Some Techniques for Intelligent Searching on Ontology-based

Knowledge Domain in e-Learning

Nhon V. Do

1,*

, Hien D. Nguyen

2,3,*,



and Long N. Hoang

Hong Bang International University, Ho Chi Minh city, Vietnam

Faculty of Computer Science, University of Information Technology, Ho Chi Minh City, Vietnam

Vietnam National University, Ho Chi Minh City, Vietnam

Keywords: Intelligent Searching, Ontology, Knowledge Engineering, e-Learning.

Abstract: E-learning is the modern way to learn by using electronic media and information and communication

technologies in education. Ontology is a useful method to organize knowledge bases for intelligent

educational systems. In this paper, a method for intelligent searching on ontology-based knowledge domain

in e-learning is presented. This method includes an ontology representing educational knowledge domains,

called Search-Onto. The foundation of this ontology is concepts, relations between concepts, operators, and

rules combining the structures of problems and their solving methods. Beside the ontology organizing the

knowledge base, the proposed method also studies some techniques for intelligent searching, such as

searching for the knowledge content, searching on the knowledge classification, and searching the related

knowledge. The method for intelligent searching based on a knowledge base has been applied to construct a

search engine for the knowledge of high-school mathematics. This engine can do searching works and retrieve

required information in mathematics for high-school students to support their learning.

1 INTRODUCTION

Nowadays e-learning is the way in which learning is

imparted to students by using electronic media and

information and communication technologies in

education (Zavyalova, 2020). For the effectiveness of

e-learning, search engines are necessary tools for

finding some information which are required for the

learning process (Veletsianos, 2010). By using this

system, learners can retrieve the knowledge of a

course to study by themselves.

Intelligent search utilizes computers to gather

insight by reading and interpreting all file types like a

human being (Bast et al., 2016, Sajja and Akerkar,

2012). Intelligent search in e-learning is a search

engine on learning resources. This engine is designed

based on a knowledge base organizing the knowledge

domain of courses. It is able to retrieve the knowledge

content matching the meaning of an inputted query,

and anticipate the user’s needs via automatic

https://orcid.org/0000-0002-8527-0602

 Corresponding author

* Equal contribution by Nhon V. Do and Hien D. Nguyen

suggesting the related knowledge to current searching

results.

The requirements of an intelligent searching

system in education are established based on criteria

of an intelligent educational system (Nguyen et al.,

2020b, Hatzilygeroudis and Prentzas, 2006):

 The system has a knowledge base representing

the knowledge of courses completely.

 The system can search information on learning

resources to support the learning. It can analyze

the meaning of a query and get results, which

are suitable for that meaning, from its

knowledge base.

 The system can predict the user’s needs and

propose the knowledge related to current

searching results.

Ontology is a useful method to organize

knowledge bases for intelligent educational systems

(Noy et al., 2013). Ontology COKB (Computational

Objects Knowledge Base) was used to represent the

Do, N., Nguyen, H. and Hoang, L.

Some Techniques for Intelligent Searching on Ontology-based Knowledge Domain in e-Learning.

DOI: 10.5220/0010174403130320

In Proceedings of the 12th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2020) - Volume 2: KEOD, pages 313-320

ISBN: 978-989-758-474-9

313

knowledge domains about discrete mathematics

(Nguyen and Do, 2017) and graph theory (Do et al.,

2018). Rela-model is an ontology representing the

knowledge of relations. This model and its extension

was also applied to represent the knowledge of

courses about Algorithms (Le et al., 2019b). One of

properties of an ontology is the communication

between people and other heterogeneous distributed

application systems (Giri, 2011); thus, ontology is

also effective to apply for represent the knowledge

base of an intelligent search engine (Grosan and

Abraham, 2011), especially the searching systems in

e-learning.

In this paper, an ontology representing an

educational knowledge domain is presented. This

ontology, called Search-Onto, can represent all

components of a course’s intellectual content. The

structure of this ontology is constructed based on the

knowledge of relations and operators (Nguyen et al.,

2020a, Nguyen et al., 2018). Its foundation includes

concepts, relations between concepts, operators, and

rules of the knowledge domain. Some problems,

which are common exercises of this course, and their

solving methods also can be represented by this

ontology. Search-Onto is a useful ontology to build

intelligent search engines. Some techniques for

intelligent searching based on this ontology have been

studied: searching for the knowledge content,

searching on the knowledge classification, and

searching the related knowledge. Besides, the

ontology Search-Onto and searching techniques have

been applied to build an intelligent search engine for

the knowledge of high-school mathematics. This

system organizes this knowledge domain completely.

It can do intelligent search works and retrieve

required information for students to support their

learning.

2 RELATED WORK

There are many methods for building search engines

which are semantic searching systems on text and

knowledge bases (Bast et al., 2016). However, those

systems have not yet equipped a reality knowledge

base with some techniques for intelligent search. For

supporting e-learning, the search engine is an

important function to help learners being able to

retrieve necessary information for their studying.

Docxonomy (2020) is a search engine by

extracting the meaning from files automatically. It

gathers knowledge, understanding, and context by

using artificial intelligence to read and interpret data

just as a human being would. However, this system

has not a knowledge base to store the knowledge

domain, and it is too general to apply in a course of

learning.

Yext’s Intelligent Search Tracker allows users to

track up to keywords based on the data stored in the

Yext Knowledge Manager (Yext, 2020). This tool

helps to understand how a business performs in

intelligent search. The knowledge manager of Yext

contains the information of businesses. It does not

have a reality knowledge. Therefore, this searching

system is only effective for some simple works in e-

commerce.

In (Le et al., 2019a), authors described a method

to build a data advising system, which is an intelligent

recommender system for Data Science. It can search

courses, events, books, and connect users with

employers/businesses in Data Science (DamSanX,

2020). Nonetheless, this system does not have a

knowledge base to support some intelligent search

based on analyzing the query’s meaning. Besides, it

also does not tend to users who are students finding

the knowledge of their course; thus, it cannot support

the learning of students.

The framework of Multimodal Attention-based

Neural Network is applied to find similar exercises in

online education systems (Liu et al., 2018). It is

worked by learning a unified semantic representation

from the large data. However, those systems did not

search the content of knowledge about a course with

the query as natural language.

3 ONTOLOGY FOR SEARCHING

KNOWLEDGE DOMAIN

Fig. 1 shows an architecture of an intelligent search

engine based on the knowledge base. The knowledge

base of this system is extracted from the knowledge

domain collected in the real-world. The user will

input a query to the system. Firstly, the system

analyzes the semantic of query and classifies it to a

searching problem suitably. If the system cannot

understand the query, it will generate some questions

to determine the exactly searching content. Secondly,

the search engine is worked to do some intelligent

searching techniques based on the organized

knowledge base. Finally, the system outputs results

and their related knowledge to the user.

In practical, the knowledge domain of a course has

many components. Thus, the knowledge model for

this domain needs to be performed those components.

One of ontology methods for representing the

knowledge base of an intelligent search system is

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

314

Figure 1: The architecture of an intelligent search engine based on knowledge base.

constructed based on the structure of the knowledge

model of relations combining some other intellectual

components. In this study, this ontology is an

integration between the Rela-Ops model (Nguyen et

al., 2018) and structures of problems with their

methods for solving them (Do et al., 2017).

3.1 Knowledge Model of Relations and

Operators

Rela-Ops model is an ontology representing the

knowledge of relations and operations (Nguyen et al.,

2020a). This model is useful to apply in the practical

knowledge domains, which require the combination

between those kinds of knowledge when solving

problems on them (Nguyen et al., 2020b).

Definition 3.1 (Nguyen et al., 2018): Rela-Ops

model is a tube including 5 components:

(C, R, Ops, Rules)

In which, C is a set of concepts, each concept is a

class of objects. R is a set of other relations between

concepts, objects and objects’ attributes. Those

relations are some special relations, such as “is-a” and

“has-a”, and other binary relations. Ops is a set

consisting of operators on C. Set Rules represents

deductive rules of the knowledge domain. The rules

represent statements, theorems, principles, formulas,

and so forth. The detail of structures of each Rela-

model’s component has been presented in (Nguyen et

al., 2020a).

3.2 Ontology for Searching Knowledge

Domain

On a knowledge domain for searching, besides the

basic knowledge of a course, it also has the

knowledge of exercises and methods for solving them

(Do et al., 2017). Those knowledge is useful for

learners when they want to search its application for

solving some common exercises in that course.

Ontology for searching knowledge domain of a

course is the combination between ontology Rela-

Ops model and the knowledge of common exercises

and their solving methods.

Definition 3.2: Ontology for searching

knowledge domain, called Search-Onto, is a tube:

(C, R, Ops, Rules) + (Problems, Methods)

In which, (C, R, Ops, Rules) is an ontology as

Rela-Ops model; however, each component has been

improved its structure representing its knowledge

more exactly, especially the knowledge for the

interpretation of its meaning.

Problems is a set of problems/exercises of a

course, this component collects all basic exercises

which are characteristics of a course. Methods is a set

of methods for solving problems.

3.2.1 Problems–set of Problems/Exercises

Problems-set is a set of problems/exercises in a

course. Each problem is a tube of five elements:

(Name, Category, Content, Hypo, Goal)

 Name: it is a string used to identify the

problem. Some problems can be referred to by

different names, so this element will only take

the most commonly used keyphrases to identify

the problem as its value.

 Category: the category of the problem.

 Content: describe the problem as HTML text.

 Hypo: The specification of the problem’s

hypothesis.

 Goal: The specification of the problem’s goals.

Some Techniques for Intelligent Searching on Ontology-based Knowledge Domain in e-Learning

315

3.2.2 Methods–set of Methods for Solving

Problems

Methods-set is a set of methods for solving problems.

Each method is bound to solve a specific problem. It

can be represented as a tube of four elements:

(Name, Content, Problem, Illustration)

 Name: it is a string used to identify the method.

Some methods can be referred to by different

names, so this element will only take the most

commonly used keyphrases to identify the

problem as its value.

 Problem: this element represents the problem

which this method is bound to solve.

 Content: it is a HTML text describing the

content of this method.

 Illustration: it is a set of illustrations on some

practical problems for this method.

4 SOME TECHNIQUES FOR

INTELLIGENT SEARCHING

ON ONTOLOGY

There are many requirements for searching on the

knowledge domain. A system of intelligent searching

has the ability to search the knowledge as each kind

of it, such as searching for concepts, relations of

knowledge, or related knowledge to current results.

This section presents some techniques to solve some

problems for intelligent searching on the knowledge

domain based on ontology.

(1) Searching for the content of knowledge: this

searching returns the results based on the meaning of

the user’s query.

(2) Searching on the classification of a knowledge

domain: it returns the results of searching based on

types of knowledge. Some categories of educational

knowledge are: concepts, relations, rules, kind of

problems as well as methods for solving them.

(3) Searching for related knowledge: this kind of

searching returns some knowledge which has

relations with the current searching automatically.

The related knowledge helps users to catch required

information.

4.1 Searching for the Knowledge

Content

The searching of the knowledge content returns a set

of knowledge based on the meaning of an inputted

query. The system determines the meaning of this

query from its extracted keywords. This searching is

worked through the determination of similarly

meaning between keywords of the query and stored

knowledge. There are two problems for designing this

kind of searching.

Definition 4.1 (): Let K be a knowledge domain

as ontology Search-Onto or a component of ontology

Search-Onto, and an object o.

Define: Com(K) is a set of components of K.

key(o) is a set of keywords of o with o

 k and k  Com(K).

Problem 4.1: Let K be an ontology of knowledge

domain as Search-Onto, and o  Com(K) where K =

K or K  Com(K), a keyword w. Determine a

keyword w’  key(o) which is the word has the most

similar meaning with keyword w.

Denote: w’ := Mean(w, key(ob))

Problem 4.2: Let K be a knowledge domain as

ontology Search-Onto, and a query q. Determine a set

of knowledge which are appropriate to the meaning

of query q.

For problem 4.1, the measure for similarly meaning

is computed by the tube (tf(v, o), idf(

v, K)), where tf(v,

o) is the term frequency representing the frequency of

a keyword v in an object o, and idf(v, K) is the inverse

document frequency representing the specificity of

keyword v in knowledge K. The formulas of (tf(v, o),

idf(v, K)) are established as follows (Le et al., 2019b):

,:1



,





′



|′ ∈

(1)

where, n

v,o

is the number of occurrences of the

keyword v in the object o,

c  [0, 1] is a parameter which is the minimum

value for every keywords.

,:



card

⋃



∈



1card



∈|∈





(2)

Let K be a knowledge domain as ontology Search-

Onto, and a query q. This algorithm determines a set

of knowledge which are appropriate to the meaning

of the query q.

Input: The knowledge domain K as Search-Onto.

A query q.

Output: Set of knowledge being appropriate the

meaning of the query q.

Algorithm 3.1:

Step 0: Initialize

W := set of extracted keywords from query q.

Knowledge := {} // set of query results.

If (there exist a keyword about type of

knowledge in W) then

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

316

Search := set of components of K related

to the type of knowledge.

Else

Search := Com(K)

Step 1:

Extract keywords which are closest meaning

with words in W of a query sentence.

:





∈

(3)

Use problem 4.1 to compare per word w  W

with words in Keywords by their meaning

and select the word in Keywords which has

the most appropriate meaning with w.

:



,

∈

(4)

We have, MainWords  KeyWords

Step 2:

Retrieve knowledge from components in

Search based on keywords in MainWords.

Update Knowledge.

Step 3:

Unification of facts in the knowledge model

and the comparison the meaning of them by

using problem 4.1.

Update Knowledge.

Step 4: Return results in Knowledge.

4.2 Searching on the Classification of a

Knowledge Domain

The knowledge domain has many knowledge

components, such as concepts, relations, and

theorems, properties of concepts, objects. Especially,

in the educational knowledge domain, there is the

knowledge of common exercises of that course and

methods for solving them. The searching on the

classification of a knowledge domain helps users to

find necessary information which they need. This

searching is worked based on determining the type of

searched knowledge in a query. For designing this

kind of searching, the algorithm has two stages as

follows.

Stage 1: Determine the category of an inputted

query.

The determination is solved by using the corpus

of the knowledge domain. For example, if the query

is “what is a triangle”, then the category of this query

is the concept. This corpus is collected from the

knowledge source. Through that corpus, the system

can detect the type of searched knowledge. The

process for solving this problem consists of three

main parts of natural language processing: word

segmentation, chunking and selecting candidates

(Pham et al., 2020).

Word segmentation is a task to divide a string of

natural language into its component words. Chunking

is a process to extract phrases from unstructured

sentences. The core idea of this step is the choosing

of the phrases, which have chunking tags: NP (Noun

phrase), VP (Verb phrase) and QP (Question phrase),

as candidate keywords. While NP and QP contain the

key information of queries, QP determines the type of

queries for query classification that will be presented

in subsequent sections. Candidates selection is the

final step of the query processing. Based on the

structure ontology Search-Onto representing the

knowledge K, the heuristic approach is used for

selecting keywords. Using selecting keywords from

the query, the matching of them and each component

of ontology Search-Onto determines the

corresponding category for the knowledge in the

input query.

Stage 2: Matching the searching knowledge and

the knowledge in its category.

From the types determined at Stage 1, the search

engine retrieves the knowledge from components of

ontology Search-Onto. It uses Algorithm 3.1 to find

the content in the corresponding components

matching the information extracted from keywords of

the inputted query.

4.3 Searching for Related Knowledge

The related knowledge is a set of knowledge related

to the current searching knowledge. That related

knowledge help users understanding more the content

of their searching. In this paper, the related

knowledge is a set of knowledge which have relations

with the current searching results in the ontology

Search-Onto representing the knowledge domain.

Let K be a knowledge domain as ontology

Search-Onto, a query q, and the results of query q.

This algorithm determines a set of knowledge related

to the knowledge results of query q.

Input: The knowledge domain K as Search-Onto.

A query q.

Output: Set of knowledge being appropriate the

meaning of the query q.

Algorithm 3.2:

Step 1:

Some Techniques for Intelligent Searching on Ontology-based Knowledge Domain in e-Learning

317

Knowledge := set of knowledge results for the

query q. It is determined by using

algorithm 3.1 or algorithm 3.2,

Step 2:

For each o in Knowledge do

Determine the set of knowledge related to o

Related(o) := {o’|  r



K.R, r(o, o’)};

(5)

Step 3:

Determine the set of knowledge related to the

Knowledge-set:

Classify the knowledge in Relate(Knowledge)

based on types of knowledge of ontology

Search-Onto model.

Step 4: Show results in Relate(Knowledge) by

that classification.

5 INTELLIGENT SEARCH

ENGINE FOR THE

KNOWLEDGE OF

HIGHSCHOOL

MATHEMATICS

Mathematics is an important course for high school

students. The demand of students for supporting to

learn this course is very high, especially when they

want to review for preparing their final exams. Based

on the ontology Search-Onto represented in section 2

and the searching problems in section 4, an intelligent

search engine for the knowledge of high-school

mathematics in Vietnam has been constructed. This

system can support the studying of students via the

supplying of required contents of the course based on

the user’s searching. The system has a knowledge

base which is extracted from this knowledge domain

collected in (Vietnam Ministry of Education and

Training, 2017).

5.1 Design the Knowledge Base

The high-school pupils are the main kind of people

using our system. They will use our program to

support their studying. Knowledge base, which stores

the knowledge domain about high-school

mathematics, is organized by the ontology Search-

Onto as Section 3.2.

(C, R, Ops, Rules) + (Problems, Methods)

There are some examples for geometric knowledge

represented in the searching system.

a) C–set of concepts of geometric knowledge

The set C consists of concepts such as “Point”,

“Segment”, “Line”, “Triangle”, “Quadrilateral”,

“Plane”, “Pyramid”, etc.

b) R–set of relations between concepts

The set R includes two kinds of relations:

- The relations “is-a” between concepts.

- Other binary relations, such as:

 Relations intersect: relations about intersection

between two lines, two planes, a line and a plane, etc.

 Relations parallel (//): relations about parallel

between two lines, two planes, a line and a plane, etc.



Relations perpendicular (): relations about

perpendicular between two lines, two planes, a line

and a plane, etc.

c) Ops–set of operators

In knowledge about plane geometry, some

operators are represented:

Ops := {Projection, Symmetric}

 Projection of a point on a line or a plane,

projection of a line on a plane.

 Symmetrical point of a point through a line.

d) Rules–set of inference rules

Almost all properties, clauses, theorems in high-

school geometry can be represented by structures of

rules in Search-Onto. Some particular rules are:

: {a: segment, b: segment, c: segment},

{a // b, c



a}  {c  b}

: {A: point, B: point, C: point},

{BC = AC}{ABC is an isosceles

triangle at C}.

: {A: point, B: point, C: point},

{AB



BC}{ angle ABC = 90

e) Problems–set of Sample Problems:

There are some common problems about Right

Triangle, Rectangle, Circle, etc., such as:

 Solving a right triangle.

 The problem about determining relations

between diameter and chord of circle.

 Problems about the properties of a pyramid.

f) Methods–set of methods for solving some

problems:

There are some methods for solving problems,

such as:

 Method for solving a right triangle.

 Method for solving an isosceles triangle.

 Method for solving the problem about

determining relations between diameter and chord of

circle.

 Method for solving the problem about

properties of an inscribe quadrilateral.

Relate( ) : Related( ) (6)

o Knowledge

Knowledge o







KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

318

Table 1 shows some topics for algebraic

knowledge at high-school in this system:

Table 1: Some topics for algebraic knowledge at high-

school.

Order Group of topics

“set”, “functions”, “equations”, “inequalities”,

“graph”.

“trigonometric equations”, “trigonometric

functions”, “lines”, “plane”, “derivative”, “limit”,

“vector”.

“differentiation”, “logarithm”, “exponential”,

“analyzing functions”.

Those knowledge of ontology Search-Onto are

organized by the system of structured text files:

 Files store names of concepts and their structures.

 Files store the specification of relations and

operators between concepts.

 A file stores inference rules.

 Files store sample problems.

 Files store the specification of methods for

solving problems.

Example 5.1: For each concept, the corresponding

file has the file name <concept_name>.txt. That file

specifies the structure of the concept. The file’s

structure as follows:

begin_concept: <concept name> [based objects]

specification of based objects

begin_variables

<attribute name> : <attribute type>;

...

end_variables

begin_ relations

begin_relation

specification of a relation

end_relation

...

end_ relations

begin_rules

begin_rule

kind_rule = "<kind of rule>";

hypothesis_part:

{facts}

goal_part:

{facts}

end_rule

...

end_rules

end_concept

5.2 Testing and Experimental Results

The searching system for the knowledge of high-

school mathematics has been tested by 97 students at

two high-schools. There are 31 students studying at

-grade, 29 students at 11

-grade, and 37 students

at 12

-grade.

Each student inputs 03 queries for searching the

knowledge of mathematics. The queries belong to:

searching definitions of mathematics in the curriculum,

searching properties and rules, searching kinds of

exercises and searching the method for solving some

kinds of exercises. The related knowledge has been

shown automatically with the results of the searching. The

results of the inputted queries have been double checked:

• If the student recognizes the results being

suitable for them, they will check on the system.

• Besides, those results are noted. They will be

evaluated by the lecturers in mathematics later.

• If the results are checked by students and

lecturers, they will be correct.

The experimental results are shown in Table 2 and

Fig. 2. In this table, some duplicate queries were omitted.

Table 2: Experimental results on tested queries.

Search content

Number of

queries

Correct Rate

Definitions in

mathematics

101 81 80.2%

Properties, rules 62 42 67.7%

Kinds of exercises 59 43 72.8%

Solving methods 53 38 71.7%

Total 275 204 74.2%

Figure 2: Compare the results between search contents.

Students usually search definitions of

mathematics and their properties. They have a need to

understand the knowledge more clearly. Concepts in

mathematics are separate, the searching on each

definition is more exactly than others. When

searching for rules and properties, because there are

many rules which have the same kinds and same

objects, those queries do not return the correct results.

6 CONCLUSION AND FUTURE

WORK

This article presents an ontology representing the

knowledge base of an intelligent educational

Some Techniques for Intelligent Searching on Ontology-based Knowledge Domain in e-Learning

319

searching system, called Search-Onto. This ontology

is built based on the knowledge model of relations

and operators combining the structures of problems

and their solving methods. Besides, some techniques

for intelligent searching have been designed based on

the knowledge base. Those kinds of searching are:

Searching for the knowledge content, searching on

the classification of knowledge and searching for

related knowledge.

Ontology Search-Onto has been applied to

represent the knowledge domain about high-school

mathematics for an intelligent searching system on

this domain. The built system can retrieve required

information matching the meaning of an inputted

query. The system can get some related knowledge to

the search content. It is helpful to support high-school

students in learning mathematics.

In the future, the studying of combining an

intelligent search system and an architecture of

chatbot (Nguyen et al., 2019) will build an intelligent

chatbot for question-answering the knowledge of

courses. This chatbot can interact and help users more

naturally. Moreover, the intelligent search system in

e-learning will be integrated to the intelligent problem

solver in education (Nguyen et al., 2020a) and the

intelligent supporting system for multiple choices

training test (Mai et al., 2018). The integrated system

will be a complete system to support the learning of

students effectively.

REFERENCES

Bast, H., Buchhold, B., Haussmann, E., 2016. Semantic

Search on Text and Knowledge Bases. Information

Retrieval 10(2-3), 119 - 271.

DamSanX, 2020. https://damsanx.com/ (Access on 27 Aug. 2020)

Docxonomy, 2020. https://docxonomy.com/intelligent-

search/ (Access on 27 Aug. 2020)

Do, N, Nguyen, H., Mai, T., 2018. Intelligent Educational

Software in Discrete Mathematics and Graph Theory.

In SOMET 2018, 17

International Conference on

Intelligent Software Methodologies, Tools, and

Techniques, Sept. 2018.

Giri, K., 2011. Role of ontology in semantic web.

DESIDOC Journal of Library Information Technology

31, 116–120.

Grosan, C., Abraham, A., 2011. Intelligent Systems: A

modern approach, Springer, Berlin, Heidelberg.

Hatzilygeroudis, I., Prentzas, J., 2006. Knowledge

Representation in Intelligent Educational Systems. In

Web-Based Intelligent E-Learning Systems:

Technologies and Applications, IGI publisher.

Le, P., Nguyen, L., Nguyen, T., Vo, K., Hoang, S., 2019a.

Toward a Smart Platform for Data Science Carreer. In

KSE 2019, 11

IEEE International Conference on

Knowledge and Systems Engineering, Oct. 2019.

Le, T., Luu, S., Nguyen, H., Do, N., 2019b. Knowledge

representation method for designing an Intelligent

Tutoring System in Learning of courses about

Algorithms. In APCC 2019, 25

Asia-Pacific

Conference on Communications, Nov. 2019.

Liu, Q., Huang, Z., Huang, Z., Liu, C., Chen, E., Su, Y., Hu,

G. 2018. Finding similar exercises in online education

systems. In KDD 2018, ACM SIGKDD International

Conference on Knowledge Discovery and Data Mining,

London, UK, Aug. 2018.

Mai, T., Nguyen, H., Le, T., Pham, V., 2018. An Intelligent

Support System for the Knowledge evaluation in high-

school mathematics by Multiple choices testing. In

NICS 2018, 5

NAFOSTED Conference on Information

and Computer Science, Nov. 2018.

Nguyen, H.D., Do, N.V., 2017. Intelligent Problems Solver in

Education for Discrete Mathematics. In SOMET 2017, 16

International Conference on Intelligent Software

Methodologies, Tools, and Techniques, Sept. 2017.

Nguyen, H.D., Do, N.V., Pham, V.T., 2018. Rela-Ops

model: A method for Knowledge Representation and

Application. In SOMET 2018, 17

International

Conference on Intelligent Software Methodologies,

Tools, and Techniques, Sept. 2018.

Nguyen, D.H., Pham, T.V., Tran, A.D., Le, T., 2019.

Intelligent tutoring chatbot for solving mathematical

problems in High-school. In KSE 2019, 11

IEEE

International Conference on Knowledge and Systems

Engineering, Oct. 2019.

Nguyen, H., Do, N., Pham, V., Selamat, A., Herrera-

Viedma, E., 2020a. A method for knowledge

representation to design Intelligent Problems Solver in

mathematics based on Rela-Ops model. IEEE Access 8,

76991–77012.

Nguyen, H., Do, N., Tran, N., Pham, X.H., Pham, V., 2020b.

Some criteria of the Knowledge Representation method

for an Intelligent Problem Solver in STEM education.

Applied Computational Intelligence and Soft

Computing 2020, Article ID 9834218.

Noy, N., et al., 2013. Final Report on the 2013 NSF

Workshop on Research Challenges and Opportunities

in Knowledge Representation. National Science

Foundation Workshop Report.

Pham, T., Tran, V., Nguyen, T., Pham, V.T., Nguyen, H.,

2020. Build a search engine for the knowledge of the

course about Introduction to Programming based on

ontology Rela-model. In KSE 2020, 12

IEEE

International Conference on Knowledge and Systems

Engineering, Nov. 2020. Accepted.

Sajja, P., Akerkar, R., 2012. Intelligent technologies for

web applications. CRC Press.

Veletsianos, G., 2010. Emerging Technologies in Distance

Education. Au Press.

Vietnam Ministry of Education and Training, 2017.

Textbooks and workbooks of mathematics in high

school, Publisher of Education.

Yext, 2020. https://www.yext.com/ (Access on 27 Aug. 2020)

Zavyalova, K., 2020. Unlocking students’ motivation in the

blended higher education classroom: Lecturers’

perspectives. E-Learning and Digital Media 17(5), 425–441.

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

320