Content uploaded by Sebastian Ventura
Author content
All content in this area was uploaded by Sebastian Ventura on Jun 17, 2022
Content may be subject to copyright.
Overview
Data mining in education
Cristobal Romero∗and Sebastian Ventura
Applying data mining (DM) in education is an emerging interdisciplinary re-
search field also known as educational data mining (EDM). It is concerned with
developing methods for exploring the unique types of data that come from ed-
ucational environments. Its goal is to better understand how students learn and
identify the settings in which they learn to improve educational outcomes and to
gain insights into and explain educational phenomena. Educational information
systems can store a huge amount of potential data from multiple sources coming
in different formats and at different granularity levels. Each particular educa-
tional problem has a specific objective with special characteristics that require a
different treatment of the mining problem. The issues mean that traditional DM
techniques cannot be applied directly to these types of data and problems. As
a consequence, the knowledge discovery process has to be adapted and some
specific DM techniques are needed. This paper introduces and reviews key mile-
stones and the current state of affairs in the field of EDM, together with specific
applications, tools, and future insights. C
2012 Wiley Periodicals, Inc.
How to cite this article:
WIREs Data Mining Knowl Discov 2013, 3: 12–27 doi: 10.1002/widm.1075
INTRODUCTION
The increase of e-learning resources, instrumental
educational software, the use of the Internet in
education, and the establishment of state databases
of student information has created large repositories
of data.1All this information provides a goldmine of
educational data that can be explored and exploited
to understand how students learn.2In fact, today, one
of the biggest challenges that educational institutions
face is the exponential growth of educational data
and the use of this data to improve the quality of
managerial decisions.3
Educational data mining (EDM) is concerned
with developing, researching, and applying comput-
erized methods to detect patterns in large collections
of educational data that would otherwise be hard or
impossible to analyze due to the enormous volume
of data within which they exist.4EDM has emerged
as a research area in recent years aimed at analyz-
ing the unique kinds of data that arise in educational
settings to resolve educational research issues (Baker
and Yacef, 2009). In fact, EDM, can be defined as
the application of data mining (DM) techniques to
this specific type of dataset that come from educa-
∗Correspondence to: cromero@uco.es
Department of Computers Science and Numerical Analysis, Uni-
versity of Cordoba, Cordoba, Spain.
DOI: 10.1002/widm.1075
tional environments to address important educational
questions.5,6
EDM analyze data generated by any type of in-
formation system supporting learning or education
(in schools, colleges, universities, and other academic
or professional learning institutions providing tradi-
tional and modern forms and methods of teaching,
as well as informal learning). These data7are not re-
stricted to interactions of individual students with an
educational system (e.g., navigation behavior, input
in quizzes and interactive exercises) but might also
include data from collaborating students (e.g., text
chat), administrative data (e.g., school, school dis-
trict, teacher), demographic data (e.g., gender, age,
school grades), student affectivity (e.g., motivation,
emotional states), and so forth. These data have typ-
ical characteristics such as multiple levels of hierar-
chy (subject, assignment, question levels), context (a
particular student in a particular class encountering a
particular question at a particular time on a particular
date), fine grained (recording of data at different reso-
lutions to facilitate different analyses, e.g., recording
data every 20 second), and longitudinal (much data
recorded over many sessions for a long period of time,
e.g., spanning semester and year-long courses).
EDM is an interdisciplinary area including
but not limited to information retrieval, recom-
mender systems, visual data analytics, domain-driven
DM, social network analysis (SNA), psychopedagogy,
12 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
FIGURE 1|Main areas related to educational data mining.
cognitive psychology, psychometrics, and so on. In
fact, EDM can be drawn as the combination of three
main areas (see Figure 1): computer science, educa-
tion, and statistics. The intersection of these three ar-
eas also forms other subareas closely related to EDM
such as computer-based education, DM and machine
learning, and learning analytics (LA).
Of all the aforementioned areas (see Figure 1),
the field most related to EDM is LA, also known
as academic analytics.8LA is focused on data-driven
decision-making and integrating the technical and the
social/pedagogical dimensions of LA.9However, al-
though EDM is generally looking for new patterns
in data and developing new algorithms and/or mod-
els, LA is applying known predictive models in in-
structional systems.10 In fact, LA can be defined as
the measurement, collection, analysis, and reporting
of data about learners and their contexts, for pur-
poses of understanding and optimizing learning and
the environments in which it occurs. Although LA
and EDM can share many attributes and have some
similar goals and interests, the next key differences
can be distinguished between both communities9:
•Techniques: In LA, the most used tech-
niques are statistics, visualization, SNA, sen-
timent analysis, influence analytics, discourse
analysis, concept analysis, and sense-making
models. In EDM, the most used techniques
are classification, clustering, Bayesian model-
ing, relationship mining and discovery with
models.
•Origins: LA has stronger origins in Seman-
tic Web, intelligent curriculum, and systemic
interventions. EDM has strong origins in edu-
cational software, student modeling, and pre-
dicting course outcomes.
•Emphasis: LA has more emphasis on the de-
scription of data and results; however, EDM
has more emphasis on the description and
comparison of the DM techniques used.
•Type of discovery: In LA, leveraging human
judgment is key; automated discovery is a tool
used to accomplish this goal. In EDM, au-
tomated discovery is key; leveraging human
judgment is a tool used to accomplish this
goal.
This paper provides an updated overview of the
current state of knowledge in EDM with the objec-
tive of introducing it to researchers, instructors and
advanced students without a strong background in
the field. The paper is organized as follows. First,
the background of EDM is described. Then the main
types of educational environments and their data are
shown. The following sections describe the main goals
and the specific knowledge discovery process in EDM.
Next, the most popular methods used in EDM are
presented. Subsequently, some examples of applica-
tions or tasks in educational environments and some
examples of specific DM tools are listed. Finally,
some future lines of research and conclusions are
outlined.
BACKGROUND
EDM has emerged as an independent research area in
recent years, starting with research in intelligent tutor-
ing systems (ITS), artificial intelligence in education
(AIED), user modeling (UM), technology-enhanced
learning (TEL), and adaptive and intelligent educa-
tional hypermedia (AIEH). Its origins lie in a series of
workshops (see Table 1) organized into related con-
ferences that began in 2000. The first workshop, re-
ferred to as ‘Educational Data Mining’, took place
in 2005 and culminated in 2008 with the establish-
ment of the annual International Conference on Edu-
cational Data Mining organized by the International
Working Group on Educational Data Mining.
The first conference EDM2008 was held in
Montreal, Canada; then EDM2009 in Cordoba,
Spain; EDM2010 in Pittsburgh, USA; EDM2011 in
Eindhoven, The Netherlands; EDM2012 in Chania,
Greece; and the next EDM2013 will be held in Mem-
phis EEUU. There are some other closely related con-
ferences (see Table 2) in which EDM is colocated
most years. All of them are older than EDM with
the exception of the LAK conference (International
Conference on Learning Analytics and Knowledge),
Volume 3, January/February 2013 13
c
2012 John Wiley & Sons, Inc.
Overview wires.wiley.com/widm
TABLE 1 Educational Data Mining (EDM) Workshops
Title Acronym Location Year
Workshop on Applying Data Mining in e-Learning EC-TEL’07-ADML Crete, Greece 2007
Workshop on Educational Data Mining AIED’07-EDM California, USA 2007
Workshop on Educational Data Mining ICALT’07-EDM Niigata, Japan 2007
Workshop on Educational Data Mining AAAI’05-EDM Boston, USA 2006
Workshop on Educational Data Mining ITS’06-EDM Jhongli, Taiwan 2006
Workshop on Educational Data Mining AAAI’05-EDM Pittsburgh, USA 2005
Workshop on Usage Analysis in Learning Systems AIED’05-W1 Amsterdam, the Netherlands 2005
Workshop on Analyzing Student–Tutor Interaction Logs to
Improve Educational
ITS’04-W2 Maceio, Brazil 2004
Workshop on Applying Machine Learning to ITS
Design/Construction
ITS’00-W3 Montreal, Canada 2000
TABLE 2 Related Conferences about Educational Data Mining
Title Acronym Type Year
International Conference on Educational Data Mining EDM Annual 2008
International Conference on Learning Analytics and Knowledge LAK Annual 2011
International Conference on Artificial Intelligence in Education AIED Biannual 1982
International Conference on Intelligent Tutoring Systems ITS Biannual 1988
International Conference on User Modeling, Adaptation, and Personalization UMAP Annual 2009
which is younger. The first LAK conference, was in
Banff, Canada, in 2011 and the second in Vancouver,
Canada, in 2012.
Currently, only two books on EDM have
been published. The first, entitled Data Mining in
E-Learning,11 has 17 chapters oriented to Web-
based educational environments. The second, enti-
tled Handbook of Educational Data Mining (Romero
et al., 2010), and has 36 chapters oriented to different
types of educational settings.
There are several surveys in journals and chap-
ters in books about EDM. The first and most popular
review of EDM research was presented in a journal,5
and was followed by a more theoretical paper (Baker
and Yacef, 2009) and a more complete review.6A first
and wide-ranging book chapter review was oriented
to the application of DM in e-learning,12 a second
and shorter book chapter was more oriented to ITS13
and a third book chapter was the most generic but
the shortest.7Finally, a recent report was published
by the US Office of Educational Technology about
how to enhance teaching and learning through EDM
and LA.10
There are a wide range of international and pres-
tigious journals in which a large number of EDM pa-
pers have been published (see Table 3). Of all of them,
the most specific is the Journal of Educational Data
Mining (http://www.educationaldatamining.org/
JEDM/), which was launched in 2009 as an online
and free journal. On the other hand, a selection
of the most cited papers in EDM area is shown in
Table 4.
There are an increasing number of important
authors in EDM area as well as the authors of this
paper. Ryan Baker from Worcester Polytechnic In-
stitute, USA, that is, the president of the EDM so-
ciety. Kalina Yacef from the University of Sydney,
Australia, that is, the editor of JEDM journal and
member of the steering committee of EDM soci-
ety together with Tiffany Barnes from University of
North Calorina, USA; Joseph E. Beck from Worcester
Polytechnic Institute, USA; Michel Desmarais from
Ecole Polytechnic de Montreal, Canada; Neil Hef-
fernan from Worcester Polytechnic Institute, USA;
Agathe Merceron from Beuth University of Applied
Sciences, Germany; and Mykola Pechenizkiy from
Eindhoven University of Technology, the Nether-
lands. Other relevant authors are Osmar Zai¨
aine
from Alberta University, Canada; John Stamper from
Carnegie Mellon University, USA; Judy Kay from The
14 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
TABLE 3 Some examples of Educational Data Mining Related Journals
Journal Title Acronym Editorial Impact Factor 2011
Journal of Educational Data Mining JEDM EDM Society –
Journal of Artificial Intelligence in Education JAIED AIED Society –
Journal of the Learning Sciences JLS Taylor&Francis 2.0001
Computer and Education CAE Elsevier 2.6212
IEEE Transactions on Learning Technologies TLT IEEE –
IEEE Transactions on Knowledge and Data Engineering KDE IEEE 1.6572
ACM Special Interest Group on Knowledge Discovery
and Data Mining, Explorations
SIGKDD Explorations ACM –
User Modeling and User-Adapted Interaction UMUAI Springer 1.4002
Internet and Higher Education INTHIG Elsevier 1.0151
Decision Support Systems DCS Elsevier 1.6872
Expert Systems with Applications ESWA Elsevier 2.2032
Knowledge-Based Systems KBS Elsevier 2.4222
1JCR Social Science Edition.
2JCR Science Edition.
TABLE 4 Top 10 Most Cited Papers about Educational Data Mining until August 2012
Number of Number of
Paper Title Reference citations1citations2
Educational data mining: a survey from 1995 to 2005 5 296 158
Data mining in course management systems: Moodle case study and tutorial 14 191 83
Web usage mining for a better Web-based learning environment 15 183 –
Off-task behavior in the cognitive tutor classroom: when students game the system 16 177 25
Building a recommender agent for e-learning systems 17 168 –
Detecting student misuse of intelligent tutoring systems 18 156 20
The ecological approach to the design of e-learning environments: purpose-based
capture and use of information about learners
19 136 –
Student modeling and machine learning 20 127 –
Towards evaluating learners’ behavior in a Web-based distance learning environment 21 117 –
Smart recommendation for an evolving e-learning system: architecture and experiment 22 98 –
1Google Schoolar.
2SciVerse Scopus.
University of Sydney, Australia; Kenneth Koedinger
and Jack Mostow from Carnegie Mellon University,
USA; Rafi Nachmias from Tel Aviv University, Is-
rael; Gord McCalla from University of Saskatchewan,
Canada; Arthur Graesser from The University of
Hemphis, USA; and so forth.
There are several international societies re-
lated to EDM. The most important are the Interna-
tional Educational Data Mining Society (http://www
.educationaldatamining.org) which was founded by
the International Working Group on Educational
Data Mining in 2011; the Society for Learning Analyt-
ics Research (SoLAR) (http://www.solaresearch.org/)
which was created in 2011; and the IEEE Task
Force of Educational Data Mining (EDM-TF) (http://
datamining.it.uts.edu.au/edd/) which was created in
2012.
Finally, to demonstrate the current increasing
interest in EDM, Figure 2 shows the number of
references or results that return a freely accessible
Web search engine such as Google Schoolar and a
subscription-based tool such as SciVerse Scopus when
searching the exact term ‘Educational Data Mining’
in each year from 2004 to 2011. As can be seen, both
numbers grow in an exponential way, showing the
high interest in this topic, and in the last two years
the number of cites in SciVerse Scopus is higher than
in Google Schoolar.
Volume 3, January/February 2013 15
c
2012 John Wiley & Sons, Inc.
Overview wires.wiley.com/widm
FIGURE 2|Number of educational data mining references in Google Schoolar and cites in SciVerse Scopus by year.
BOX 1: EDM COMPETITIONS
There were two specific international EDM competitions
with the same objective of predicting whether a student
will answer the next test question correctly. The first
competition was the KDD Cup 2010 (https://pslcdatashop
.web.cmu.edu/KDDCup/) on the Educational Data Mining
Challenge with 5096 participants. The data comes from
10,000 students of Carnegie Learning Inc.’s Cognitive Tu-
tors. And the second competition was the Kaggle Com-
petition (http://www.kaggle.com/c/WhatDoYouKnow) with
252 teams and a prize pool of $5000. The data in this com-
petition comes from students studying for three tests: the
GMAT, SAT, and ACT.
TYPES OF EDUCATIONAL
ENVIRONMENTS
Nowadays, there is a wide variety of educational en-
vironments and information systems both in tradi-
tional education and computer-based education (see
Figure 3). Each one of them provides different data
sources that have to be pre-processed in different ways
depending on both the nature of available data and
the specific problems and tasks to be resolved by DM
techniques.5
Traditional Education
Traditional education or back-to-basics refers to long-
established customs found in schools that society has
traditionally deemed to be appropriate. These envi-
ronments are the most widely used educational sys-
tem, based mainly on face-to-face contact between ed-
ucators and students organized through lectures, class
discussion, small groups, individual seat work, and so
forth. These systems gather information on student
attendance, marks, curriculum goals, and individu-
alized plan data. Also, educational institutions store
many diverse and varied sources of information23: ad-
ministrative data in traditional databases (with a stu-
dent’s information, the educator’s information, class
and schedule information, etc.), online information
(online Web pages and course content pages), and
so forth. In conventional classrooms, educators nor-
mally attempt to enhance instruction by monitor-
ing students’ learning processes and analyzing their
performance on paper and through observation.24
16 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
FIGURE 3|Types of traditional and computer-based educational environments and systems.
TABLE 5 Types of Traditional Educational Environments
System Description
Infant/preschool education This provides learning to children before statutory and obligatory education, usually between the
ages of zero and three or five.
Primary/elementary education This consists of the first 5–7 years of formal, structured education.
Secondary education This comprises the formal education that occurs during adolescence.
Higher/tertiary education This is the noncompulsory educational level that follows the completion of a school providing a
secondary education.
Alternative/special education This includes not only forms of education designed for students with special needs, but also forms
of education designed for a general audience and employing alternative educational methods.
Some examples of traditional education systems are
described in Table 5. Finally, it is important to
note that all these traditional systems can also use
computer-based educational systems as a complemen-
tary tool to face-to-face sessions.
Computer-Based Educational Systems
Computer-based education (CBE) means using com-
puters in education to provide direction, to instruct
or to manage instructions given to the student. CBE
systems were originally stand-alone educational ap-
plications that ran on a local computer without using
artificial intelligence techniques for student modeling,
adaptation, personalization, and so forth. On the one
hand, the global use of Internet has led to today’s
plethora of new Web-based educational systems such
as e-learning systems, e-training systems, online in-
struction systems, and so forth. On the other hand, the
increasing use of artificial intelligence techniques has
induced the emergence of new intelligent and adap-
tive educational systems. Some of the main types of
computer-based educational systems used currently
are (see Table 6 for a description) learning and man-
agement systems (LMS),14 ITS,2adaptive and intel-
ligent hypermedia systems (AIHS),25 test and quiz
systems,26 and other types of CBE systems.
GOALS
Data mining has already been successfully applied to
other areas or domains such as business, bioinfor-
matics, genetics, medicine, and so forth. Although the
discovery methods used in all these areas can be seen
similar, the objectives are different.6For instance, in
comparing the use of data mining in e-commerce ver-
sus EDM. The main objective of data mining in e-
commerce is to increase profit. Profit is a tangible
goal that can be measured in terms of sums of money,
Volume 3, January/February 2013 17
c
2012 John Wiley & Sons, Inc.
Overview wires.wiley.com/widm
TABLE 6 Types of Computer-Based Educational Systems
System Description
Learning management
systems
Suites of software that provide course-delivery functions: administration, documentation, tracking, and
reporting of training programs—classroom and online events, e-learning programs, and training
content. They also offer a wide variety of channels and workspaces to facilitate information sharing
and communication among all the participants in a course. They record any student activities involved,
such as reading, writing, taking tests, performing tasks in real, and commenting on events with peers.
Intelligent tutoring
systems (ITS)
ITS provide direct customized instruction or feedback to students by modeling student behavior and
changing its mode of interaction with each student based on its individual model. Normally, it consists
of a domain model, student model, and pedagogical model. ITS record all student–tutor interaction
(mouse clicks, typing, and speech).
Adaptive and intelligent
hypermedia (AIH)
systems
These attempt to be more adaptive by building a model of the goals, preferences, and knowledge of each
individual student and using this model throughout interaction with the student to adapt to the needs
of that student. The data recorded by AIHs are similar to ITS data.
Test and quiz systems The main goal of these systems is to measure the students’ level of knowledge with respect to one or
more concepts or subjects by using a series of questions/items and other prompts for the purpose of
gathering information from respondents. They store a great deal of information about students’
answers, calculated scores, and statistics.
Other types Learning object repositories, concept maps, social networks, wikis, forums, educational games, virtual
reality/3D, ubiquitous computing, and so forth.
TABLE 7 Example of Users/Stakeholders and Objectives
User/Stakeholders Examples of objectives
Learners To support a learner’s reflections on the situation, to provide adaptive feedback or recommendations to
learners, to respond to student needs, to improve learning performance, and so on
Educators To understand their students’ learning processes and reflect on their own teaching methods, to improve
teaching performance, to understand social, cognitive and behavioral aspects, and so on
Researchers To develop and compare data mining techniques to be able to recommend the most useful one for each
specific educational task or problem, to evaluate learning effectiveness when using different settings
and methods, and so on.
Administrators To evaluate the best way to organize institutional resources (human and material) and their educational
offer, and so on
and which leads to clear secondary measures such
as the number of customers and customer loyalty.
As the main objective of data mining in education is
largely to improve learning, measurements are more
difficult to obtain, and must be estimated through
proxies such as improved performance. So, in general
it enables data-driven decision-making for improving
current educational practice and learning materials.
However, there are many more specific objectives in
EDM depending on the viewpoint of the final user and
the problem to resolve. Some examples of particular
problems are27
•How to (re)organize classes, or assessment,
or the placement of materials based on usage
and performance data.
•How to identify those who would benefit
from feedback, study advice, or other help
provided.
•How to decide which kind of help, feedback,
or advice would be most effective.
•How to help learners in finding and searching
useful material, individually, or in collabora-
tion with peers.
Although an initial consideration seems to in-
volve only two main groups of potential users/
stakeholders—the learners and the instructors—there
are actually more groups involved with many more
objectives, as can be seen in Table 7.
The number of possible problems or objectives
for each type of stakeholder is huge. For example,
18 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
TABLE 8 Current Topics of Interest of Educational Data Mining Research Community
Topics of Interest Description
Generic frameworks and
methods
To develop tools, frameworks, methods, algorithms, approaches, and so forth, specifically
oriented to educational data mining research.
Mining educational data Mining assessment data, mining browsing or interaction data, mining the results of educational
research (e.g., A/B tests), and so forth.
Educational process mining To extract process-related knowledge from event logs recorded by educational systems.
Data-driven adaptation and
personalization
To apply data mining methods and techniques for improving adaptation and personalization in
educational environments and systems.
Improving educational
software
Many large educational data sets are generated by computer software. Can we use our
discoveries to improve the software’s effectiveness?
Evaluating teaching
interventions
Student learning data provides a powerful mechanism for determining which teaching actions
are successful. How can we best use such data?
Emotion, affect, and choice The student’s level of interest is critical. Can we detect when students are bored and
uninterested? What other affective states or student choices should we track?
Integrating data mining and
pedagogical theory
Data mining typically involves searching a large volume of models. Can we use existing
educational and psychological knowledge to better focus our research?
Improving teacher support What types of assessment information would help teachers? What types of instructional
suggestions are both feasible to generate and would be welcomed by teachers?
Replication studies To apply a previously used technique to a new domain, or to reanalyze an existing data set with
a new technique.
Best practices Best practices for adaptation of data mining, information retrieval, recommender system, opinion
mining, and question answering techniques to educational context.
FIGURE 4|Educational knowledge discovery and data mining process.
from the point of view of EDM researchers, there
is a wide range of current topics of interest (see
Table 8).
EDUCATIONAL KNOWLEDGE
DISCOVERY PROCESS
The process of applying data mining to educational
systems can be interpreted from different points of
view (Romero et al., 2010).
On the one hand, from an educational and an
experimental viewpoint, it can be seen as an iterative
cycle of hypothesis formation, testing, and refinement
(see Figure 4). In this process, the goal is not just to
turn data into knowledge, but also to filter mined
knowledge for decision-making about how to mod-
ify the educational environment to improve student’s
learning. This is a type of formative evaluation of an
educational program while it is still in development,
and with the purpose of continually improving the
program. Analyzing how students use the system is
one way to evaluate instructional design in a forma-
tive manner and may help educational designers to
improve instructional materials. For example, EDM
techniques discover models/patterns that can be used
to assist educational designers to establish a pedagog-
ical basis for decisions when designing or modifying
an environment’s pedagogical approach.
On the other hand, from a DM viewpoint, it
can be seen very similar to the general knowledge dis-
covery and data mining (KDD) process (see Figure 4)
although there are important differences or specific
characteristics in each step as is described in the fol-
lowing subsections.
Educational Environment
Depending on the type of the educational environ-
ment (traditional classroom education, computer-
based or Web-based education) and an information
Volume 3, January/February 2013 19
c
2012 John Wiley & Sons, Inc.
Overview wires.wiley.com/widm
FIGURE 5|Different levels of granularity and their relationship to
the amount of data.
system that supports it (a learning management, in-
telligent tutoring or adaptive hypermedia system) dif-
ferent kinds of data can be collected to resolve differ-
ent educational problems.4All these data may come
from different sources including administrative data,
field observations, motivational questionnaires, mea-
surements collected from controlled experiments, fi-
nal marks, and so on. Gathering and integrating this
raw data for mining are nontrivial tasks on their own
and thus a preprocessing step is necessary.
Preprocessing
In educational contexts, it is natural for data prepro-
cessing to be a very important and complicated task,
and sometimes the data preprocessing itself takes up
more than half of the total time spent solving the
data mining problem.10 First, the educational data
available (raw, original, or primary data) to solve a
problem is not in the appropriate form (or abstrac-
tion). And second, given the heterogeneous and hi-
erarchical nature of educational data, determining
data structures and formats that represent an event
under consideration become key and the best data
structure will also depend on the type of problem
to be solved. So, it is necessary to convert the data
to an appropriate form (modified data) for solving
a specific educational problem. This includes choos-
ing what data to collect, focusing on the questions
to be answered, and making sure the data align with
the questions. On the other hand, educational envi-
ronments can store a huge amount of potential data
from multiple sources with different formats and with
different granularity levels (from coarse to fine grain)
or multiple levels of meaningful hierarchy (keystroke
level, answer level, session level, student level, class-
room level, and school level) that provide more or less
data (see Figure 5). So, it can be necessary to carry
out data integration at the appropriate granularity
level. Normally, also available are a huge number of
variables/attributes with information about each stu-
dent, which can be reduced into a summary table for
better analysis. Continuous attributes are normally
transformed/discretized into categorical attributes to
improve their comprehensibility. Issues of time, se-
quence, and context also play important roles in the
study of educational data. Time is important to cap-
ture data such as length of practice sessions or time
to learn. Sequences represent how concepts build on
one another and how practice and tutoring should
be ordered. Context is important for explaining re-
sults and knowing where a model may or may not
work. Finally, it is important to maintain and protect
the confidentiality of student information when inte-
grating all collected data by deleting some personal
information (not useful for mining purposes) such as
name, e-mail, telephone number, and so on, and thus
anonymizing data by using, for example, a numerical
sequence for identifying students.
Data Mining
The majority of traditional data mining techniques
including but not limited to classification, clustering,
and association analysis techniques have been already
applied successfully in the educational domain.13
Nevertheless, educational systems have special char-
acteristics that require a different treatment of the
mining problem. For example, methods for hierarchi-
cal data mining and longitudinal data modeling have
to be used in EDM. As a consequence, some specific
data mining techniques are needed to address learn-
ing and other data about learners. However, EDM is
still an emerging research area, and we can foresee
that its further development will result in a better un-
derstanding of the challenges specific to this field and
will help researchers involved in EDM to see which
techniques can be adopted and what new customized
techniques have to be developed. On the other hand,
there are some data mining methods that are more ap-
propriate for solving some of the types of educational
problems to resolve, as described in Methods.
Interpretation of Results
This final step is very important to apply the knowl-
edge acquired to making decision about how to im-
prove the educational environment or system.28 So
the models obtained by the DM algorithms have to
be comprehensible and useful for the decision-making
process. For example, white-box DM models such
as decision trees are preferable to black-box models
such as neural networks as they are more accurate
but less comprehensible. Visualization techniques are
also very useful for showing results in a way that is
easier to interpret. For example, it is better to show
only a subset of association rules in graphic format
instead of showing all the rules discovered (normally
hundreds or thousands) in a traditional text format.
20 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
Finally, recommender systems can be the best way
to display results, information, explanations, recom-
mendations and comments to a nonexpert user in DM
such as instructors. Thus instead of showing the ob-
tained DM model, a list of suggestions or conclusions
about the results and how to apply them are shown
to the users.
METHODS
There are a number of popular methods within
EDM.5,7,13,29 Some of them are widely acknowl-
edged to be universal across types of data mining,
such as prediction, clustering, outlier detecting, rela-
tionship mining, SNA, process mining, and text min-
ing. And others have particular prominence within
EDM, such as the distillation of data for human judg-
ment, discovery with models, knowledge tracing (KT)
and nonnegative matrix factorization.
Prediction
The goal of prediction is to infer a target attribute
or single aspect of the data (predicted variable) from
some combination of other aspects of the data (pre-
dictor variables). Types of predictions methods are
classification (when the predicted variable is a cate-
gorical value), regression (when the predicted vari-
able is a continuous value), or density estimation
(when the predicted value is a probability density
function). In EDM, prediction has been used for fore-
casting student performance30 and for detecting stu-
dent behaviors.31
Clustering
The goal of clustering is to identify groups of in-
stances that are similar in some respect. Typically,
some kind of distance measure is used to decide how
similar instances are. Once a set of clusters has been
determined, new instances can be classified by deter-
mining the closest cluster. In EDM, clustering can be
used for grouping similar course materials or group-
ing students based on their learning and interaction
patterns.32
Outlier Detection
The goal of outlier detection is to discover data points
that are significantly different than the rest of data.
An outlier is a different observation (or measurement)
that is usually larger or smaller than the other values
in data. In EDM, outlier detection can be used to
detect students with learning difficulties, deviations
in the learner’s or educator’s actions or behaviors,
and for detecting irregular learning processes.33
Relationship Mining
The goal of relationship mining is to identify rela-
tionships between variables and normally to encode
them in rules for later use. There are different types
of relationship in mining techniques such as asso-
ciation rule mining (any relationship between vari-
ables), sequential pattern mining (temporal associa-
tions between variables), correlation mining (linear
correlations between variables), and causal data min-
ing (causal relationship between variables). In EDM,
relationship mining has been used to identify rela-
tionships in learners’ behavior patterns and diagnos-
ing students’ learning difficulties or mistakes that fre-
quently occur together.34
Social Network Analysis
The goal of SNA is to understand and measure the
relationships between entities in networked informa-
tion. SNA views social relationships in terms of net-
work theory consisting of nodes (representing indi-
vidual actors within the network) and connections or
links (which represent relationships between the in-
dividuals, such as friendship, kinship, organizational
position, sexual relationships, etc.). In EDM, SNA
can be used for mining to interpret and analyze the
structure and relations in collaborative tasks and in-
teractions with communication tools.35
Process Mining
The goal of process mining is to extract process-
related knowledge from event logs recorded by an
information system to have a clear visual representa-
tion of the whole process. It consists of three subfields:
conformance checking, model discovery, and model
extension. In EDM, process mining can be used for
reflecting students behavior in terms of their examina-
tion traces consisting of a sequence of course, grade,
and timestamp triplets for each student.36
Text Mining
The goal of text mining, also referred to as text data
mining or text analytics, is to derive high-quality in-
formation from text. Typical text mining tasks include
text categorization, text clustering, concept/entity ex-
traction, production of granular taxonomies, senti-
ment analysis, document summarization, and entity
relation modeling. In EDM, text mining has been used
to analyze the content of discussion boards, forums,
chats, Web pages, documents, and so forth.37
Distillation of Data for Human Judgment
The goal is to represent data in intelligible ways us-
ing summarization, visualization and interactive in-
terfaces to highlight useful information and support
Volume 3, January/February 2013 21
c
2012 John Wiley & Sons, Inc.
Overview wires.wiley.com/widm
FIGURE 6|Example of nonnegative matrix factorization and
Q
-matrix interpretation.
decision-making. On the one hand, it is relatively easy
to obtain descriptive statistics from educational data
to obtain global data characteristics and summaries
and reports on learner behavior. On the other hand,
information visualization and graphic techniques help
to see, explore, and understand large amounts of ed-
ucational data at once. In EDM, it also known as dis-
tillation for human judgment13 and it has been used
for helping educators to visualize and analyze the stu-
dents’ course activities and usage information.38
Discovery with Models
The goal of discovering with models is to use a pre-
viously validated model of a phenomenon (using pre-
diction, clustering, or manual knowledge engineering)
as a component in another analysis such as prediction
or relationship mining.29 It is particularly prominent
in EDM and it supports the identification of relation-
ships between student behaviors and students’ char-
acteristics or contextual variables, the analysis of re-
search questions across a wide variety of contexts,
and the integration of psychometric modeling frame-
works into machine-learning models.10
Knowledge Tracing
KT is a popular method for estimating student mas-
tery of skills that has been used in effective cognitive
tutor systems.39 It uses both a cognitive model that
maps a problem-solving item to the skills required,
and logs of students’ correct and incorrect answers
as evidence of their knowledge on a particular skill.
KT tracks student knowledge over time and it is pa-
rameterized by four variables. There is an equivalent
formulation of KT as a Bayesian network.
Nonnegative Matrix Factorization
Nonnegative matrix factorization (NMF) is a tech-
nique that allows a straightforward interpretation in
terms of a Q-Matrix, also termed a transfer model.40
There are many NMF algorithms and they can yield
different solutions. NMF consists of a matrix of posi-
tive numbers, as the product of two smaller matrices.
For example, in the context of education, a matrix M
that represents the observed examinee’s test outcome
data (see Figure 6) that can be decomposed into two
matrices: Qthat represents the Q-matrix of items and
Sthat represents each student’s mastery of skills.
APPLICATIONS
There are many examples of applications or tasks
in educational environments that can be resolved
through DM.3,4Between all of them, predicting stu-
dents’ performance is the oldest and most popular
application of DM in education. However, in recent
years, EDM has been applied to address a large num-
ber of new and different problems (see Table 9).
SPECIFIC DATA MINING TOOLS
Nowadays, there are a lot of general free and com-
mercial DM tools and frameworks49 thatcanbeused
for mining datasets from any domain or research area.
However, all these tools are not specifically designed
for pedagogical/educational purposes and problems.
So they are cumbersome for an educator to use be-
cause they are designed more for power and flexibility
than for simplicity. However, an increasing number
of mining tools have been developed that are specifi-
cally oriented to solve different educational problems
(see Table 10).
CONCLUSIONS AND FUTURE
INSIGHTS
EDM brings together an interdisciplinary community
of computer scientists, learning scientists, psycho-
metricians, and researchers from other fields. EDM
applies techniques coming from statistics, machine
learning, and data mining to analyze data collected
during teaching and learning, tests learning theories,
and informs decision-making in educational practice.
The field of EDM has grown substantially in re-
cent years with two related annual conferences, two
books, several surveys in books and journals, two
22 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
TABLE 9 Some Examples of Educational Data Mining Tasks or Applications
Task/Application Description/Goal Reference
Predicting student
performance
To estimate the unknown value of a student’s performance, knowledge, score or mark 26
Scientific inquiry To develop and test scientific theories on technology-enhanced learning, to formulate new
scientific hypotheses, and so on
9
Providing feedback for
supporting instructors
To provide feedback to support educators in decision-making about how to improve students’
learning and enable them to take appropriate proactive and/or remedial action
26
Personalizing to
students
To adapt automatically learning, navigation, content, presentation, and so forth, to each
particular students
41
Recommending to
students
To make recommendations to students with respect to their activities or tasks, links to visits,
problems or courses to be done, and so forth.
42
Creating alerts for
stakeholders
To monitor students’ learning progress for detecting in real time undesirable student behaviors
such as low motivation, playing games, misuse, cheating, dropping out, and so forth
43
User/Student modeling To develop and tune cognitive models of human students that represent their skills and
declarative knowledge
44
Domain modeling To describe the domain of instruction in terms of concepts, skills, learning items and their
interrelationships
45
Grouping/Profiling
students
To create groups of students according to their customized features, personal characteristics,
personal learning data, and so forth
46
Constructing
courseware
To help instructors and developers to carry out the construction/development process of
courseware and learning content automatically
28
Planning and scheduling To plan future courses, student course scheduling, planning resource allocation, admission
and counseling processes, developing curriculum, and so forth
47
Parameter estimation To infer parameters of probabilistic models from given data to predict the probability of
events of interest
48
TABLE 10 Examples of Educational Data Mining Tools
Tool Goal Reference
EPRules To discover prediction rules to provide feedback for courseware authors 50
GISMO To visualize what is happening in distance learning classes 38
TADA-ED To help teachers to identify relevant patterns in students’ online exercises 51
O3R To retrieve and interpret sequential navigation patterns 52
Synergo/ColAT To analyze and produce interpretative views of learning activities 53
LISTEN Mining tool To explore large student–tutor interaction logs 54
MINEL To analyze navigational behavior and the performance of the learner 55
LOCO-Analyst To provide teachers with feedback on the learning process 56
Measuring tool To measure the motivation of online learners 57
DataShop To store and analyze click-stream data, fine-grained longitudinal data generated by
educational systems
1
Decisional tool To discover factors contributing to students’ success and failure rates 58
CIECoF To make recommendations to courseware authors about how to improve courses 28
SAMOS To browse student activity using overview spreadsheets 59
PDinamet To support teachers in collaborative student modeling 60
AHA! Mining Tool To recommend the best links for a student to visit next 61
EDM Visualization Tool To visualize the process in which students solve procedural problems in logic 62
Meerkat-ED To analyze participation of students in discussion forums using social network analysis
techniques
35
MMT tool To facilitate the execution of all the steps in the data mining process of Moodle data for
newcomers
63
SNAPP To visualize the evolution of participant relationships within discussions forums 64
AAT To access and analyze students’ behavior data in learning systems 65
DRAL To discover relevant e-activities for learners 66
E-learning Web Miner To discover student’s behavior profiles and models about how they work in virtual courses 67
Volume 3, January/February 2013 23
c
2012 John Wiley & Sons, Inc.
Overview wires.wiley.com/widm
competitions (see Box 1), an increasing number of
specific tools, and so forth. Time will tell if this evolu-
tion continues or not, and if other communities such
as KDD perceive it as yet another application domain
of data mining or really as a new subfield. Currently,
there is a large amount of work to be done in the
EDM community in order for it to be considered as
a mature area. EDM has to be much more widely
used and applied, not only by researchers but also
by teachers and institutions. Although EDM has been
used in some courses and institutions with success,
it is necessary to move from the lab to the general
market, and to achieve this objective it is necessary to
carry out the next stages of future work:
On the one hand, it is essential that EDM tools
are open source or freely available to download in or-
der for them to be used by a much wider and broader
population. In fact, most of the current specific EDM
tools (see Table 10) are not available for download.
EDM tools must be included and integrated into their
own computer-based educational systems alongside
another tools such as course designer tools, test gener-
ator tools, report tools, and so forth. EDM tools must
also be easier for educators to use. Usually, they are
required to select the specific DM method/algorithm
they want to apply/use. And these DM algorithms
usually require parameters and they have to pro-
vide appropriate values in advance to obtain good
results/models. So, the educators must possess a cer-
tain amount of expertise to find the right settings. A
solution to this problem is the use of decision sup-
port systems, wizard tools, recommendation engines
and free-parameter DM algorithms to automate and
facilitate all the EDM processes for instructors.
On the other hand, educators and institutions
should develop a data-driven culture of using data
for making instructional decisions and improving in-
struction. Results from EDM research are typically
achieved in the narrow context of specific research
projects and educational settings. However, it is nec-
essary to obtain more general results, for instance,
whether the same student model parameters also can
be used with other student populations, or whether
a predictive model is still reliable when used in a dif-
ferent context. There is therefore an increasing need
for replication studies to test for broader generaliza-
tions. As a practical consequence of this need, EDM
researchers have become increasingly more interested
in open data repositories and standard data formats
to promote the exchange of data and models.
ACKNOWLEDGMENTS
This research is supported by projects of the Regional Government of Andalucia and the Ministry
of Science and Technology, P08-TIC-3720 and TIN-2011-22408, respectively, and FEDER
funds.
REFERENCES
1. Koedinger K, Cunningham K, Skogsholm A, Leber
B. An open repository and analysis tools for fine-
grained, longitudinal learner data. In: First Interna-
tional Conference on Educational Data Mining. Mon-
treal, Canada; 2008, 157–166.
2. Mostow J, Beck J. Some useful tactics to modify, map
and mine data from intelligent tutors. J Nat Lang Eng
2006, 12:195–208.
3. Bala M, Ojha DB. Study of applications of data mining
techniques in education. International J Res Sci Tech-
nol 2012, 1: 1–10.
4. Romero C, Ventura S, Pechenizky M, Baker R. Hand-
book of Educational Data Mining. Data Mining and
Knowledge Discovery Series. Boca Raton, FL: Chap-
man and Hall/CRC Press; 2010.
5. Romero C, Ventura S. Educational data mining: a sur-
vey from 1995 to 2005. J Expert Syst Appl 2007,
1:135–146.
6. Romero C, Ventura S. Educational data mining: a re-
view of the state-of-the-art. IEEE Trans Syst Man Cy-
bern C: Appl Rev 2010, 40:601–618.
7. Scheuer O, McLaren BM. Educational data mining.
In: The Encyclopedia of the Sciences of Learning.New
York, NY: Springer; 2011.
8. Baepler P, Murdoch CJ. Academic analytics and data
mining in higher education. Int J Schoolarship Teach
Learn 2010, 4:1–9.
9. Siemens G, Baker RSJd. Learning analytics and educa-
tional data mining: towards communication and col-
laboration. In: Proceedings of the 2nd International
24 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
Conference on Learning Analytics and Knowledge.
Vancouver, British Columbia, Canada; 2012, 1–3.
10. Bienkowski M, Feng M, Means B. Enhancing teach-
ing and learning through educational data mining and
learning analytics: an issue brief. Washington, D.C.:
Office of Educational Technology, U.S. Department of
Education; 2012, 1–57.
11. Romero C, Ventura S. Data Mining in E-learning.
Southampton, UK: Wit-Press; 2006.
12. Castro F, Vellido A, Nebot A, Mugica F. Applying data
mining techniques to e-learning problems. In: Evolu-
tion of Teaching and Learning Paradigms in Intelligent
Environment. Studies in Computational Intelligence.
Vol. 62. Berlin, Germany: Springer-Verlag; 2007, 183–
221.
13. Baker RSJd. Data mining for education. In McGaw B,
Peterson P, Baker E, eds. International Encyclopedia of
Education. 3rd ed. Vol. 7. Oxford, UK: Elsevier; 2010,
112–118.
14. Romero C, Ventura S, Salcines E. Data mining in
course management systems: Moodle case study and
tutorial. Comput Edu 2008, 51:368–384.
15. Za¨
ıane O. Web usage mining for a better web-based
learning environment. In: Proceedings of Conference
on Advanced Technology for Education. Madison, WI;
2001, 60–64.
16. Baker RS, Corbett AT, Koedinger KR, Wagner AZ.
Off-task behavior in the cognitive tutor classroom:
when students game the system. In: Proceedings of the
SIGCHI Conference on Human Factors in Computing
Systems. Vienna, VA; 2004, 383–390.
17. Za¨
ıane O. Building a recommender agent for e-learning
systems. In: Proceedings of the International Con-
ference on Computers in Education. Auckland, New
Zealand; 2002, 55–59.
18. Baker RS, Corbett AT, Koedinger KR. Detecting stu-
dent misuse of intelligent tutoring systems. In: Proceed-
ings of the 7th International Conference on Intelligent
Tutoring Systems. Macei `
o, Alagoas, Brazil; 2004, 531–
540.
19. McCalla G. The ecological approach to the design of e-
learning environments: purpose-based capture and use
of information about learners. J Interact Media Edu
2004, 7:1–23.
20. Sison R, Shimura M. Student modeling and machine
learning. IntJArtifIntellEdu1998, 9:128–158.
21. Za¨
ıane O. Towards evaluating learners’ behaviour
in a web-based distance learning environment. In:
Proceedings of the IEEE International Conference
on Advanced Learning Technologies. 2001, 357–
360.
22. Tang T, Mcalla G. Smart recommendation for an
evolving e-learning system: architecture and experi-
ment. Int J E-Learn 2005, 105–129.
23. Ma Y, Liu B, Wong C, Yu P, Lee S. Targeting the right
students using data mining. In: KDD ‘00: Proceed-
ings of the Sixth ACM SIGKDD International Confer-
ence on Knowledge Discovery and Data Mining. 2000,
457–464.
24. Marquez-Vera C, Cano A, Romero C, Ventura S. Pre-
dicting student failure at school using genetic pro-
gramming and different data mining approaches with
high dimensional and imbalanced data. Appl Intell.In
Press.
25. Merceron A, Yacef K. Mining student data captured
from a web-based tutoring tool: initial exploration and
results. J Interact Learn Res 2004, 15:319–346.
26. Romero C, Zafra A, Luna JM, Ventura S. Associa-
tion rule mining using genetic programming to provide
feedback to instructors from multiple-choice quiz data.
Expert Syst J Knowl Eng. In Press.
27. Calders To, Pechenizkiy M. Introduction to the special
section on educational data mining. ACM SIGKDD
Explor 2011, 13:3–6.
28. Garcia E, Romero C, Ventura S, Castro C. Collabora-
tive data mining tool for education. In: International
Conference on Educational Data Mining. Cordoba,
Spain; 2009, 299–306.
29. Baker RSJd, Yacef K. The state of educational data
mining in 2009: a review and future visions. JEdu
Data Min 2009, 3–17.
30. Romero C, Espejo P, Zafra A, Romero J, Ventura S.
Web usage mining for predicting marks of students that
use Moodle courses. Comput Appl Eng Edu J. In Press.
31. Baker RSJd, Gowda SM, Corbett AT. Automatically
detecting a student’s preparation for future learning:
help use is key. In: Fourth International Conference
on Educational Data Mining. Eindhoven, The Nether-
lands; 2011, 179–188.
32. Vellido A, Castro F, Nebot A. Clustering Educational
Data. Handbook of Educational Data Mining.Boca
Raton, FL: Chapman and Hall/CRC Press; 2011, 75–
92.
33. Ueno M. Online outlier detection system for learn-
ing time data in e-learning and its evaluation. In: In-
ternational Conference on Computers and Advanced
Technology in Education. Beijiing, China; 2004, 248–
253.
34. Merceron A, Yacef K. Measuring correlation of strong
symmetric association rules in educational data. In
Romero C, Ventura S, Pechenizkiy M, Baker RSJd, eds.
Handbook of Educational Data Mining. Boca Raton,
FL: CRC Press; 2010, 245–256.
35. Rabbany R, Takaffoli M, Za¨
ıane O. Analyzing partic-
ipation of students in online courses using social net-
work analysis techniques. In: International Conference
on Educational Data Mining. Eindhoven, The Nether-
lands; 2011, 21–30.
36. Trˇ
cka N, Pechenizkiy M, van der Aalst W. Process min-
ing from educational data. Handbook of Educational
Data Mining. Boca Raton, FL: CRC Press; 2011, 123–
142.
Volume 3, January/February 2013 25
c
2012 John Wiley & Sons, Inc.
Overview wires.wiley.com/widm
37. Tane J, Schmitz C, Stumme G. Semantic resource man-
agement for the web: an e-learning application. In:
International Conference of the WWW. New York;
2004, 1–10.
38. Mazza R, Milani C. GISMO: a graphical interactive
student monitoring tool for course management sys-
tems. In: International Conference on Technology En-
hanced Learning. Milan, Italy; 2004, 1–8.
39. Corbett A, Anderson J. Knowledge tracing: modeling
the acquisition of procedural knowledge. User Model
User-Adapted Interact 1995, 4:253–278.
40. Desmarais MC. Mapping question items to skills with
non-negative matrix factorization. ACM SIGKDD Ex-
plor 2011, 13:30–36.
41. Romero C, Ventura S. Preface to the special issue on
data mining for personalised educational systems. User
Model User-Adapted Interact 2011, 21:1–3.
42. Tang T, Daniel BK, Romero C. Preface to the special
issue on recommender systems for and in social and on-
line learning environments. Expert Syst J Knowl Eng.
In Press.
43. Kotsiantis S, Patriarcheas K, Xenos MN. A combina-
tional incremental ensemble of classifiers as a tech-
nique for predicting student’s performance in dis-
tance education. Knowl-Based Syst 2010, 23:529–
535.
44. Frias-Martinez E, Chen S, Liu X. Survey of data mining
approaches to user modeling for adaptive hypermedia.
IEEE Trans Syst Man Cybern C, 2006, 36, 6, 734–
749.
45. Pavlik P, Cen H, Koedinger K. Learning factors transfer
analysis: using learning curve analysis to automatically
generate domain models. Int Conf Edu Data Min 2009,
121–130.
46. Ayers E, Nugent R, Dean N. A comparison of student
skill knowledge estimates. In: International Conference
On Educational Data Mining. Cordoba, Spain; 2009,
1–10.
47. Hsia t, Shie A, Chen L. Course planning of extension
education to meet market demand by using data mining
techniques—an example of Chinkuo Technology Uni-
versity in Taiwan. Expert Syst Appl J 2008, 34:596–
602.
48. Wauters K, Desmet P, Noortgate W. Acquiring item
difficulty estimates: a collaborative effort of data and
judgment. In: International Conference on Educational
Data Mining. Eindhoven, The Netherlands; 2011,
121–128.
49. Mikut R, Reischl M. Data mining tools. WIREs: Data
Min Knowl Discov 2011, 1,5, 431–443.
50. Romero C, Ventura S, De Bra P. Knowledge discovery
with genetic programming for providing feedback to
courseware author. User Model User-Adapted Interact
2004, 14:425–464.
51. Merceron A, Yacef K. Educational data mining: a case
study. In: International Conference on Artificial Intel-
ligence in Education. Amsterdam; 2005, 1–8.
52. Becker K, Vanzin M, Ruiz D. Ontology-based filtering
mechanisms for web usage patterns retrieval. In: Sixth
International Conference on E-Commerce and Web
Technologies. Copenhagen, Denmark; 2005, 267–
277.
53. Avouris N, Komis V, Fiotakis G, Margaritis M, Voyi-
atzaki E. Why logging of fingertip actions is not enough
for analysis of learning activities. In: Workshop on Us-
age Analysis in Learning Systems, AIED Conference.
Amsterdam; 2005, 1–8.
54. Mostow J, Beck J, Cen H, Cuneo A, Gouvea E, Heiner
C. An educational data mining tool to browse tutor-
student interactions: time will tell! In: Proceedings of
the Workshop on Educational Data Mining.Amster-
dam; 2005, 15–22.
55. Bellaachia A, Vommina E. MINEL: a framework for
mining e-learning logs. In: Fifth IASTED International
Conference on Web-based Education. Mexico; 2006,
259–263.
56. Jovanovic J, Gasevic D, Brooks C, Devedzic V,
Hatala M. LOCO-Analyist: a tool for raising teacher’s
awareness in online learning environments. In: Euro-
pean Conference on Technology-Enhanced Learning.
Crete,Greece; 2007, 112–126.
57. Hershkovitz A, Nachmias R. Developing a log-based
motivation measuring tool. In: First International
Conference on Educational Data Mining. Montreal,
Canada; 2008, 226–233.
58. Selmoune N, Alimazighi Z. A decisional tool for qual-
ity improvement in higher education. In: International
Conference on Information and Communication Tech-
nologies. Damascus, Syria; 2008, 1–6.
59. Juan A, Daradoumis T, Faulin J, Xhafa F. SAMOS: a
model for monitoring students’ and groups’ activities
in collaborative e-learning. Int J Learn Technol 2009,
4:53–72.
60. Gaudioso E, Montero M, Talavera L, Hernandez-del-
Olmo F. Supporting teachers in collaborative student
modeling: a framework and an implementation. Expert
Syst Appl 2009, 36:2260–2265.
61. Romero C, Ventura S, Zafra A, De Bra P. Apply-
ing Web usage mining for personalizing hyperlinks in
Web-based adaptive educational systems. Comput Edu
2009, 53, 828–840.
62. Johnson M, Barnes T. EDM visualization tool: watch-
ing students learn. In: Third International Conference
on Educational Data Mining. Pittsburgh, PA; 2010,
297–298.
63. Pedraza-Perez R, Romero C, Ventura S. A Java desktop
tool for mining Moodle data. In: International Con-
ference on Educational Data Mining. Eindhoven, The
Netherlands; 2011, 319–320.
26 Volume 3, January/February 2013
c
2012 John Wiley & Sons, Inc.
WIREs Data Mining and Knowledge Discovery Data mining in education
64. Bakharia A, Dawson S. SNAPP: a bird’s-eye view of
temporal participant interaction. In: Proceedings of
the 1st International Conference on Learning Ana-
lytics and Knowledge. Vancouver, British Columbia,
Canada; 2011, 168–173.
65. Graf S, Ives C, Rahman N, Ferri A. AAT: a tool for
accessing and analysing students’ behaviour data in
learning systems. In: First International Conference
on Learning Analytics and Knowledge. Banff, Alberta,
Canada; 2011, 174–179.
66. Zafra A, Romero C, Ventura S. DRAL: a tool for dis-
covering relevant e-activities for learners. Knowl Inf
Syst. In Press.
67. Garc´
ıa-Saiz D, Zorrilla ME. A service oriented archi-
tecture to provide data mining services for non-expert
data miners. Decis Support Syst. In Press.
Volume 3, January/February 2013 27
c
2012 John Wiley & Sons, Inc.
