International Journal of Technical Research and Applications e-ISSN: 2320-8163,

www.ijtra.com Volume-2, Special Issue 1 (July-Aug 2014), PP. 46-50

46 | P a g e
CONCEPT DIAGRAM OF ALGEBRA LEARNING

YUAN-HORNG LIN
Department of Mathematics Education,
National Taichung University of Education
Taichung, Taiwan
JENG-MING YIH
Center of General Education,
Min-Hwei College of Health Care Management
Tainan, Taiwan
AbstractEach cluster of data can easily describe features of
knowledge structures. The purpose of this study is to analyze the
concept diagram of Abstract Algebra concepts for university
students and clustering based on concept proficiency. Besides,
fuzzy clustering on concept proficiency expresses the cognitive
characteristics. Methodology in this study is CAISM. it shows
CAISM can provide useful information for cognition diagnosis.
According to the results, some suggestions and recommendations
for future investigation are discussed.
I ndex TermsConcept Structure, CAISM, Cognition
Diagnosis.
I. INTRODUCTION AND MOTIVATION
Empirical data is the paper-and-pencil assessment of
Abstract Algebra course. This method can not only present the
personal concept structure by hierarchical diagram, but also
calculate the magnitude of mastery on each concept. Concept
structure analysis (CSA) could provide individualized
knowledge structure. CSA algorithm is the major methodology
and it is based on fuzzy logic model of perception (FLMP) and
interpretive structural modeling (ISM). CSA could display
individualized knowledge structure and clearly represent
hierarchies and linkage among concepts for each examinee.
Concept diagram is one method of assessment and it will
improve remedial instruction. One is to analyze total score and
the other is to analyze concept diagram. Total score of test is
adopted by classical assessment to indicate achievement of
students. However, this method has some shortcomings. For
example, total score is hard to provide information of cognitive
diagnosis and total score can not distinguish features of
concepts involved in the items. Therefore, analysis of total
score exist limitations. Moreover, task-takers of same total
score with different response pattern have distinct concept
diagram.
Teaching and learning is an interactive process between
teachers and students. Many courses of abstract algebra are
involved in teaching Abstract Algebra either formally in a
college classroom or informally in an industrial setting.
Regardless of the setting, a major concern of those who teach
statistics in how to ensure that the student understand abstract
algebras ideas and are able to apply what they learn to real-
world situations. Although teachers of abstract algebra often
express frustration about difficulties students have learning and
applying course material, many may be unaware of the growing
body of research related to teaching and learning abstract
algebra. It is needed to design a systematic way to realize
student's understanding about their concept diagrams. However,
each task-taker should have his own knowledge structure based
on the viewpoint of cognition science.
Clustering base on concept proficiency in knowledge
structures should be helpful for remedial instruction. Therefore,
methodological development for both individualized concept
structure and clustering is an important issue. In this study,
foundations of clustering operation from fuzzy theory will be
used. The empirical data is abstract algebras test.
In this paper, Knowledge management of abstract algebras
Concepts was essential in educational environment. The
purpose of this study is to provide an integrated method of
fuzzy theory basis for individualized concept structure analysis.
This method integrates Fuzzy Logic Model of Perception
(FLMP) and Interpretive Structural Modeling (ISM). The
combined algorithm CAISM (concept advanced interpretive
structural modeling) can reveal individualized concept diagram
[1] [2] [3], only brief discussion is considered in this study.
Fuzzy clustering based on proficiency of concepts, which is
derived from results of individualized concept structure, is
adopted so that task-takers will be properly clustered. Students
of the same total score with different response pattern will also
be chose and discussed. Results of this research are prospective
on the cognition diagnosis for courses of abstract algebra
learning and help teachers design remedial instruction.
However, CAISM could analyze individualized concepts
structure based on the comparisons with concept structure of
expert. It has great benefits if analysis of concept diagram is
adopted. How to analyze concept diagram is a prospective
methodology. It is because diagram could be easily understood
and individualized concept diagram can improve realization for
characteristics of concepts for each task-taker.
Cognition diagnosis can help represent knowledge structure
[4] [5] [6]. It is a common viewpoint that human knowledge is
stored in the form of structural relationship among concepts
and their subordinate relationship is fuzzy, not crisp. There are
some methodologies for concept structure analysis but little is
known about methodologies of individualized concept structure
[7] [8] [9] [10]. Therefore, the development for methodology of
individualized concept structure is an important issue and it is
essential for cognition diagnosis and pedagogy [11].
II. LITERATURE REVIEW
In this study, the integrated method of individualized
concept structure based on fuzzy logic model of perception
(FLMP) and interpretive structural modeling (ISM) will be
developed. An example of empirical test data of abstract
algebras concept for students of learning deficiencies will also
be analyzed and discussed. For the feasibility of remedial
instruction based on the cognition diagnosis, clustering method
is needed so that students within the same cluster own similar
knowledge structures and students among different clusters
have the most variance on knowledge structures [12] [13].
The algorithm of interpretive structural modeling is the
foundation of this study and other topics will also discuss.
A. Interpretive Structural Modeling
The algorithm of interpretive structural modeling is the
foundation of this study. They will be discussed as follows.
Moreover, issue on the statistics assessment will be also
discussed. The theory of interpretive structural modeling
(ISM) is based upon discrete mathematics and graph theory. J.
N. Warfield provided ISM and it aims to arrange elements in
a hierarchical relation. For any set that contains elements, we
International Journal of Technical Research and Applications e-ISSN: 2320-8163,
www.ijtra.com Volume-2, Special Issue 1 (July-Aug 2014), PP. 46-50
47 | P a g e
CSA
AMC ASC AFISM
can make a hierarchical graph of all elements if the binary
relationship among elements is known. Namely, the
relationship of
i
A and
j
A must be acquired in advance.
The relationship could be expressed in the form of
matrix
K K ij
a A

) ( . If 1
ij
a exists,
i
A is the precondition
of
j
A . On the other hand, if it is 0
ij
a ,
i
A is not the
precondition of
j
A . The analytical procedure of ISM is as
follows.
The ISM adopts Boolean operation. The transitive closure is
P
A A A A A
3 2
and
P
I A I A R ) (

is reachability
matrix. With transitive closure A

and reachability matrix R ,

the hierarchical graph of elements in matrix
K K ij
a A

) ( could
be plotted. For example, let the
K K ij
a A

) ( be as follows. Its
relationship and hierarchies of elements is depicted in Fig. 1.



Fig. 1 The linkage of elements in hierarchies
B. Fuzzy Logic Model of Perception
Suppose there be a combination of two factors and there are
levels within each factor. There are levels and levels for
factor and respectively. The fuzzy true values are expressed.
Fuzzy truth value and express the degree that the combination
of and will support prototype [14] [15]. The probability that
the combination of could be viewed as prototype can be
expressed as follows [16]. The fuzzy true value is to express
the degree that the combination of two levels from distinct
factor will support prototype. The probability that this
combination could be viewed as prototype is as follows
1
)] 1 )( 1 ( [ ) , (


j i j i j i j i
o c o c o c o c p (1)
C. Clustering and Fuzzy Theory
Cluster analysis is a data analysis tool to find proper clusters
of a data set with the most homogeneity in the same cluster
and the most heterogeneity among clusters. Statistics
approaches include two categories of clustering, which are
hierarchical and non-hierarchical. The k-means algorithm is
a popular example of non-hierarchical clustering whereby
the data is partitioned into -classes with the value of known
a priori. Zadeh provides the fuzzy theory and it flourishes
well and bring contributions to clustering analysis. Fuzzy
clustering is based on the membership of sample points and
fuzzy c-means is the most popular technique. In this study,
Bezdeks algorithm is applied.
Suppose there be a data matrix with subjects and variables.
The data matrix M N nm x X is known. The membership
matrix N C cn u U and the group center matrix M C cm v V
are unknown with group number C . The objective function
with optimization problem is applied and membership matrix
and group center matrix are derived by way of iteration.
) , ( ) ( ) , (
1 1
2
n c
d u
V U J
N
n
C
c
cn
q

(2)

C
l
q
cn
n l d
n c d
u
1
2
2
1
1
] [
) , (
) , (
1
and

N
n
cn
q
N
n
nm cn
q
cm
u
x u
v
1
1
) (
) ( ) (
(3)
Partition entropy ) ( ln
1
) ; (
1 1
cn
N
n
C
c
cn u u
N
C U H

and partition
coefficient

N
n
C
c
cn
q
u
N
C U F
1 1
) (
1
) ; ( are used to decide the
number of cluster. Larger partition coefficient and smaller
partition entropy means the better partition number C will be
used.
D. Integrated model and Procedure
The procedure of the integrated model is depicted in Figure 2.
Firstly, FCM is to classify examinee based on their response
pattern. Secondly, concept structure analysis (CSA) will
analyze individualized knowledge structures. CSA includes
three algorithms, which are AMC (algorithm for mastery of
concept), ASC (algorithm for subordination of concepts) and
AFISM (algorithm for fuzzy ISM). By the integrated
procedure, Examinee within the same cluster represent
similar knowledge structures and remedial instruction could
be feasible based on the information of cognition diagnosis
within the same cluster.












Fig. 2 Procedure of the Integrated Model

Three algorithms, AMC, ASC and AFISM, are combined in
order to analyze individualized knowledge structure. Some
basic definitions are as follows [1]. The integrated algorithms
consist of three steps of algorithms, AMC, ASC and AFISM.
E. Abstract Algebras Education and Assessment
In this study, the authors provide the empirical data for
concepts of the course of abstract algebra from university
students. To sum up, the methodology can improve knowledge
management in classroom more feasible. It has great benefits
if analysis of concept diagram is adopted. However, each task-
taker should have his own knowledge structure based on the
viewpoint of cognition science. Teaching and learning is an
interactive process between teachers and students.
Students learn by constructing knowledge and active
involvement in learning actives. It should not underestimate
the difficulty students have in understanding basic concepts of
FCM
International Journal of Technical Research and Applications e-ISSN: 2320-8163,
www.ijtra.com Volume-2, Special Issue 1 (July-Aug 2014), PP. 46-50
48 | P a g e
abstract algebra. It is often neglected how students understand
basic concepts. Students will learn better if they receive
consistent and helpful feedback on their performance.
Therefore, it is important to evaluate what students know.
Learning will be enhanced by investigating concept diagrams
so that misconception and mastery of concepts will be
understood. A better understanding of students learning not
only can be important references in determining success in
abstract algebra, but also be beneficial to both teachers and
students. Assessment will helps teachers understand the
learning styles of students and adjust their own instructions
accordingly. Furthermore, adaptively remedial instruction
could depend on the information of students shortcomings. In
addition, students will also benefit by managing their learning
process more efficiently. However, most researches related to
abstract algebra learning and measurement adopt descriptive
methods. Little is known about the analysis of knowledge
structures. Therefore, the methodology to analyze concept
diagrams about university students should be prospective.
In this paper, abstract algebra consists of three parts.
(A).Group Theory:
1.Understand the definitions and the properties of groups,
and have the ability to distinguish different groups.
2.Understand the definitions and the properties of
subgroups.
3.Understand the definitions and the properties of
homomorphism and auto-morphism.
(B).Ring Theory:
1.Understand the definitions and the properties of rings.
2.Understand the definitions and the properties of integral
domains, division rings, fields, ideals and quotient rings.
3.Understand the definitions and the properties of
homomorphism and auto-morphism, and know some
familiar ideals and rings.
4.Understand the definitions and the properties of
polynomial rings.
(C).Field Extension and Galois Theory:
1.Understand the definitions and the properties of field
extensions.
2.Know some familiar field extensions.
3. Understand the significance of the fundamental theorem
of Galois theory.
F. METHODOLOGY AND ALGORITHM
In t.
Fig. Fig. 3 The process of CAISM algorithm

In this study, CAISM is developed to integrate method of
fuzzy logic model of perception (FLMP) and interpretive
structural modeling (ISM). The method analyzes concept
diagram based on response data matrix and item attribute
matrix. It will reveal the individualized concept diagram and
relationship. The process of CAISM algorithm is depicted in
Fig. 3
III. EMPIRICAL ANALYSIS
The course of abstract algebra assessment is designed by the
authors. There are 251 university students and take elementary
linear algebra and advance of abstract algebra courses. There
are 10 concept attributes within each item and they are depicted
in Table1.
TABLE I. CONCEPT ATTRIBUTES OF TEST
Concepts Concept Attributes
1 Groups and distinguish different groups
2 The properties of subgroups
3 Homomorphism of group
4 The Properties of rings
5 Division rings, and quotient rings
6 Homomorphism of ring
7
8
9
10
The properties of polynomial rings.
Properties of field extensions.
Know some familiar field extensions.
Fundamental theorem of Galois theory
The test includes 25 items. The data set comes from the
National Taichung University of Education used in the
International Journal of Technical Research and Applications e-ISSN: 2320-8163,
www.ijtra.com Volume-2, Special Issue 1 (July-Aug 2014), PP. 46-50
49 | P a g e
empirical study with learning Abstract Algebra. The item
concept matrix is depicted in Table 2. For example, from item
1 to item 7, these items measure concept 1. From item 3 to item
7, these items measure concept 2. From item 5 to item 7, these
items measure concept 3. Similarly, From concept 1 to concept
10, they are random sampling, sampling distribution, sampling
distribution of mean, central limit theorem and sampling
distribution of ratio. All these items are dichotomous and is
decided in the CAISM algorithm.
TABLE II. ITEM CONCEPT MATRIX
Items
Concept
Items
Concept
1 2 3 4 5

6 7 8 9
1
0
A1 1 0 0 0 0 B7 1 0 0 0 0
A2 1 0 0 0 0 B8 1 0 0 0 0
A3 1 1 0 0 0 B9 1 0 0 0 0
A4 1 1 0 0 0 B10 0 1 0 0 0
A5 1 1 1 0 0 B11 0 1 1 0 0
A6 1 1 1 0 0 C1 0 1 1 1 0
A7 1 1 1 0 0 C2 0 0 1 0 0
B1 0 0 0 1 0 C3 0 0 1 0 0
B2 0 0 0 0 1 C4 0 1 1 0 1
B3 0 0 0 0 1 C5 0 0 1 0 1
B4 0 0 0 1 1 C6 0 1 1 1 1
B5 0 0 0 1 1 C7 0 1 1 1 1
B6 0 0 0 1 1

The data of task-takers ID 106 is depicted in Table 3. In
Fig. 4, the symbol of each concept and magnitude of mastery
on concept of task-taker ID 106 is shown. In task-taker ID 106,
the mastery of all concepts except concept 10 is 1.00 and these
are equivalent relation. These concepts are the precondition of
concept 10. From the item concept matrix of table 2, we can
Obtain verification.


TABLE III. RESPONSE PATTERN AND TOTAL SCORE OF ID 106
ID Response Pattern Total Score
106 1111111111111111111110000 21



Fig. 4 Concept diagram of task-taker ID 106

In task-taker ID 76, the mastery of all concepts except
concept 1, concept 2 and concept 3 is 1.00 and these are
equivalent relation. These concepts are the precondition of
concept 1, concept 2 and concept 3. The mastery of concept
and concept 2 is 0.50 and these are equivalent relation. These
concepts are the precondition of concept 1.
TABLE IV. RESPONSE PATTERN AND TOTAL SCORE OF ID 76
ID Response Pattern Total Score
76 0000111111111111111111111 21



Fig. 5 Concept diagram of task-taker ID 76

Compare to task-taker ID 76 and task-taker ID 106, they
get same total score, but Concept diagram of task-taker ID 76
and task-taker ID 106 is different, It implies that difficulties of
these concepts are quite different. Although CAISM could
provide the concept diagram of each task-taker.
To sum up, this integrated algorithm could improve the
assessment methodology of cognition diagnosis and manage
the knowledge structures of abstract algebras concepts easily.
In task-taker ID 9, the mastery of concept 8 is .51. It is
the basic concept of all concepts. For task-taker ID 9, The
mastery of concept 7 is 0.46. The precondition of concept 7 is
concept 1, concept 2, concept 3, concept 4, concept 6, concept
9 and concept 10 and Only concept 1, concept 2 and concept 3
are equivalent relation. The concept 5 is isolated concept.
TABLE V. RESPONSE PATTERN AND TOTAL SCORE OF ID 9
ID Response Pattern Total Score
9 0110110101100001000110100 11



Fig. 6 Concept diagram of task-taker ID 9

In task-taker ID 205, the mastery of concept 1 is .52. It is
the basic concept of all concepts. For task-taker ID 205, The
mastery of concept 8 is 0.45. The precondition of concept 8 is
concept 2, concept 5, concept 6 and concept 9. None concepts
is equivalent relation. The concept 4 is isolated concept.
TABLE VI. RESPONSE PATTERN AND TOTAL SCORE OF ID 205
ID Response Pattern Total Score
205 1011111110110000001010010 13

International Journal of Technical Research and Applications e-ISSN: 2320-8163,
www.ijtra.com Volume-2, Special Issue 1 (July-Aug 2014), PP. 46-50
50 | P a g e


Fig. 7 Concept diagram of task-taker ID 205

Compare to task-taker ID 9 and task-taker ID 205,
Concept diagram of task-taker ID 9 and task-taker ID 205 is
different, Knowledge Management of Mathematics Concepts
was essential in educational environment. The purpose of this
study is to provide an integrated method of fuzzy theory basis
for individualized concept structure analysis. In the task-taker
ID 9 of Fig. 6, concept 8 is more easily than others. But In the
task-taker ID 205 of Fig. 7, concept 8 is more difficultly than
others.
In addition, empirical test data of abstract algebra for
university students are discussed. It shows that knowledge
structures will be feasible for remedial instruction [17].
This procedure will also useful for cognition diagnosis.
To sum up, this integrated algorithm could improve the
assessment methodology of cognition diagnosis and manage
the knowledge structures of Abstract Algebra Concepts easily.
An example of empirical test data of linear algebra
concept for students of learning deficiencies will also be
analyzed and discussed. For the feasibility of remedial
instruction based on the cognition diagnosis, clustering method
is needed so that students within the same cluster own similar
knowledge structures and students among different clusters
have the most variance on knowledge structures [18].
IV. CONCLUSIONS
However, each task-taker should have own knowledge
structure based on the viewpoint of cognition science. In this
paper, each cluster of data can easily describe features of
knowledge structures. We can manage the knowledge structures
of Abstract Algebra Concepts to construct the model of features
in the pattern recognition completely. An integrated method of
FLMP and ISM for analyzing individualized concept structure is
provided. With this integrated algorithm, the graphs of concept
structures will display the characteristics of knowledge structure.
This result corresponds with foundation of cognition diagnosis in
psychometrics [11]. The correct ration of concepts varies. It
implies that difficulties of these concepts are quite different.
Although CAISM could provide the concept diagram of each
task-taker, it is unfeasible to display the concept diagrams of all
task-takers respectively.
As shown in Table 3 and Table 4, two task-takers , ID 106
and ID 76, have the same total score 21 with different response
pattern. As shown in Fig. 4 and Fig. 5, these two task-takers have
varied concept diagrams. In addition to the distinction on mastery
of concepts, the linkages and prerequisite relationship among
concepts also differ.
This phenomenon supports viewpoints of cognitive
psychology and psychometrics that response pattern could
distinguish characteristics of concept diagram, but not total score.
An integrated method of individualized concept structure and
fuzzy clustering on proficiency of concepts could correspond
with issue of cognition diagnosis in psychometrics. This study
investigates the concept diagrams of university students. One is
concluded that the study help understand personal concepts
diagram. Thus, teachers could design the adaptively remedial
instruction. Another finding is that concept diagram reveals
personal misconceptions or erroneous linkage among concepts.
Future research could apply CAISM to another fields or subjects.
Furthermore, it is prospective to integrate CAISM with
instructional technology so that e-learning approach of remedial
instruction becomes feasible.
ACKNOWLEDGMENT
This paper is partially supported by Min-Hwei College of
Health Care Management of Taiwan, The projects title is
Construct Remedial Instructions Interface for the Sample of
Nursing Background to Learning Foundational Mathematics in
Longitudinal Study ( NSC 103).
REFERENCES
[1] J. N. Warfield, Societal Systems Planning(1976). Policy and
Complexity, Wiley.
[2] J. N. Warfield, Interpretive Structural Modeling (ISM). In S. A.
Olsen (Eds.), Group Planning & Problem Solving Methods in
Engineering (1982).pp.155-201, Wiley,
[3] L. A. Zadeh, Information and Control (1965). Vol. 8, p.338.
[4] M. Smithson, and J. Verkuilen, Fuzzy Set Theory: Applications
in the Social Sciences, Sage Publications (2006).Thousand Oaks,
CA.
[5] M. Smithson, and J. Verkuilen, Fuzzy Set Theory: Applications
in the Social Sciences, Sage Publications (2006).Thousand Oaks,
CA.
[6] T. Sato, The S-P Chart and The Caution Index, NEC
Educational Information Bulletin 80-1, C&C Systems Research
Laboratories (1980). Nippon Electic Co., Ltd,. Tokyo, Japan.
[7] J. P. Doignon and J. C. Falmagne, Knowledge Space(1999).
Springer-Verlag.
[8] K. VanLehn, Journal of the Learning Sciences(1999). Vol.8,
p.71.
[9] R. W. Schvaneveldt, Pathfinder Associative Networks (1991).
Ablex.
[10] W. P. Jr. Fisher, Rasch Measurement Transactions (1995).
Vol.9, p. 442.
[11] R. J. Mislevy and N. Verhelst, Psychometrika (1990). Vol. 55,
p.195.
[12] Y. H. Lin, M. W. Bart, and K. J. Huang, Generalized
Polytomous Ordering Theory(2006). [manual and software],
National Taichung University, Taiwan.
[13] T. Sato, Introduction to S-P Curve Theory Analysis and
Evaluation (1985). Tokyo, Meiji Tosho.
[14] G. Klir and B. Yuan, Fuzzy Sets and Fuzzy Logic, Theory and
Applications(1995). Prentice Hall.
[15] D. W. Massaro and D. Friedman, Psychological Review(1990).
Vol. 97, p.225.
[16] C. S. Crowther, W. H. Batchelder and X. Hu, Psychological
Review (1995). Vol.102, p.396.
[17] K. K. Tatsuoka, and M. M. Tatsuoka, Computerized Cognitive
Diagnostic Adaptive Testing: Effect on Remedial Instruction as
Empirical Validation, Journal of Educational
Measurement(1997).Vol. 34, , pp. 3-20.
[18] Y. H. Lin, M. W. Bart, and K. J. Huang, Generalized polytomo-
us Ordering Theory(2006). [manual and software], National
Taichung University, Taiwan.