CONCEPTS NETWORK

Text Mining Technology allied with Pedagogical Practices to Qualify the Collective

writing Process in Distance Learning

Alexandra Lorandi Macedo, Patricia Alejandra Behar and Eliseo Berni Reategui

PPGEDU/PPGIE - UFRGS, Av. Paulo Gama, 110, 90040-060, Porto Alegre, RS, Brazil

Keywords: Collective Text Editor, Text Mining, Pedagogical Practices.

Abstract: The main goal of this article is to identify possibilities of pedagogical practices from the information

generated by a tool called the Concepts Network. Such information is extracted from texts produced by

participants / students in a Collective Text Editor - ETC. Thus, this study presents a tool and, from it,

alternative teaching practices. The intention is to create conditions so that such practices are developed with

quality and minimize the large work load required in digital work spaces in distance learning practices.

1 INTRODUCTION

This article presents the development of a study that

had as its objective the construction of a tool which

allows teachers to monitor the process of collective

construction of texts in the Collective Text Editor -

ETC (available at: www.nuted.ufrgs.br/etc2). The

ETC is an online editor, which was developed with

an educational purpose by the Centre of Digital

Technology applied to Education at the Federal

University of Rio Grande do Sul (NUTED /

UFRGS). The first version was built in 2001 and the

system has since been applied and perfected,

boasting, as a reference, validations and evaluations

made by teachers and student users. Some of the

principal adjustments and implementations

developed during its application can be checked in

Macedo et al. (2009). In this period, it was also

noted that teachers frequently reported the difficulty

they had in following the process of students' written

construction, due to the high volume of data

generated in this practice.

From this context began the development of this

study, which developed a tool to support the

monitoring of collective text production in the ETC.

Therefore, the next sections describe the theoretical

foundation that supported the educational

perspective in the use of the tool. The developed tool

is also presented along with the validation process,

the possible applications from the results obtained

and the final considerations.

2 THEORETICAL PERSPECTIVE

TO SUPPORT EDUCATIONAL

PRACTICE

Advances in science and technology are ongoing and

take place in an ever more dynamic form. In this

context begins the teacher’s challenge, which is to

develop strategies that enable the formation of a

subject that takes account of this complexity. In

order to achieve this, on the basis of studied theory,

the following assumptions were elected in this

research: the construction of autonomy, critical

thinking, the process of learning to learn and

knowing how to articulate oneself in collective

contexts, cooperating and communicating with all

involved in the process.

The process of the student's personal

development involves, among other elements, the

development of autonomy. It is this that may offer

conditions to allow the subject to realize the

complexity and the challenges they encounter

throughout the learning process.

For Piaget (1973), autonomy is directly related to

the type of respect that the subject has for rules or

standards. According to the author, there are two

kinds of respect: heteronomy and autonomy.

For Piaget (1994), only cooperation leads to

autonomy. Thus, it is necessary to favor cooperation

in the educational context, not only restricting the

social exchange between teacher and student, but

also encouraging the exchange among peers as "[...]

Lorandi Macedo A., Alejandra Behar P. and Berni Reategui E..

CONCEPTS NETWORK - Text Mining Technology allied with Pedagogical Practices to Qualify the Collective writing Process in Distance Learning.

DOI: 10.5220/0003801500710076

In Proceedings of the 4th International Conference on Computer Supported Education (CSEDU-2012), pages 71-76

ISBN: 978-989-8565-07-5

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

criticism is born out of discussion and discussion is

only possible among equals, therefore, only

cooperation will realize what intellectual constraint

is incapable of realizing" (Piaget, 1994, p. 298-299).

Based on these conditions, it is understood that

to achieve the highest levels of autonomy

development, it is necessary that, throughout the

educational process, teachers and students take on

distinct actions in which the former must create for

the latter, ever more challenging situations for

resolution and development, offsetting practices

which are limited to reproduce or copy. The

intention is for the students to be able to support

their actions within the principles of critical thinking

about their own process. In this study, critical

thinking is considered as a sense of experience and

need for logical consistency that are placed in the

service of an autonomous reasoning, common to all

individuals and not depending on any external

authority (Piaget, 1998). Thus, social exchange acts

as a tool that encourages creativity and critical spirit.

The construction of critical spirit is fundamental to

the intellectual and social development of the

individual (Piaget, 1986). It is from the social

relations which follow the cognitive conflict that

subjects are enticed to question, doubt and criticize

the different points of view and from there are

motivated to propose solutions and alternatives that

are shown as more viable for conflict resolution.

This set of elements shows the complexity of

students' reflection, the complexity that is found in a

critical thought (Parrat-Dayan, 2007). Thus, Piaget’s

perspective (1986) indicates that pedagogical

practice must insist on the exchange of different

points of view in order to provide mutual enrichment

between subjects. Such a perspective consists of

leading each individual to think for themselves and

position themselves relative to another.

Different challenges can make the subject feel

the need to seek elements, information that will help

him/her to realize the given situation. The measure

to which the student can realize their limitations,

procure and articulate new information to enable

him/her to overcome the challenges posed is that we

can talk, beyond autonomy and critical thinking, of

learning to learn.

Piaget’s perspective (1973) on learning focuses

on the reflexive action of each individual with the

world and exchanges between individuals. For the

author, real learning is that which generates

knowledge.

It is noteworthy that, on a permanent basis, inter-

individual exchanges touch on the basic assumptions

of pedagogical practice highlighted in this study. In

this sense, each subject is responsible for their

production and learning and may ultimately

contribute to the learning and production of others,

since the interaction can enable favorable conditions

for this process. Hence the need for pedagogical

practice to consider and contemplate actions that

allow the articulation of the subject in collective

contexts.

This study understands the articulation of the

subject in collective contexts from the social

exchanges dealt with by Piaget (1973).

3 THE CONCEPTS NETWORK

TOOL

It is noteworthy that the focus of the problem

identified by teachers who utilize collective text

practices is the amount of time required to monitor

the high volume of data generated in the writing

construction process. It was at this point, from

numerous analyses and investigations that we chose

Text Mining technology for the development of the

tool, here called the Concepts Network, in order to

meet the demands posed.

When based on statistical methods, Text Mining

(Feldman and Sanger, 2006) depends on the

frequency with which the terms appear in the texts.

To generate the Concepts Network, the first

processing step comprises the lexical analysis, where

the produced text is broken down word by word.

Following this, all extracted concepts are subjected

to statistical analysis. At this point, based on the

statistical data, a base of concepts is created, which

will then assist in building the Network. In the next

step, the system removes words that do not add

meaning to the text, such as articles, conjunctions of

the verb to be and have, as well as pronouns. This

process was based on the method used by Schenker

(2003).

The representation of information extracted from

the texts requires specific data structures such as

Vector Space Model (VSM). The VSM-based

representation is basically a list of keywords

commonly used in Information Retrieval - IR –

(Greengrass, 2001). In the process of indexing, each

document is represented by a list of keywords. When

the user submits a query to the system, it is

converted also into a list of terms. Both lists

subsequently undergo a process of comparison,

using the scaling method (Russel and Norving,

2003). Thus, the search may return documents in

order of relevance and of similarity. However, the

VSM has, in its own model, some undesirable

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

characteristics. One such characteristic relates to the

way in which words are stored, as it prevents one

from knowing the order in which they appear in the

text and their relation to the context (Greengrass,

2001). Given these considerations, an alternative

approach is highlighted that permits the organization

of words extracted from the text and the relationship

between them – graphs.

Graphs form part of a line of research known as

'Graph Theory'. One of the most famous problems

involving the use of graphs is the problem of the

bridges of Königsberg, in the XVIII century, and

formulated by Swiss mathematician Leonhard Euler

(Berry and Linoff, 1997). In short, graphs are

abstractions created to represent relationships. In

essence, the graphs are comprised of two distinct

parts, as described below:

Vertices (singular Vertex) – contain information

which generally represent the inter-related points;

Edges – represent the relationship between two

vertices, so two vertices that are in some way related

will be connected by an edge.

The process of implementing and testing the tool

in the Collective Text Editor was initiated using the

results generated by the Concepts Network. It was

then that, with the objective of meeting the demands

identified in the study, the first contact between the

results of the Network and the original texts were

delineated. The next step involved the application of

the tool.

At the start of the analysis, without the original

texts having been read, the Networks which showed

a similarity in the presentation of terms were

grouped. They were separated into two groups: one

with a higher incidence of loose terms and the other

with a higher incidence of related /connected terms.

The analysis showed that the first group is made up

of texts that need improvement, the absence of

linkage coming from the productions lacking in

logical sequence. Meanwhile, the second group, with

an incidence of terms related to each other, is made

up of texts presenting coherence, sequence and logic

in the development.

In parallel to the applications described, one in

particular (Klemann, et al, 2011) was developed in

order to ascertain the degree of precision in the

correspondence between the text and the result

presented by the Concepts Network. In this

experiment, students were challenged to construct a

production based on a text made available on the

curricular subject proposed in class. Subsequently,

the text base was mined, and from there, a

comparative analysis was done between the concepts

extracted by the Concepts Network and the concepts

from the reference text used by the students in their

own productions. In this process, it was noted that

the texts produced by students contained 61.6% of

the words highlighted by the Concepts Network.

This statistic shows that the tool was able to

emphasize a considerable number of relevant terms

from the base text.

Once in possession of the collected data,

pedagogical practice notes were created with the

objective of offering support to the teacher who uses

the tool.

4 EDUCATIONAL

APPLICATIONS BASED ON

THE RESULTS OF THE

CONCEPTS NETWORK

This section aims to establish relationships between

the results obtained with the Concepts Network tool

and the theoretical perspective chosen in this study.

4.1 Focus on the Subject Studied

The results returned by the Concepts Network

showed that, depending on the written structure of

the text, if well written or not (presence or lack of

cohesion and coherence), it is possible to identify the

theme of production even without having read the

text before.

It is worth highlighting that based on the

interpretation of data, it was noted that the result of

the Network is closely related to the structure and

content developed in the text. Thus, if the Network

does not give indications of the subject, this may be

an indication that the text produced did not focus on

the proposed subject to be developed. Faced with

this observation, care should be given in relation to

the exchanges. It is known that the Concepts

Network is the result of social changes in the

Collective Text Editor during the construction of a

text.

When there is an imbalance in this exchange, it

can also be directly related to the type of respect that

students have regarding the rules constructed in the

group. Thus, if in the construction of the text, the

exchange between participants is supported by a

type of heteronomous respect, where obedience

governs the development of the writing, the chances

that the group cannot adjust to the perspectives of

the participants regarding the clarity and objectivity

of the production increase and may compromise the

quality.

CONCEPTSNETWORK-TextMiningTechnologyalliedwithPedagogicalPracticestoQualifytheCollectivewriting

ProcessinDistanceLearning

Figure 1: Concepts Network with incidence of loose terms.

Figure 2: Concepts Network with incidence of interrelated terms.

Only cooperation leads to autonomy and the

scope for mutual respect, where individuals can

understand the result of agreements between

different points of view and in this way integrate, a

group that transcends the perspective of a single

individual (Piaget, 1973).

Freire (2003) points out that autonomy is

founded upon the experience of many decisions. In

this sense, the pedagogical practice should focus on

stimulating experiences of decision and

responsibility (Freire, 2003). So, when the Concepts

Network does not give evidence of the approached

subject, indicating that the production did not focus

on the proposed theme, it is suggested that the

interaction between the teacher and student group

adjusts to that focus.

The challenges and questions proposed, in

addition to promoting the development of students,

may also serve as a reference to the teacher who

needs to know the potential and limitations of each

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

individual.

It is further highlighted that, in a context where,

from the Concepts Network it is possible identify the

developed theme, indicating that the written

production gives focus to the proposed issue, the

teacher may use this information as a basis for

investigating students’ level of knowledge on the

topic. Based on this, the teacher can provide

materials and discussions that allow theoretical

deepening about the subject. Thus, such conditions

can leverage the knowledge and relationing of the

theme in question with other aspects, encouraging

new developments, increasing knowledge and

improving social exchanges that support the

collective writing of text.

A pedagogical practice which encourages and

creates favorable conditions for research, reading,

analysis, inter-individual discussion and the

production of new knowledge, can provide the

development of autonomy, forming a subject with

critical spirit, with initiative go in search of elements

that meet their needs (learning to learn), articulating

and confronting their perspectives with others, who

think and argue from different points of view. Thus,

in this study it is believed that, having the Concepts

Network as support for detecting if the text

production focused on the proposed theme, can

qualify the teaching practice in order to point out the

need for new strategies which meet the needs posed.

4.2 Focus on Quality of Text

In addition to identifying the theme dealt with in the

text, the Concepts Network differentiates

productions which require perfecting and those

textual productions well developed. The Networks in

which isolated terms or small groups of terms

prevailed, were drawn from texts that needed

perfecting. Figure 1 shows a Network with these

characteristics.

In this case, the studied texts showed a greater

need, in regard to clarity in development and

objectivity, as well as in the sequence and

consistency in the development of the writing. On

the other hand, Networks in which related terms

prevailed were also identified. Here, in contrast to

the previous situation, the studied texts showed

clarity, objectivity, sequence and consistency. Figure

2 shows a Network with an incidence of interrelated

terms.

Subjects build their knowledge from their

interaction with the physical and social environment

(Piaget, 1973). In this sense, reading, reflecting,

building a perspective on the selected theme and

discussing it with other subjects is a learning

exercise. It is this direction that it is believed that the

pedagogical practice which is concerned with

creating conditions to foster interactions should be

supported, buoyed in conditions of equilibrium as

predicted by Piaget (1973).

In addition to the conditions already considered,

another important factor to investigate associated

with Concepts Networks which show evidence of

the need for improvement in the text, is the presence

or absence of relationships based on self-

centeredness, coercion and cooperation.

When in the presence of self-centeredness, the

subjects fail to coordinate their points of view, once

they understand the things and the other individuals

from their actions. The movement that this writing

requires depends fundamentally on the articulation

and coordination of the propositions of the subjects

involved. Otherwise this relationship is governed by

an unbalanced situation.

In this perspective, it is understood that the

teacher must recognize when to intervene to foster

the relationship the student must make between the

object of knowledge and the level of development

that the student finds himself/herself in that moment.

This does not imply a single method of work. On the

contrary, it is understood that the pedagogical

practice should adopt different ways of working for

different necessities, always with the goal of

building the student's knowledge.

It is worth noting that this study does not intend to

reduce the teaching strategies that were

contemplated in this writing, as it is aware of the

diversity of situations and variables that a learning

process may involve.

5 FINAL CONSIDERATIONS

This study shows that the Concepts Network can

both indicate the theme developed in the text and

provide qualitative indicators of this production.

Qualitative indicators differentiate the texts

requiring improvement from those which were

developed with clarity and objectivity. All these

situations can be identified through the Network,

without the prior reading of the original text. It is

understood that this option only makes sense if it

serves to support a teaching practice committed to

the process of student learning and equally

committed to its own qualification. From this

perspective, four assumptions were taken as

minimum to account for the complex relationship

that involves the process of collective text

CONCEPTSNETWORK-TextMiningTechnologyalliedwithPedagogicalPracticestoQualifytheCollectivewriting

ProcessinDistanceLearning

production, they are: autonomy, critical thinking,

learning to learn and knowing how to articulate

oneself in collective contexts. It is noteworthy also

that this research brings, as a support perspective for

these assumptions, the reflective practice of the

teacher who must deal with diversity, with differing

needs and different demands resulting from the

process of collective construction and in this sense, a

“single only” or an “ideal model” does not make

sense.

Bearing in mind the above, the principal

contributions that result from this study are

presented in an objective form. First, there is the

significant reduction of reading time required of the

teacher to monitor students' collective text

production. As a result, there is an increase in the

length of direct interaction time between students

and teacher, which can provide significant

qualification in the teaching-learning process. In

addition, the indicators shown by the Concepts

Network can help the teacher to focus on his/her

actions, acting directly on the needs and potentiality

of students. Such conditions can extend the

possibilities of knowledge building and skills of

written production. Finally, it is emphasized that the

possible strategies of pedagogical practices may

qualify the practice of the teacher who cares and

always aims to achieve ever higher and significant

levels of excellence in what he/she does.

ACKNOWLEDGEMENTS

This study has been partially supported by CAPES,

through grant No 2737/2010.

REFERENCES

Berry, J. A., and Linoff, G., 1997. Data Mining

Technniques for Marketing, Sales and Customer

Support.: Wiley.

Feldman, R. and Sanger, J., 2006. Text Mining Handbook.

Inglaterra: Universidade de Cambridge.

Freire, P., 2003. Pedagogia da Autonomia: saberes

necessários à prática educativa. Paz e Terra, São

Paulo.

Greengras, E., 2001. Information retrieval, A survey.

Available:http://citeseer.ist.psu.edu/greengrass00infor

mation.html. Accessed: March 2011.

Klemann, M., Reategui, E., et al., 2011. Sobek: a Text

Mining Tool for Educational Applications.

International Conference on Data Mining, Las Vegas,

Estados Unidos.

Macedo, A. L., Behar, P., et al. 2009. Collective Text

Editor: a new interface focused on interaction design.

In: Arthur Tatnall; Anthony Jones. (Org.). “Education

and Technology for a better world”. 1 ed. Berlin /

Germany: Springer, v. 1, p. 331-339.

Parrat-Dayan, S., 2007. A discussão como ferramenta para

o processo de socialização e para a construção do

pensamento. In: Educação em Revista, nº 45, Belo

Horizonte. Available in: http://www.scielo.br/scielo.

php?pid=S0102-46982007000100002&script=sci_art

text&tlng=es. Accessed: março 2011.

Piaget, J., 1973. Estudos Sociológicos. Forense, Rio de

Janeiro.

Piaget, J., 1986. A linguagem e o pensamento da criança.

Martins Fontes, São Paulo.

Piaget, J., 1994. O juízo moral na criança. Summus, São

Paulo.

Piaget, J, 1998. Sobre a Pedagogia. Casa do Psicólogo,

São Paulo.

Schenker, A., 2003. Graph-Theoretic Techniques for Web

Content Mining. Tese de Doutorado em Ciência da

Computação, University of South Florida.

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation