An Ontological Model for Assessment Analytics
Azer Nouira, Lilia Cheniti-Belcadhi and Rafik Braham
PRINCE Research Lab ISITCom, H- Sousse, Sousse University, Sousse, Tunisia
Keywords: Learning Analytics, Learning Analytics Models, Assessment Analytics, Ontology.
Abstract: Today, there is a growing interest in data and analytics in the learning environment resulting in a highly
qualified research concerning models, methods, tools, technologies and analytics. This research area is
referred to as learning analytics. Metadata becomes an important item in an e-learning system, many
learning analytics models are currently developed. They use metadata to tag learning materials, learning
resources and learning activities. In this paper, we firstly give a detailed injection of the existing learning
analytics models in the literature. We particularly observed that there is a lack of models dedicated to
conceive and analyze the assessment data. That is why our objective in this paper is to propose an
assessment analytics model inspired by the Experience API data model. Hence, an assessment analytics
ontology model is developed supporting the analytics of assessment data by tracking the assessment
activities, assessment result and assessment context of the learner.
1 INTRODUCTION
Learning analytics (LA) analyses the educational
data derived from student interaction with the
learning environment such as LMS (Learning
Management System) and MOOC (Massive Open
Online Courses) who generate large amounts of data
(Big Data). In the literature, many definitions have
been proposed for the term learning analytics. For
example, according to (Siemens., 2011) learning
analytics is the measurement, collection, analysis
and reporting of data about learners and their
contexts, for purposes of understanding as well as
optimizing learning and the environments in which it
occurs. The majority of learning analytics definitions
share a particular emphasis on converting
educational data into useful actions to improve the
learning processes (Lukarov et al., 2014).
Learning modeling is a key task in the emerging
research areas of learning analytics (LA). A learner
model represents information about a learner's
characteristics, states and activities, such as
knowledge, motivation and attitudes. In this research
work, we investigate the existing learning analytics
models by specifying some of their characteristics.
In fact, these models focus essentially on modeling
learning data (traces). However learning
environments generate different types of educational
data. Indeed, beyond the learning data there is
assessment data and communication data, etc.
Assessment is one of the major steps in the learning
process. In addition, assessment traces must be well
conceived in the same way as the learning traces, to
build a flexible and correct assessment model that
can support the analytic of assessment data. Our
research questions can be summarized in two major
questions:
 How can we model assessment data to build a
flexible assessment analytics model?
 What are the different assessment data that can
be conceived to build our assessment analytics
model?
This paper is structured as follows: in section 2 we
identify the most well known learning analytics
models by detailing some of their features. In section
3 we explore the assessment analytics concept. In
section 4 we present our assessment analytics
scenario and its analysis and requirements then, we
describe the assessment analytics process. Finally, in
section 5 and 6, we propose more details and we
describe semantically our proposed ontological data
model for assessment analytics.
2 RELATED WORK
Actually there are various formats of data
representation for usage data and they focus on the
Nouira, A., Cheniti-Belcadhi, L. and Braham, R.
An Ontological Model for Assessment Analytics.
DOI: 10.5220/0006284302430251
In Proceedings of the 13th International Conference on Web Information Systems and Technologies (WEBIST 2017), pages 243-251
ISBN: 978-989-758-246-2
Copyright © 2017 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
243

user’s activities. The user activities and their usage
of learning data objects in the different learning
environment are called usage metadata. Indeed
different models are proposed by researchers to
characterize usage data across learning systems.
Based on this data representation via models,
learning analytics can realize different analysis and
provide personalized and meaningful information to
improve the learning process. (Niemann et al., 2012;
Lukarov et al., 2014) present in their work the most
commonly used data models representations such a
CAM (Contextualized Attention Metadata) (Schmitz
et al., 2012), Activity Streams (J. Snell, M. Atkins,
W. Norris, C. Messina, M. Wilkinson, and R. Dolin.,
2015), Learning Registry Paradata (Paradata
specification., 2011) and NSDL Paradata (NSDL’s
TSPE., 2012).
2.1 CAM (Contextualized Attention
Metadata)
The CAM Model (Schmitz et al., 2012) allows
people to control the user interactions with the
learning environments. This model focuses on the
event itself rather than on the user or data object. So
many attributes are assigned for each event such as
its id, the event type, the timestamp, and a sharing
level reference. Each entity and also each session
can be described in a different and suitable way and
no information is duplicated. Each event can be
conducted in a session. The information can be
stored with different formats such XML, RDF,
JSON or in a relational database. In the literature we
found various researches concerning CAM. (Najjar
et al., 2006) discuss how CAM enables the
collection of rich usage to enhance user’s models,
predict usage patterns and feed personalization.
(Ochoa and Duval., 2006) provide a study showing
how CAM can be used to rank and recommend
learning objects. (Wolpers et al., 2007) propose a
CAM framework that is able to capture the
observations about the user activities with digital
content from different applications such as web
browser, multimedia player, LMS etc.
2.2 Activity Streams
Activity Streams (Snell et al., 2015) is a data format
for encoding and transferring activity/event metadata
published in 2011. An Activity Stream is a
collection of one or more individual activities
carried out by users. Each activity comprises a
certain number of attributes such as verbs, ids and
contents. An activity has three properties e.g. the
actor, the object, and the target. Many social
networks like facebook actually use Activity
Streams to store and manage user’s activities. Some
learner management systems (e.g. Canvas) are also
starting to develop analytics tools to generate
activity stream profiles for learners and teachers.
2.3 Learning Registry Paradata
Learning Registry Paradata (Paradata specification.,
2011) is an extended or modified version of Activity
Streams for storing aggregated usage information
about resources such as description, measure, and
date. The three main elements of Learning Registry
Paradata are actor, verb, and object. The verb refers
to a learning action and detailed information can be
stored. It’s important here to clear the difference
between metadata and paradata. Metadata describes
what a resource is, while paradata records how the
resource is being used. The learning registry is a
metadata agnostic. This means that it adds the
paradata about resource which is used, reused,
adapted, contextualized, tweeted, shared, etc.
Paradata complements metadata by providing an
additional layer of contextual information.
2.4 Learning Context Data Model with
Interest
The Learning Context Data Model with interest
(LCDM) (Thus et al., 2015). It mainly describes the
learner’s activities and the characteristics of the
learning environment. The learning context data
model is based on CAM representation. According
to (Thus et al., 2015) this data model considers two
points: the first one is to take into account which
type of learning activities that should be filtered and
the second point is how to mountain the semantic of
context information. The interest extension of
LCDM takes into account the weights of the
interests as well as their evolution over time. The
authors follow an iterative approach to develop a
new version of LCDM which satisfies specific
interest because context and interests represent
important features in the lifelong learner model.
2.5 The Experience API (xAPI)
xAPI called also Tin Can API is developed by
Advanced Distributed Learning Initiative (ADL)
(Experience API Working Group, 2013), and is
aimed at defining a data model for logging data
about students’ learning paths (Kelly and Thorn.,
2013). The xAPI presents a flexible data model for
WEBIST 2017 - 13th International Conference on Web Information Systems and Technologies
244

logging data about the learner experience and
performance. The xAPI specification is suitable with
the learning analytics purpose, since it tracks and
stores the experience and the performance of the
learner (learning traces).
The xAPI specification is based on two main
parts. The first part is the format of learning activity
statement and the second part is the Learning Record
Store (LRS). LRS is the element responsible for
storage and exchange of learning activities traces
presented as activities statements. The activity
statement is a key part of the xAPI data model. All
learning activities are stored as statements such as:
“I did this” of the form actor, verb and object and it
can be extended with some optional properties like
result and context.
The xAPI specification is flexible. Hence among
web-based formal learning, xAPI is capable of
tracking informal learning, social learning, and real
world experiences. A wealth of
examples related to
the learning
activities that can be tracked include
reading an article, watching a training video or
having a conversation with a mentor. As a result the
LRS stores various statements concerning content
view, video consumption and assessment result. As a
result, it is possible to access and query the data
stored in Learning Record Store (LRS) and therefore
we could provide different services such as
statistical service, reporting service, assessment
service and semantic analysis.
In the literature xAPI has been widely
implemented. Hence, we found a several research
works related to learning analytics using xAPI. For
instance (Kitto et al., 2015) present a solution for
Learning Analytics beyond the LMS which is the
Connected Learning Analytics (CLA) toolkit, which
enables data to be extracted from social media such
as Google+, Twitter, Facebook, etc and imported
into a Learning Record Store (LRS), the way it is
defined by the new xAPI standard. Many other
works also can be founded in (Brouns et al., 2014),
(Corbi and Burgos, 2014), (Del Blanco et al, 2013).
All data models mentioned above are learning
analytics centric. That is to say that they focus on
how to present well and to conceive learning
activities, users and data objects. These models will
be used for analytical purpose to improve the quality
of the learning process. However learning
environments generate different types of educational
data. Among them there is assessment data and
communication data, etc. Assessment is one of the
major steps in the learning process. In addition,
assessment data must be tracked as well and
processed
According to our research, there is a lack of models
that focus on assessment analytics, that is to say, a
model which is interested in assessment data. The
only learning analytics data model which is
previously detailed and which can support analyzing
assessment data is the TIN CAN API (xAPI) since it
contains an optional property in its sentence format
named result that records information about
assessment result. But xAPI is presented as an e-
learning standard for tracking data interoperability in
the whole learning process and does not focus
particularly on assessment. Besides, during our
research we did not found any paper that focuses on
tracking assessment data with xAPI specification. Is
this due to the weakness of xAPI standard in
tracking assessment data? When we investigate the
result property which is an optional property, we
noticed that is described with different metadata that
can record information about the assessment result
such as the score, the success, the completion the
duration and the response. All these assessment
results are very important, but according to our point
of view, these results are insufficient and need to be
extended and annotated to ensure a several
assessment result tracked that can help later for
assessment analytics. The investigation of the
context property of the xAPI data model leads as to
deduce that the context metadata of xAPI data model
are not related to assessment context. In fact, all of
them represent information about the context of
learning activity such as the instructor and the team
that the statement is related to, the platform used, the
language of the statement recorded, etc. Any
information is recorded about the context of
assessment such as the type of assessment, the form
of assessment and the technique of assessment.
Our contribution is based on the weakness of the
existing xAPI data model dedicated to assessment
data summarized into two major points:
 Insufficiency of information dedicated to the
track of the assessment result.
 Lack of information dedicated to the assessment
context.
3 ASSESSMENT ANALYTICS
One of the most important steps in the learning
process is assessment; a successful learning
environment must provide effective assessment of
learners. Assessment is both ubiquitous and very
meaningful as far as students and teachers are
concerned (Ellis., 2013). Actually the new learning
environments such as MOOCs generate big
An Ontological Model for Assessment Analytics
245

assessment data (Big Data) given the massive
number of courses proposed and the great number of
learners enrolled. These assessment data must be
processed and analyzed too.
When we focus on assessment data that means,
we study the assessment activities and the
assessment result left by learners, we can launch a
new source of data that can be analyzed and give
new and different indicators to be interpreted and
hence contribute to the improvement of the field of
learning analytics. This research area is called
assessment analytics. According to (Ellis., 2013)
The role that assessment analytics could play in the
learning process is significant. Yet it is
underdeveloped and underexplored. Assessment
analytics has the potential to make valuable
contribution to the field of learning analytics by
extending its scope and increasing its usefulness.
The assessment analytics is the analytics of
assessment data within learning analytics strategy.
(Cooper., 2015) offers an assessment analytics
definition, which is based on the learning analytics
definition of (Siemens., 2011) with a little
modification: assessment analytics is the
measurement, collection, analysis and reporting of
data about learners and their contexts, for purposes
of understanding and optimizing learning and the
environments from which the data derives
assessment.
This open research area and
development topic is addressed by this paper in
order to propose an assessment analytics model.
4 ASSESSMENT ANALYTICS
SCENARIO
4.1 Assessment Analytics Scenario
Description
This section illustrates the context of our research
using a use case scenario that includes an assessment
analytics scenario. Let’s consider the following
situation:
Peter is very passionate of computer sciences, so
he decides to register in a MOOC environment for
studying a course on object oriented modeling. After
the sign stage, Peter can start the learning process. A
set of different learning activities are made by Peter
when he starts interaction with the learning content
such as watching course video, reading texts,
commenting etc. Then Peter is invited to start the
stage of assessment (formative assessment,
summative assessment or it can be also diagnostic
assessment and in this case it will be done before the
learning step). Hence a set of assessment activities
are triggered such as answering, completing, failing
and scoring during the interaction with quizzes or
MCQ (multiple choice question). The MOOC
environment that Peter uses must keep track about
every activity done by this learner including the
learning activities and the assessment activities.
In our case we will focus only on the assessment
data that are particularly the assessment activities
and the assessment results left by Peter. Assessment
data will be conceived and analyzed properly using
analytics strategy, then interpreted for improving the
quality of the learning process and identifying the
opportunities for feedbacks, interventions,
adaptations, recommendation, personalization, etc.
4.2 Assessment Analytics Scenario
Analysis
According to the above scenario, we can identify the
following challenges.
We suppose here that the learning standard chosen
for tracking data interoperability is xAPI, hence all
learning and assessment activities are stored in the
LRS (Learning record store).
Since our objective here is to analyze the
assessment activities, we will focus only on the set
of assessment activities stored in the LRS.
Now here is an example of a set of an assessment
activities stored in the LRS of the learner Peter:
1. Peter attempted quiz_1
2. Peter completed the quiz_1 with a passing score
75%
3. Peter passed the quiz_1
4. Peter completed the quiz_1 with completion false
5. Peter completed the quiz_1 in 6 minutes
6. Peter attempted quiz_2
7. Peter completed the quiz _2 with a passing score
15%
8. Peter failed the quiz_2
9. Peter attempted quiz_2
10. …
To ensure a consistent assessment analytics
engine, it is necessary to track and manage a set of
metadata in relation with assessment activities and
results, the tracked assessment result by the xAPI
specification seems to be insufficient and need to be
more annotated. Besides the assessment activities
and the assessment result it’s necessary to track and
manage also the assessment context data such as: the
assessment types, forms and techniques, the
assessment environment and session. Here is an
WEBIST 2017 - 13th International Conference on Web Information Systems and Technologies
246

example of the metadata values corresponding to the
different assessment analytics that can be recorded
by the specification of our proposed assessment
analytics model:
Environment: MOOC
Assessment Form: Automated Assessment
Assessment Type: Formative Assessment
Assessment Technique: Closed question format:
MCQ
Session: The date and the hour of the logging.
ID: a unique number given to each assessment
activity
All these assessment data must be conceived in
an appropriate format supporting analytics in such a
way that it can help providing users with efficient
personalization and adaptation services. An
important question can be posed now, how can we
conceive assessment data to build a correct and
complete assessment analytics model?
5 ONTOLOGICAL MODEL FOR
ASSESSMENT ANALYTICS
Our objective in this section is to propose an
assessment analytics model dedicated to conceive
assessment data such as assessment activity,
assessment result and assessment context. This
model will be inspired from xAPI data model
specification. In this section, we propose an
assessment analytics model based on the data
derived from assessment. This model allows people
to control user interaction with assessment
resources. More precisely we will focus only on
assessment data. According to (Moody., 1998) there
are eight general requirements which describe data
models. These are: completeness, correctness,
integrity, flexibility, understandability, simplicity,
integration and implementability.
To ensure a consistent representation of our
proposed model for assessment analytics, it will be
interesting to develop an ontological model for
assessment analytics considering the several
advantages given by the use of ontologies like
improving reusability and interoperability,
aggregation of the scattered data in the web,
permitting inferences and contribute coherence and
consistency rules. The proposed model is called
assessment analytics ontology (AAO). In order to
develop our ontology, we follow the most important
steps detailed in (Noy and McGuiness., 2005), the
first step is to enumerate the most important terms in
our ontology through specification of classes such
as: assessment type, assessment statement, etc. Then
it’s necessary to define the classes and the class
hierarchy. After that, we need to define the class
properties and attributes and finally determine the
facets of attributes. To develop our AAO model,
several tools are available and can be used such as
SWOOP, Protégé and WebOnto. In our case, we
used Protégé that offers a simple, complete and
expressive graphical formalism. It also facilitates the
design activity. All the figures below are designed
with VOWL (Lohmann and al., 2014) plug-in,
which can be integrated very easily to protégé and
used to represent graphically the different types of
properties such as object properties, data type
properties and subclasses relations. Figure 2 below
shows a graphical representation of our assessment
analytics ontology with the tool Protégé.
Figure 1: Graphical representation of assessment analytics
ontological model.
The developed ontology gives us the opportunity
to model the main concepts of our assessment
analytics model in terms of its classes and its
hierarchy. The main class of our assessment
analytics ontology is the assessment activity class.
This class is linked with has-a relations to a set of
classes describing the assessment activity context
metadata and the assessment activities. For instance,
we cite the assessment environment class and his
possible instances such as MOOC (Massive Open
Online courses), LMS (Learning Management
system) or PLE (Personal Learning Environment).
Technically, it can be described by using the element
of enumeration OWL:one Of.
<owl:class rdf: id = “Asses_Envr”
<owl:oneOf rdf:parseType="Collection">
<owl:Thing rdf:about="#MOOC"/>
<owl:Thing rdf:about="#LMS"/>
<owl:Thing rdf:about="#PLE"/>
An Ontological Model for Assessment Analytics
247

</oneOf>
The second class is the assessment session class that
contains information about the assessment session
such as activity id and the date of logging as
datatype properties. It serves to ensure that no
information is duplicated.
Also we have the assessment technique class and his
possible instances such as MCQ (Multiple Choice
Question), MRQ (Multiple Response Question), T/F
(True or False question) and Fill in Blanks question.
<owl:class rdf: id = “Asses_Tech”
<owl:oneOf rdf:parseType="Collection">
<owl:Thing rdf:about="#MCQ"/>
<owl:Thing rdf:about="#MRQ"/>
<owl:Thing rdf:about="#T/F"/>
<owl:Thing rdf:about="#Fill in Blanks"/>
</oneOf>
The third class is the assessment form class and its
three possible instances: thediagnostic assessment,
the formative assessment or the summative
assessment.
<owl:class rdf: id = “Asses_Form”
<owl:oneOf rdf:parseType="Collection">
<owl:Thing rdf:about="#Diagnostic"/>
<owl:Thing rdf:about="#Formative"/>
<owl:Thing rdf:about="#Summative"/>
</oneOf>
Then, we have the assessment type class that may
consist of automated assessment, self assessment or
peer assessment.
<owl:class rdf: id = “Asses_Type”
<owl:oneOf rdf:parseType="Collection">
<owl:Thing rdf:about="#Automated"/>
<owl:Thing rdf:about="#Peer assessment"/>
<owl:Thing rdf:about="#Self assessment"/>
</oneOf>
It’s important to mention that all the instances of the
classes cited above have exactly one value for a
particular property such as has_Environment,
has_session, has_form, has_type and has_technique.
Technically, we can use the cardinality constraint
owl:cardinality.
<owl:Restriction>
<owl:onProperty rdf:resource="#Has_Form" />
<owl:cardinality rdf:datatype =
"&xsd;nonNegativeInteger">1</owl:cardinality>
</owl:Restriction>
All these metadata are very helpful for enriching our
assessment analytics model. They are useful later in
the stage of analytics.
Finally, we have the most important class and the
core of our ontological model which is the
assessment statement class used to represent the
assessment experience of the learner. The
assessment statement class of the proposed ontology
is able to capture and formulate sentences of the
form: Peter completed the quiz with a passing score
of 80%, Jane completed the quiz in 10 minutes, and
Daniel failed the quiz with 20%. This leads us to
conclude that each assessment activity has at least
one or more assessment statements.
<owl:Restriction>
<owl:onProperty
rdf:resource="#Has_Statement" />
<owl:minCardinality rdf:datatype =
"&xsd;nonNegativeInteger">1</owl:minCardinality
> </owl:Restriction>
In the next section, we will focus on describing in
details the different properties and data type
properties of the assessment statement class of our
proposed ontological assessment analytics model.
6 SEMANTIC DESCRIPTION OF
THE ONTOLOGICAL
ASSESSMENT ANALYTICS
MODEL (OAA)
In this section we will present the semantic
description of some classes of our ontological model
for assessment analytics.
6.1 Assessment Statement Class
From figure 3 below, we can observe that the
assessment statements have 4 required properties
conform to its format such as Peter failed the quiz
with 20%. From this example we can extract four
properties: the verb, the actor, the object and the
result. Let us began with the first property which is
the actor property that refers to whom? e.g., Peter
that means the learner in our context. The second
one is the verb property which is a key part of an
assessment analytics sentence; it describes the action
performed by the learner when he interacts with an
assessment resource. As examples of assessment
verbs we may cite: answered, completed, and
scored.
The third property is the object that form the
third part of the statement, which refers to what was
experienced in the action defined by the verb e.g. the
quiz. And finally we have one of the most important
properties of the assessment analytics engine which
is the result property, that record information about
assessment result such as score, completion and
duration. Besides we can annotate the assessment
statement class more thoroughly by some data
WEBIST 2017 - 13th International Conference on Web Information Systems and Technologies
248

properties like the record time of the assessment
statement in the LRS (Learning Record Store).
Figure 2: The semantic description of the assessment
statement model.
6.1.1 Semantic Network of Agent Model
Figure 3: The semantic description of the actor model.
As described in the figure 4, the assessment actor
class is described by using some classes of FOAF
ontology (Brickley and Miller., 2014) which is an
ontology describing individuals, their activities and
relations with other people to define agent and
group. Hence the assessment actor is a subclass of
FOAF:Agent and is a member of FOAF:Group
which is also a subclass of FOAF:Agent. Each
assessment actor holds an online account which
represents the provision of some form of online
service. Each assessment actor has two different
type of levels, the second one is the unsupervised
level which is the level of the learner before starting
the assessment process and the second one is the
supervised level which is the real and the concrete
level of the learner related to his performance in the
assessment process identified automatically by
analyzing the assessment traces of the learner.
6.1.2 Semantic Network of Verb Model
The verb class describes the action performed during
the learning experience and more specifically in the
assessment experience. Figure 5 below, shows that
the verb class is described by 2 attributes. The first
one is the language of the verb and the second one is
the value of the verb. The assessment verb is an
important class of this model. Its possible instances
are the verbs related to the assessment experience
like scored, failed and passed.
Figure 4: The semantic description of the verb model.
6.1.3 Semantic Description of Assessment
Object Model
The assessment object can be annotated and
described through using the standards LOM
(Learning Object Metadata) (IEEE Learning
Technology Standards Committee,. 2002) and
Dublin Core (The Dublin Core Metadata Initiative.,
2001)
and can be described more thoroughly by
some other attributes like the coefficient , the
module and the level of difficulty of each
assessment object.
Figure 5: The semantic description of the assessment
object model.
An Ontological Model for Assessment Analytics
249

6.1.4 Semantic Description of the
Assessment Result Model
Figure 6: The semantic description of the assessment
result model.
The assessment result class is one of the main
classes in our assessment analytics models since it
contains a several datatype properties that records
information about assessment results such as score,
completion, duration, success inspired from the
xAPI specification and others assessment result
metadata proposed such as the attempt, number of
correct answer, number of wrong answer and
number of unanswered question. This is particularly
helpful for future assessment analytics. The
assessment result class should be described with a
rich metadata. These bits of information are
important and we should record and use them since
our objective is to conceive a complete, correct and
flexible assessment analytics models. These
attributes can be very helpful later in the stage of
analytics.
The design of our AAO model brings some
advantages for future use. For instance, one of the
most important advantages of using ontology resides
in its structure that is consistent with the logic
description and its coupling with the inference
engines allowing expert system to deduce logical
reasoning and conclusions. In our case, the AAO
model can be a consistent model supporting the
assessment analytics purpose by deducing logical
reasoning and conclusions about the assessment
activities, result and context of each learner.
7 CONCLUSION AND FUTURE
WORK
According to (Lukarov et al., 2014) learning
analytics can perform different analysis and provide
personalized and meaningful information to improve
the learning and teaching process. In this paper, we
presented a detailed presentation of the set of
learning analytics models existing in the literature.
The existing learning analytics models focus
essentially on learning data, according to our
research there is a remarkable lack of models that
focus on assessment data. That means a model that
is interested in data derived from assessment. That is
why we tried to propose an ontological assessment
analytics model (AAO) inspired by the Experience
API data model. This model focus essentially on
assessment data, meaning that it tracks the
assessment activities of the learner, then tries to
conceive them in a flexible and consistent way and
finally store them in a specific module for
assessment data storage to be later accessed for
analytics purposes.
Concerning our further work we will try to extend
our ontology with additional metadata and hence
capture the semantic description to deduce logical
reasoning and conclusions for assessment analytics.
REFERENCES
Brickley, D. and L. Miller (2014). FOAF Vocabulary
Specification 0.99.
Brouns, F., Tammets, K., and Padrón-Nápoles, C. L.
(2014). How can the EMMA approach to learning
analytics improve employability?.
Corbi, A., and Burgos, D. (2014). Review of Current
Student-Monitoring Techniques used in eLearning-
Focused recommender Systems and Learning
analytics. The Experience API & LIME model Case
Study. IJIMAI, 2(7), 44-52.
Cooper, A. (2015). Assessment Analytics. In Eunis E-
learning task force workshop.
Del Blanco, Á., Serrano, Á., Freire, M., Martínez-Ortiz, I.,
and Fernández-Manjón, B. (2013). E-Learning
standards and learning analytics. Can data collection
be improved by using standard data models?.
In Global Engineering Education Conference
(EDUCON), 2013 IEEE (pp. 1255-1261). IEEE.
Ellis, C. (2013). Broadening the scope and increasing the
usefulness of learning analytics: The case for
assessment analytics. British Journal of Educational
Technology, 44, 4, 662-664.
Experience API Working Group (2013). Experience API.
Version 1.0.1.
IEEE Learning Technology Standards Committee (2002).
WEBIST 2017 - 13th International Conference on Web Information Systems and Technologies
250

Draft standard for learning object.
J. Snell, M. Atkins, W. Norris, C. Messina, M. Wilkinson,
and R. Dolin (2015). Activity streams 2.0 w3c
working draft. W3C Working Draft, W3C.
http://www.w3.org/TR/2015/WD-activitystreamscore/
Kelly, D. and Thorn, K. (2013). Should Instructional
Designers care about the Tin Can API? In eLearn, 3, 1.
Kitto, K., Cross, S., Waters, Z., and Lupton, M. (2015).
Learning analytics beyond the LMS: the connected
learning analytics toolkit. In Proceedings of the Fifth
International Conference on Learning Analytics And
Knowledge (pp. 11-15). ACM.
Lohmann, S., Negru, S., and Bold, D. (2014). The
ProtégéVOWL plugin: ontology visualization for
everyone. In European Semantic Web Conference,
395-400. Springer.
Lukarov, V., Chatti, M. A., Thus, H., Kia, F.S., Muslim,
A., Greven, C., and Schroeder, U. (2014).Data Models
in Learning Analytics. In DeLFI Workshops 88-95.
Moody, D. L. (1998). Metrics for evaluating the quality of
entity relationship models. In International
Conference on Conceptual Modeling, 211-225.
Niemann, K., Scheffel, M., and Wolpers, M. (2012). An
overview of usage data formats for recommendations
in TEL. In 2 nd Workshop on Recommender Systems
for Technology Enhanced Learning, 95.
NSDL’s Technical Schema for Paradata Exchange.
(2011). Retrieved from
http://nsdlnetwork.org/stemexchange/paradata/schema.
Noy, N. F. and McGuinness, D. L. (2005). Développement
d’une ontologie 101: Guide pour la création de votre
première ontologie. Université de Stanford. Rapport
technique.
Ochoa, X. and Duval, E. (2006). Use of contextualized
attention metadata for ranking and recommending
learning objects. In Proceedings of the 1st
international workshop on Contextualized attention
metadata: collecting, managing and exploiting of rich
usage information, 9-16.
Paradata Specification. v1.0, (2011). Retrieved from
https://docs.google.com/document/d/1IrOYXd3S0FU
wNozaEG5tM7Ki4_AZPrBn-pbyVUz-
Bh0/edit?hl=en_US.
Schmitz, H. C., Wolpers, M., Kirschenmann, U., and
Niemann, K. (2012). Contextualized Attention
Metadata’. Human Attention in Digital Environments,
186-209.
Siemens, G. (2011). What Are Learning Analytics? In 1st
international conference on Learning Analytics and
Knowledge. 2015.
The Dublin Core Metadata Initiative. (2001) DC-Library
Application Profile, Dublin Core.
Thus, H, Chatti, M.A, Brandt, R., and Schroeder, U.
(2015) Evolution of Interests in the Learning Context
Data Model. In Design for Teaching and Learning in a
Networked World. Springer.
Wolpers, M., Najjar, J., Verbert, K., and Duval, E. (2007).
Tracking actual usage: the attention metadata
approach. Educational Technology & Society, 10, 3,
106-121.
An Ontological Model for Assessment Analytics
251