KR102316837B1 - Public system of studying abroad information - Google Patents

Abstract

The present invention relates to a study abroad information sharing system, and the study abroad information sharing system according to an embodiment of the present invention includes: an information server; and a terminal connected to the information server online and used by international students, wherein the information server includes: a study abroad information collection unit for collecting information on study abroad through online; an international student information unit that collects information about the international student through the terminal used by the international student; a matching information guide unit that provides the information collected through the study abroad information collection unit to international students through the terminal; And the information collected by the study abroad information collection unit is analyzed based on an artificial neural network, and the information matching the information collected through the foreign student information unit is analyzed, including any one or more of a topic and a schedule, and the analyzed information is matched with the matching information. It may include a processing unit for transmitting to the information guide. According to the present invention, it is possible to classify information that is fragmented and indiscriminately scattered by international students according to topics and guide them according to a schedule, so that international students can check the necessary information according to their schedule.

Description

Study Abroad Information Sharing System {PUBLIC SYSTEM OF STUDYING ABROAD INFORMATION}

The present invention relates to a study abroad information sharing system, and more particularly, to a study abroad information sharing system capable of providing necessary information to students living abroad and sharing information among international students.

Students and parents preparing for study abroad or training require a wide variety of information such as the nature and contents of the educational institution's program, facilities, cultural conditions of the area where it is located, and tuition and living expenses. This is not only true for overseas students, but also for foreign students who come to Korea to study abroad. Therefore, it is necessary to collect considerable information for studying abroad.

Recently, information on studying abroad can be found on the Internet rather easily, but generally it is only a guide to procedures necessary for study abroad procedures, and information related to study abroad is also provided in the form of a list of information related to studying abroad.

In addition, although there are programs supported by schools, the government, and local governments, in most cases, only limited information is provided or only simple information is provided. Therefore, it is not easy for international students to find information that applies to them.

Republic of Korea Patent Publication No. 10-2000-0050035 (2000.08.05.) Republic of Korea Patent Publication No. 10-2018-0055100 (2018.05.25.)

An object of the present invention is to provide a study abroad information sharing system capable of sharing customized study abroad information to international students.

Study abroad information sharing system according to an embodiment of the present invention, an information server; and a terminal connected to the information server online and used by international students, wherein the information server includes: a study abroad information collection unit for collecting information on study abroad through online; an international student information unit that collects information about the international student through the terminal used by the international student; a matching information guide unit that provides the information collected through the study abroad information collection unit to international students through the terminal; And the information collected by the study abroad information collection unit is analyzed based on an artificial neural network, and the information matching the information collected through the foreign student information unit is analyzed, including any one or more of a topic and a schedule, and the analyzed information is matched with the matching information. It may include a processing unit for transmitting to the information guide.

The international student information unit may collect information on any one or more of ID, nationality, affiliated school, major, accommodation, and field of interest of the international student.

The study abroad information collection unit may input information related to study abroad by an administrator through offline.

The information server may further include a terminal used by a local student and connected to the information server online, and the information server may further include a local student information unit for collecting information about the local student through the terminal used by the local student. .

The processing unit may analyze information using a restricted Boltzmann machine model.

According to the present invention, it is possible to classify information that is fragmented and indiscriminately scattered by international students according to topics and guide them according to a schedule, so that international students can check the necessary information according to their schedule.

In addition, necessary information between international students and local students can be exchanged through the information server, which has the effect of allowing international students to more easily adapt to the local area.

1 is a diagram illustrating a study abroad information sharing system according to an embodiment of the present invention.
2 is a block diagram illustrating an information server of a study abroad information sharing system according to an embodiment of the present invention.
3 is a diagram for explaining a concept for performing deep learning of the study abroad information sharing system according to an embodiment of the present invention.
4 is a diagram for explaining an example of performing deep learning according to an embodiment of the present invention.
5 is a diagram illustrating classification according to a schedule of the study abroad information sharing system according to an embodiment of the present invention.
6 is a view showing the classification according to the subject of the study abroad information sharing system according to an embodiment of the present invention by way of example.

A preferred embodiment of the present invention will be described in more detail with reference to the accompanying drawings.

1 is a diagram illustrating a study abroad information sharing system according to an embodiment of the present invention, and FIG. 2 is a block diagram illustrating an information server of the study abroad information sharing system according to an embodiment of the present invention. And Figure 3 is a view for explaining a concept for performing deep learning of the study abroad information sharing system according to an embodiment of the present invention, Figure 4 is for explaining an example of performing deep learning according to an embodiment of the present invention It is a drawing. In addition, FIG. 5 is a view exemplarily showing classification according to the schedule of the study abroad information sharing system according to an embodiment of the present invention, and FIG. It is a diagram showing the classification by way of example.

Referring to FIG. 1 , the study abroad information sharing system according to an embodiment of the present invention includes an information server 100 and a terminal 200 .

The information server 100 is provided to collect information to be provided to international students, and to provide the collected information to the terminal 200 possessed by the international students. To this end, as shown in FIG. 2 , the information server 100 includes a study abroad information collection unit 110 , an international student information unit 120 , a local student information unit 130 , a processing unit 140 , and a matching information guide unit 150 . include

The study abroad information collection unit 110 collects various types of information necessary for international students scattered on the Internet. The study abroad information collection unit 110 searches for and collects information necessary for international students from among various information on the Internet. The information collected by the study abroad information collection unit 110 may be processed by deep learning in the processing unit 140 to be classified according to the subject, and the classified information may be transmitted to the study abroad information collection unit 110 again. .

In this case, the study abroad information collection unit 110 may separately input information collected on the Internet or offline by an administrator in addition to the information searched for online as described above. The information inputted offline in this way is processed by deep learning through the processing unit 140 as necessary, so that classification can be made according to the subject. However, the present invention is not limited thereto, and may be input to the study abroad information collection unit 110 in a state classified according to the subject in the process of being input by the administrator.

The international student information unit 120 stores various information about international students, and may be classified according to unique information such as an ID of an international student, and the information may be stored. In the international student information unit 120 , information may be input through the terminal 200 possessed by the international student, and information on the environment to which the international student belongs may be stored. That is, the international student information unit 120 may store personal information such as the nationality, school, major, and accommodation of the international student, and also information about the international student's interest in countermeasures may be stored.

Various information about international students stored in the international student information unit 120 may be processed through deep learning in the processing unit 140, and if necessary, the processing unit 140 is used to find information matching the international student information unit 120 can be

The local student information unit 130 may store various types of information about local students, and may be classified according to unique information such as an ID of a local student and the information may be stored. That is, the local student information unit 130 stores information about local students who want to provide necessary information to international students. A local student may store information about a school and a major or information on his/her field of interest in the local student information unit 130 for exchange with international students through the terminal 200 possessed by the local student.

For example, when a user (international student or local student) accesses the information server 100 through the terminal 200 and registers as a member, it is possible to select whether an international student or a local student and enter the necessary information. have. Through this process, the corresponding information may be stored in the international student information unit 120 or the local student information unit 130 depending on whether the user is an international student or a local student. In this case, the user whose information is stored in the local student information unit 130 is not limited to only local students, and may be a general public living in the local area, not a student, if necessary.

The processing unit 140 uses the information stored in the study abroad information collection unit 110, the international student information unit 120, and the local student information unit 130 to classify information to be provided to international students based on deep learning according to subject and schedule. The analysis is performed according to the learning model. Accordingly, the processing unit 140 may be configured of hardware including computing resources (CPU, GPU, RAM, storage, etc.) in order to perform such an analysis. In addition, information such as information on the type of vehicle, a parking surface learning model, and an analysis program may be stored in the storage device included in the processing unit 140 .

As described above, the processing unit 140 categorizes the information stored in the study abroad information collection unit 110 , the international student information unit 120 , and the local student information unit 130 according to topic and schedule based on deep learning information about all information. can be checked. However, the present invention is not limited thereto, and deep learning may be performed based on information such as keywords or hashtags for each piece of information.

In this embodiment, the processing unit 140 may use an artificial neural network (ANN) for deep learning, and the artificial neural network is a statistical learning algorithm inspired by neural networks of biology in machine learning and cognitive science. it means. An artificial neural network may refer to an overall model in which artificial neurons (nodes) that form a network by combining synapses change the bond strength of synapses through learning and have problem-solving ability.

In this embodiment, a restricted Boltzmann machine model among artificial neural network models may be used. The constrained Boltzmann machine model is a model proposed by Geoff Hinton as an algorithm that can be used for dimensionality reduction, classification, linear regression analysis, collaborative filtering, feature learning, and topic modeling.

The matching information guide unit 150 is provided in order to provide the information classified by topic and schedule based on deep learning in the processing unit 140 to the terminal 200 possessed by the international student. To this end, the matching information guide unit 150 may generate a classification table according to a subject so that international students can check the necessary information, and may also display information using a calendar according to a schedule.

The terminal 200 is a terminal 200 that a user (international student or local student) can use, and is an electronic device that can access an information book and input information or check information of the information server 100 . In this embodiment, the terminal 200 may be a smartphone, a tablet PC, a notebook computer, or the like.

3 and 4, the deep learning process will be described in more detail using the drawings shown in FIGS. 3 and 4, and in this embodiment, a Boltzmann machine model will be described as an example.

The restricted Boltzmann machine model consists of a visible layer and a hidden layer, the visible layer includes a plurality of visible nodes, and the hidden layer includes a plurality of hidden nodes. And hidden nodes located in all hidden layers are connected to visible nodes located in visible layers, and visible nodes in all visible layers are connected to hidden nodes in hidden layers. That is, in the limited Boltzmann machine model, nodes in the same layer are not connected to each other, but to nodes in different layers.

When data is input, the visible node of the visible layer makes a stochastic decision based on the probability of how much of the input data will be transmitted to the hidden layer. That is, if it is determined that the input data is to be transmitted according to the probability, it is displayed as 1, and if it is determined not to transmit the data, it is displayed as 0.

Referring to FIG. 3 , when the input data v is input to the visible node i of the visible layer, the multiplication value of the input data of the visible node and the weight wij is added to each other. Then, it is sampled to a value of 0 or 1 through an activation function (eg, a sigmoid function), and an output value h is output to the hidden node j.

The limited Boltzmann machine model learns important features of the input data by adjusting the weights through unsupervised learning. Here, the weight is the input data (v), the output value of the hidden node (h), the corrected input data (v') calculated through the reconstruction process, and the corrected output value (h) calculated through the regeneration process ') can be adjusted by calculating the error value (v'h'-vh).

An example of adjusting weights in a constrained Boltzmann machine model will be described with reference to FIG. 4 . As shown in (a) of FIG. 4 , in the reconstruction process, the multiplication value of the output value (h) of each hidden node (j) of the hidden layer and the weight value (wji) are all summed, and then 0 or 1 through the activation function. It is sampled with the value of and output as input data (v') corrected at the input node (i).

As shown in (b) of FIG. 4 , in the reproduction process, the multiplication value of the corrected input data v′ of the visible node i and the weight wij is added to each other and then passes through an activation function to a value of 0 or 1 is sampled and output as the corrected output value (h') in the hidden node (j).

Therefore, the calculation of the output value (h), the corrected input data (v'), and the corrected output value (h') can be expressed as Equations 1, 2 and 3, respectively, as follows.

[Equation 1]

[Equation 2]

[Equation 3]

Here, P is the sampling function, hcj is the output value of the hidden node (j) for the input case (c), Nv is the number of visible nodes, and vci is the input data of the input case (c) input to the visible node (i). , wij is the weight between the visible node (i) and the hidden node (j), bhj is the input bias value of the hidden node (j), and σ is the activation function (eg, a logistic function).

And Nh is the number of hidden nodes, hcj is the output value of the input case (c) output from the hidden node (j), wji is the weight between the hidden node (j) and the visible node (i), and bhi is the visible node ( The output bias value of i), h'cj, is the corrected output value of the hidden node (j) for the input case (c).

Using the calculated output value (h), corrected input data (v'), and corrected output value (h'), weight (wij), output bias (bh), and input bias (bv) are It may be corrected to learn a feature, and the correction of weight, output bias, and input bias may be corrected by Equation 4, Equation 5, and Equation 6.

[Equation 4]

[Equation 5]

[Equation 6]

Here, Nc denotes the number of input cases constituting the input arrangement, and ε denotes a learning rate. In order to increase computation time and learning efficiency in an artificial neural network, weight correction is performed based on an input batch, which is a set of input cases (c) made of a large number of input data.

As described above, using the Boltzmann machine model, the processing unit 140 in this embodiment classifies the information collected through online and offline in the study abroad information collection unit 110 by subject, as shown in FIG. 6 , In addition, as shown in FIG. 5 , information classified according to a schedule may be transmitted to the matching information guide 150 . Accordingly, as the information transmitted to the matching information guide unit 150 is classified and organized by subject, the international student can access the information server 100 and check necessary information according to the subject.

In addition, the international student can access the information server 100 and check information organized according to a required schedule. The information according to the schedule may include, for example, a date of applying for a scholarship according to the academic schedule of a specific school, or information about a health insurance application schedule according to an international student's visa.

As described above, for the schedule classified according to the schedule, the processing unit 140 collects through the international student information unit 120 for the same group of international students, the hobbies of the international student, etc., and matches the schedule according to the collected information through the matching information guide unit ( 150) can be provided to international students.

In this case, if necessary, the information provided to international students through the matching information guide 150 may be expanded to a business associated with a company that conducts a predetermined event.

As described above, the detailed description of the present invention has been made by the embodiments with reference to the accompanying drawings, but since the above-described embodiments have only been described with preferred examples of the present invention, the present invention is limited only to the above embodiments. It should not be understood, and the scope of the present invention should be understood as the following claims and their equivalents.

100: information server
110: Study Abroad Information Collection Department
120: International Student Information Department
130: Local Student Information Department
140: processing unit
150: matching information guide
200: terminal

Claims (5)

information server; and
It is connected to the information server online and includes a terminal used by international students,
The information server,
a study abroad information collection unit that collects information about studying abroad online or enters information related to study abroad through an administrator offline;
an international student information unit that collects information about the international student (information on any one or more of ID, nationality, affiliated school, major, accommodation, and field of interest) about the international student through the terminal used by the international student;
a matching information guide unit that provides information collected through the study abroad information collection unit to international students through the terminal; and
The information collected by the study abroad information collection unit is analyzed based on an artificial neural network, and information matching the information collected through the international student information unit is analyzed, including any one or more of a topic and a schedule, and the analyzed information is analyzed as the matching information Including a processing unit for transmitting to the guide,
The processing unit checks the information on all information to classify the information stored in the study abroad information collection unit, the international student information unit, and the local student information unit according to topic and schedule based on deep learning, or obtain keyword or hashtag information for each information. based on deep learning,
The matching information guide unit generates a classification table according to the subject so that international students can check the necessary information, and also displays the information using a calendar according to the schedule,
The processing unit, but analyzes information using a restricted Boltzmann machine model,
The Boltzmann machine model is
Consists of a visible layer and a hidden layer, the visible layer includes a plurality of visible nodes, the hidden layer includes a plurality of hidden nodes, and the hidden node of the hidden layer is It is connected to visible nodes located in the visible layer (nodes in the same layer are not connected to each other, but to nodes in different layers), and the visible node in the visible layer determines how much data will be transmitted to the hidden layer when data is input. A stochastic decision is made according to the probability, and if the input data is determined to be delivered according to the probability, it is displayed as 1, and if it is decided not to transmit the data, it is displayed as 0,
When the input data (v) is input to the visible node (i) of the visible layer, the multiplication value of the input data of the visible node and the weight (wij) is added to each other, and the value of 0 or 1 is sampled through the activation function to be a hidden node. The output value (h) is output to (j),
The input data is learned by adjusting the weights through unsupervised learning, and the weights are the input data (v), the output value of the hidden node (h), and the corrected input data ( v') and adjusted by calculating the error value (v'h'-vh) from the corrected output value (h') calculated through the regeneration process,
In the reconstruction process, the output value (h) of each hidden node (j) of the hidden layer and the multiplication value of the weight (wji) are all summed up, sampled as 0 or 1 through an activation function, and corrected at the input node (i) output as input data (v'),
In the reproduction process, the multiplication value of the corrected input data (v') of the visible node (i) and the weight (wij) are added to each other and then sampled as 0 or 1 values through the activation function and calibrated again in the hidden node (j). output as an output value (h'),
Calculation of the output value (h), the corrected input data (v'), and the corrected output value (h') is expressed as Equation 1, Equation 2 and Equation 3,
Using the calculated output value (h), the corrected input data (v'), and the corrected output value (h'), the weight (wij), the output bias (bh), and the input bias (bv) determine the characteristics of the input data. Corrected for learning, the correction of weights, output bias, and input bias is corrected by equations 4, 5 and 6, and an input batch that is a set of input cases (c) consisting of a plurality of input data Weight correction is performed based on
[Equation 1]

[Equation 2]

[Equation 3]

(Where P is the sampling function, hcj is the output value of the hidden node (j) for the input case (c), Nv is the number of visible nodes, and vci is the input data of the input case (c) input to the visible node (i). , wij is the weight between the visible node (i) and the hidden node (j), bhj is the input bias value of the hidden node (j), σ is the active function, Nh is the number of hidden nodes, hcj is the input bias value of the hidden node (j). The output value of the output input case (c), wji is the weight between the hidden node (j) and the visible node (i), bhi is the output bias value of the visible node (i), and h'cj is the input case (c). The corrected output value of the hidden node (j) for
[Equation 4]

[Equation 5]

[Equation 6]

(Nc is the number of input cases constituting the input batch, ε is the learning rate)
Further comprising a terminal connected to the information server through online and used by local students,
The information server,
Studying abroad information sharing system further comprising a local student information unit that collects information about local students (information about schools and majors or information about their fields of interest) through the terminal used by the local students.
delete delete delete delete

Images