add synonyms to the keyword dictionary based on similarity with keywords for each subject and keywords for each detailed course of the learning data 711. . When data for analysis 712 expressed numerically is extracted, similarity may be determined based on the corresponding numerical value.

Therefore, when there are other similar keywords determined to be similar to the specific keyword in the data pre-processing process 710 for a specific keyword in the keyword dictionary used to collect the free learning content C1, the keyword dictionary in FIG. 4 ( As in the example of 420), similar keywords may be added to the keyword dictionary as synonyms.

When the data pre-processing process 710 is performed on the training data 711, that is, when TF-IDF calculation or word2vec algorithm is applied to the training data 711 and data for analysis 712 is extracted, therefrom An artificial intelligence classification model 720 may operate. The artificial intelligence classification model 720 may perform content classification by using data for analysis as an input.

When the processor 120 generates the assignment curriculum 730, the artificial intelligence classification model (720 ) to process data for analysis. That is, the artificial intelligence classification model 720 may be formed by learning by machine learning such as CNN, SVM, or Bayesian classification.

Meanwhile, when generating the assignment curriculum 730, the processor 120 may classify whether each of the free learning contents is advertisement content by using the artificial intelligence classification model 720. That is, the artificial intelligence classification model 720 may be trained to additionally classify advertising content in the process of learning to perform content classification. For example, the artificial intelligence classification model 720 may be trained to classify whether each of the free learning contents includes an advertisement or to numerically classify the degree of advertisement.

According to the artificial intelligence classification model 720, the free learning contents C1, C2, and C3 are distributed to the unassigned curriculum classified by subject and detailed course, so that the distribution curriculum 730 can be created. That is, the AI classification model 720 can classify each of the free learning contents C1, C2, and C3 to which subject and detailed course of the unassigned curriculum correspond.

8 is a diagram for explaining a process of generating a score calculation model to calculate a satisfaction score according to some embodiments.

Referring to FIG. 8 , a rating calculation model creation 810 is shown to explain a process of generating a rating calculation model to calculate a satisfaction rating. According to the rating calculation model generation 810, the rating calculation model can be created through a series of processes, and after being created, a satisfaction rating for each of the free learning contents can be calculated.

The process of creating a score calculation model according to the generation 810 may include a process of setting a data set (creation of data variables) and a process of deriving a score calculation model by performing multiple regression analysis (model development). To this end, when calculating the satisfaction rating, the processor 120 sets a data set including at least one of the number of views, the number of subscribers, update cycle, and expert rating of free learning contents; In addition, a rating calculation model can be derived by performing multiple regression analysis on the data set.

In addition, as shown, in the generation process according to the rating calculation model generation 810, the process of identifying the characteristics of the business for model development (understanding of the business), understanding and preprocessing before generating the data set The process (data understanding, data preprocessing), the process of dividing the data set into training, test, and verification data (data segmentation), exploratory data analysis (EDA) and statistical analysis to exclude inappropriate data process (data variable selection), process of evaluating the rating calculation model derived by multiple regression analysis through a statistical method (model evaluation), process of verifying the rating calculation model by applying test data (model validation), and evaluation and The process of managing and monitoring the verified model (end of model development) may be further included.

When the rating calculation model is generated by the generation process according to the rating calculation model generation 810, satisfaction ratings for the free learning contents assigned to the assignment curriculum may be calculated through the rating calculation model. Since the satisfaction with the free learning contents can be statistically quantified and provided, the satisfaction rating can be presented as a more objective reference material in the process of planning the learning of the free learning contents by the user.

9 is a diagram for explaining a process of providing an assignment curriculum and satisfaction rating to a user terminal according to some embodiments.

Referring to FIG. 9 , as an example of a free learning curation service provided to the user terminal 300, a first UI 910 displaying at least a part of the assignment curriculum and a detailed process included in at least a part of the assignment curriculum A second UI 920 displaying a satisfaction rating is shown.

The user may receive information such as the first UI 910 and the second UI 920 from the server 100 through the user terminal 300 . For example, the user may transmit a query input such as a technical field, subject, detailed course, level of difficulty, etc. to the server 100 through the user terminal 300, and the server 100 may send an assignment curriculum corresponding to the user's query input. At least a part of and the satisfaction rating may be transmitted to the user terminal 300 in the same manner as the first UI 910 and the second UI 920 .

On the other hand, the processor 120, when provided to the user terminal 300, based on the log data including the user's inquiry history, course taking history, progress rate and attendance rate, and user preference data including the user's content preference, customized for the user. management services can be provided.

According to the log data and user preference data accumulated in the process of using the free learning curation service by the user, the tendency of the user to use the free learning curation service can be identified. Therefore, when an artificial intelligence classification model and a rating generation model are updated by utilizing user tendencies based on log data and user preference data, a user-customized management service that supports a more systematic service tailored to the user can be provided.

10 is a flowchart illustrating steps constituting a method of providing a free learning curation service according to some embodiments.

Referring to FIG. 10 , the method for providing a free learning curation service may include steps 1010 to 1060. However, it is not limited thereto, and other general-purpose steps other than the steps shown in FIG. 10 may be further included in the method of providing free learning curation service.

The method of FIG. 10 may consist of steps processed time-sequentially in the server 100 described with reference to FIGS. 1 to 9 . Therefore, even if the content of the method of FIG. 10 is omitted below, the content described above for the server 100 of FIGS. 1 to 9 may be equally applied to the method of FIG. 10 .

In step 1010, the server 100 may build a keyword dictionary by selecting keywords for each subject and keywords for each detailed course for an unassigned curriculum that classifies service target fields by subject and detailed course.

In step 1020, the server 100 may perform data crawling of the free learning platforms 200 based on keywords for each subject and keywords for each detailed course to collect learning data of free learning contents for unassigned curriculum. there is.

In the process of collecting learning data, the server 100 provides the title, subject, introduction, subtitles, comments, instructor information, URL, number of views, number of affiliated channels, number of subscribers, At least one of the number of likes, content size, and voice data may be collected.

In the course of collecting learning data, if the learning data includes voice data, the server 100 may perform speech recognition (STT, Speech To Text) on the voice data to collect text data corresponding to the voice data. .

In step 1030, the server 100 may extract data for analysis from the learning data by performing data pre-processing on the training data.

In the process of extracting data for analysis, the server 100 may extract word data from the training data by performing natural language processing (NLP)-based morphological analysis and text mining on the training data, and may extract word data from the training data. At least one of TF-IDF (Term Frequency-Inverse Document Frequency) calculation and application of the word2vec algorithm may be performed.

In the process of extracting data for analysis, the server 100 may add synonyms to the keyword dictionary based on similarity with keywords for each subject and keywords for each detailed course of the learning data.

In step 1040, the server 100 processes data for analysis using an artificial intelligence classification model that is learned to classify content through machine learning, thereby distributing free learning contents to unassigned curricula to create an assigned curriculum. .

In the process of generating the assignment curriculum, the server 100 may classify whether each of the free learning contents is advertising contents by using an artificial intelligence classification model.

In the process of generating the assignment curriculum, the server 100 utilizes an artificial intelligence classification model learned by at least one of CNN (Convolutional Neural Network), SVM (Support Vector Machine), and Bayesian Classification for analysis. data can be processed.

In step 1050, the server 100 may calculate a satisfaction rating for each of the free learning contents based on the rating calculation model.

In the process of calculating the satisfaction rating, the server 100 may set a data set including at least one of the number of views, the number of subscribers, the update cycle, and expert ratings of free learning contents, and perform multiple regression analysis on the data set. Analysis) to derive a score calculation model.

In step 1060, the server 100 may provide the user terminal 300 with at least a part of the assignment curriculum along with a satisfaction rating in response to the user's query input on the service target field.

In the process of providing the user terminal 300, the server 100 provides a user-customized management service based on log data including the user's inquiry history, course taking history, progress rate and attendance rate, and user preference data including the user's content preference. can provide.

Meanwhile, the service target field may be a fourth industrial revolution technology field including at least one of big data, artificial intelligence, machine learning, and software.

The method of providing the free learning curation service of FIG. 10 may be recorded on a computer-readable recording medium on which at least one program or software including instructions for executing the method is recorded.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and floptical disks such as Hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like, may be included. Examples of program instructions may include high-level language codes that can be executed by a computer using an interpreter or the like as well as machine language codes generated by a compiler.

Although the embodiments have been described in detail above, the scope of rights according to the present invention is not limited thereto, and various modifications and improvements of those skilled in the art using the basic concepts of the present invention described in the following claims are also included in the present invention. included in the scope of rights according to

Claims (5)

A method for providing a free learning compass service to a user terminal by a server including a memory and a processor,
Building a keyword dictionary by selecting keywords for each subject and keywords for each detailed process for an unassigned curriculum that classifies service target fields by subject and detailed process;
Collecting learning data of free learning contents for the unassigned curriculum by performing data crawling on free learning platforms based on the keywords for each subject and the keywords for each detailed course;
extracting data for analysis from the learning data by performing data pre-processing on the learning data;
generating an assigned curriculum by distributing the free learning contents to the unassigned curriculum by processing the data for analysis using an artificial intelligence classification model learned to classify contents through machine learning;
Calculating a satisfaction rating for each of the free learning contents based on a rating calculation model, the rating calculation model comprising:
a data preprocessing step of constructing a data set by receiving internal and external data related to ratings, recognizing missing values based on actual and average values, preprocessing the missing values, and removing outliers and erroneous input noise;
Generating a variable that affects lecture satisfaction, wherein the variable includes an update cycle, a number of views, a number of subscribers, and a number of comments, data variable creation step;
a data division step of dividing the data into construction, verification, and performance verification;
A data variable selection step of excluding inappropriate variables from the data and selecting variables to be used in the analysis, wherein the data variable selection is to check trends for each variable by exploratory data analysis, and to select variables through statistical analysis. the data variable selection step;
developing the scoring model from logistic regression analysis;
Evaluating the scoring model through KS statistics, ROC curves, and AR accuracy; and
Calculating the satisfaction rating, which is generated from the step of applying test data and verifying the rating calculation model through gap analysis;
In response to a user's query input regarding the service target field, providing at least a part of the assignment curriculum together with the satisfaction rating to the user terminal;
The generating of the assignment curriculum may include generating the assignment curriculum from the artificial intelligence classification model based on a user's feedback on at least a part of the assignment curriculum provided to the user terminal, and utilizing the artificial intelligence classification model to provide the free assignment curriculum. Classifying whether each of the learning contents is advertising content,
The user terminal includes first UI information including a study map filtering function for selecting a technical field, subject, detailed course, difficulty level, age group, and form, and second UI information displaying a detailed course and satisfaction rating of the assignment curriculum A method for providing a learning compass service, further comprising each step provided. According to claim 1,
The step of extracting the data for analysis by performing data preprocessing may include extracting word data from the learning data by performing natural language processing (NLP)-based morpheme analysis and text mining on the learning data. ; and
A method for providing a learning compass service, comprising performing at least one of calculating a term frequency-inverse document frequency (TF-IDF) for the word data and applying a word2vec algorithm. According to claim 1,
The extracting of the data for analysis may further include adding synonyms to the keyword dictionary based on similarities between the keywords for each subject and the keywords for each detailed course of the learning data. How to provide. According to claim 1,
The step of collecting learning data of the free learning contents by performing the data crawling,
A preliminary step of creating an unassigned curriculum using basic textbooks related to the learning subject, using the process and detailed process included in the basic textbook, and selecting a search keyword according to the frequency of derivation for each of the process and detailed process;
Audio data is extracted from the entire video lecture data source including both video and audio, not meta data or meta information included in the video, and speech recognition is performed on the audio data to extract text data corresponding to the audio data. doing;
Crawling subtitle data included in video lectures separately from the audio data;
A method of providing a learning compass service comprising the step of storing text data including both the text data and the subtitle data in a basic database of lectures. A computer readable storage medium storing instructions for performing the method according to claims 1 to 4.

Images