=Paper=
{{Paper
|id=Vol-2519/paper8
|storemode=property
|title=Measurement Task Ontology
|pdfUrl=https://ceur-ws.org/Vol-2519/paper8.pdf
|volume=Vol-2519
|authors=Lucas Santos,Monalessa Barcellos,Ricardo Falbo,Cássio Reginato,Patrícia Campos
|dblpUrl=https://dblp.org/rec/conf/ontobras/SantosBFRC19
}}
==Measurement Task Ontology==
https://ceur-ws.org/Vol-2519/paper8.pdf
Measurement Task Ontology
Lucas A. Santos, Monalessa P. Barcellos, Ricardo A. Falbo, Cássio. C. Reginato,
Patrícia C. M. Campos
Ontology and Conceptual Modeling Research Group (NEMO), Department of
Computer Science, Federal University of Espírito Santo– Vitória – ES – Brazil
{lucasaugsantos,patmarcal}@gmail.com,{falbo,monalessa,
cassio.reginato}@inf.ufes.br
Abstract. Measurement is a key process in several domains. Although it has
particularities according to the application domain, part of the knowledge
related to the measurement process is common to all of them. This paper
presents the Measurement Task Ontology (MTO), which establishes a common
conceptualization regarding core aspects of measurement. MTO addresses
behavioral aspects of the measurement process and extends the Core Ontology
on Measurement, combining both domain and task-related aspects of
measurement. As evaluation of MTO, we described a real-world case scenario
regarding a laboratory test, showing that MTO is capable of representing real
world situations.
Keywords: Measurement, Ontology, Task Ontology
1. Introduction
Measurement is an important subject in various domains, since it provides information
for getting conclusions and making decisions. Part of the knowledge related to
measurement remains unchanged across different application domains. However,
depending on the application domain, measurement presents some particularities.
There are several standards and references about measurement. Some of them
focus on general aspects, such as VIM (2012), which determines a vocabulary about
measurement as an attempt to standardize terminology across different domains. Others
focus on specific domains (e.g., [ISO, 2007], which concerns Software Measurement), to
address particularities when measurement is applied in their context. When analyzing
different standards and references, it is possible to identify common concepts, although
the terminology used is largely distinct. It is not rare finding in different references the
same concept being designated by different terms and the same term being used to refer
to different concepts.
To deal with these semantic problems, we need a common conceptualization
about measurement. Ontologies are acknowledged as being quite appropriate to solve
conceptual ambiguities and knowledge vagueness. An ontology is a formal representation
of a common conceptualization of a universe of discourse [Guarino, 1998]. Therefore, it
is a useful instrument for reducing conceptual ambiguities and inconsistencies, and for
making knowledge structures clearer. Ontologies can focus on describing the concepts of
a domain (domain ontologies) or describing general knowledge about processes that may
occur in several application domains (task ontologies).
Copyright © 2019 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Aiming to establish a common conceptualization regarding the core aspects of
measurement, Barcellos et al. (2014) proposed the Core Ontology on Measurement
(COM). COM is a core ontology in the sense that it provides a precise definition of the
structural knowledge in the measurement field that spans across several application
domains. However, COM does not cover aspects that are common in more complex
measurements, such as sampling and measurement analysis through successive data
analysis. Moreover, COM represents only structural knowledge and does not clearly
address behavioral aspects describing the measurement process, by identifying its
activities and the flow between them, their inputs, outputs and actors involved in their
execution. To reach a semantic agreement in a broader sense, it is important to achieve a
common understanding regarding both domain (structural) and task-related (behavioral)
aspects of measurement [Barcellos et al., 2013]. Therefore, in this paper we propose the
Measurement Task Ontology (MTO), which describes behavioral aspects of the
measurement process and extends COM to deal with application domains in which
measurement involves sampling and measurement analysis through successive data
analysis.
This paper is organized as follows. Section 2 provides the background for the
paper, discussing briefly measurement and ontologies. Section 3 presents the
Measurement Task Ontology. Section 4 evaluates MTO by showing that the ontology is
capable of describing factual situations. Section 5 discusses related works. Finally,
Section 6 presents our final considerations.
2. Background
Measurement can be defined as a set of actions aiming to characterize an entity by
analyzing values attributed to its properties. The main activities involved in measurement
are planning, execution and analysis. During planning, the entities to be measured are
identified (e.g., objects, phenomena), as well as the properties to be measured (e.g., size,
cost), the measures to be used to quantify those properties (e.g., size in meters), and how
measurement of each measure should be carried out (e.g., the area of a square object must
be measured by applying the formula a=s2, where s is the side length). Decisions
regarding which entities and properties are to be measured and which measures are to be
used should be driven by goals. Once measurement is planned, it can be performed.
Measurement execution consists in collecting data for the measures by applying
measurement procedures. Finally, measurement analysis comprises analyzing collected
data, thus providing basis for problem solving or decision making [VIM, 2012] [ISO,
2007].
Measurement occurs in various application domains. Thus, there are key concepts
and activities present in all of them. Ontologies can be built to make explicit this common
conceptualization, since they describe the information semantics and turn its content
explicit [Wache, 2001].
According to Guarino (1998), ontologies can be classified according to their
generality level into: foundational ontologies, which describe very general concepts, such
as object, event, etc.; domain ontologies, which describe the conceptualization related to
a generic domain (e.g., medicine, law); task ontologies, which describe the
conceptualization related to a generic task (such as diagnosis and sale); and application
ontologies that describe concepts dependent on a particular domain and task. Although
the use of domain ontologies has become increasingly common, the use of task ontologies
is rarer. Task ontologies are important because they contextualize the concepts defined in
domain ontologies assigning meaning to services, functionalities, activities, flows, and
related information [Martins and Falbo, 2008].
Task ontologies should capture two major kinds of knowledge [Martins and Falbo,
2008]: (i) task decomposition, including control flow, and (ii) knowledge roles played by
entities from the domain in the fulfillment of the task. These two kinds of knowledge are
very inter-related, although they capture different views of a task. In fact, they represent
different modeling aspects, i.e. different dimension of modeling that emphasizes
particular views of the same portion of the reality. Thus, we need different models for
representing them. Martins and Falbo (2008) proposed the use of two UML diagrams for
representing task ontologies: activity diagrams, capturing task decomposition into sub-
tasks and how knowledge roles act in their fulfillment; and class diagrams, modeling the
concepts involved and their relation. We follow this proposal to represent our ontology.
In the next section, we present the Measurement Task Ontology, which describes
aspects of these two perspectives of the measurement process. It is worthwhile to point
out that, although we use the term “task ontology”, which is already consecrated in the
field of ontologies, in fact we are talking about a process ontology, in the sense that we
are interested in describing the measurement process as a whole, and not tasks with low
granularity level. Moreover, we should emphasize that our ontology is a reference
ontology, i.e., a special kind of conceptual model representing a model of consensus
within a community. It is a solution-independent specification with the aim of making a
clear and precise description of entities in the universe of discourse, for the purposes of
communication, learning and problem-solving. We are not interested in an
implementation of this ontology for purposes of reasoning, for instance. As advocated by
Guizzardi (2007), a reference ontology should be developed taking truly ontological
distinctions into account, i.e. a reference ontology should be grounded in a foundational
ontology. Thus, our task ontology is developed grounded in the Unified Foundational
Ontology (UFO) [Guizzardi, 2005] [Guizzardi, 2008]. Due to space limitations, in this
paper we present only the core fragment of our ontology. Moreover, deeper discussions
about the use of UFO to ground our ontology are out of scope of this paper. Discussions
in this regard can be found in [Barcellos et al., 2014].
3. MTO: Measurement Task Ontology
As a process ontology, MTO is supposed to answer the following competency questions:
(i) Which are the activities of the measurement process? (ii) Who is responsible for
performing them? (iii) How the activities are decomposed into sub-activities? (iv) What
is the control flow between them? (v) What are the inputs and outputs of each activity?
Following the guidelines given in [Martins et al., 2008], for capturing the
conceptualization involved in the measurement process, we developed two conceptual
models: a behavioral model (Subsection 3.1) and a structural conceptual model
(Subsection 3.2).
3.1 MTO Behavioral Model
Figure 1 presents the main activities of the measurement process and the roles responsible
for performing each activity. Although four out of five activities in the diagram (activities
identified with ) are decomposed into sub-activities, in this paper, due to space
limitation, we show only the decomposition of the “Plan Measurement Process” activity
(see Figure 2). In fact, all the activities shown in both diagrams (figures 1 and 2) are
complex actions in the sense of UFO. In UFO [Guizzardi et al., 2008], actions are
intentional events. Complex actions are actions involving the participation of different
agents and objects. Every activity of the MTO behavioral models has the participation of
one agent and of one or more objects.
Figure 1. Overview of the Measurement Process
The models also present some stereotypes alongside to the object flows to capture
distinctions made in UFO [Guizzardi et al., 2008] related to object participation in
actions, namely: creation, indicating that an object is created by the action; change,
indicating that some property of the object changed; and usage, when the object is used
without changing any of its properties. Since the behavioral model presents a complex
process with non-linear flows, there are also decision nodes responsible for bypassing
some activities and also returning to previous ones if needed. Next, the measurement
process is described. In the text, activities are written in bold. Italics is used to identify
actors that perform the activities and objects participating in the activities as input or
output. The objects involved in the behavioral models represent instances of concepts of
the MTO structural model. Some of the objects cited in the text are not illustrated in the
figures for sake of legibility.
The measurement process starts with the Measurement Process Planner
performing the Plan the Measurement Process activity. In this activity, the
Measurement Process Planner plans how other activities of the measurement process
(namely, Perform Sampling, Perform Measurement and Analyze Measurements) should
be performed. Its main result is a set of Measurement Planning Items. A Measurement
Planning Item defines the plan to be followed to perform the measurement process by
specifying, according to the goals to be achieved, what is to be measured (i.e., the
Measurable Entity Type (e.g., Person, River) or the Measurable Entity (e.g., John,
Amazonas River) and its Measurable Element (e.g., cholesterol, turbidity)); which
Measure is to be used (e.g., cholesterol in mg/dl, turbidity in Nephelometric Turbidity
Units); which procedures are to be adopted in each measurement-related activity; which
device types (or devices) are to be used in them; and who are responsible for performing
each of them. Figure 2 details this activity.
Figure 2. Detailing of Plan the Measurement Process
Measurement should be driven by goals [VIM, 2012] [ISO, 2007]. Therefore, the
first activity of Plan the Measurement Process is Establish Measurement Goals, where
the goals to be achieved through measurement are defined (for example, a doctor can have
the goal “check John’s health”). The output of this activity is a set of Measurement
Planning Items, which at this time specify only the Measurement Goals to be achieved.
The next activity is Identify Information Needs, when information needed to achieve
the established goals are identified and added to the respective Measurement Planning
Items.
For example, to achieve the goal “check John’s health”, the doctor may need to
know “what is John’s cholesterol level” and “what is John’s blood pressure”. The inputs
for this activity are Measurable Entity Types (or Measurable Entities) and Measurable
Elements, which indicate, respectively, the entities that can be characterized in the
measurement process and their properties that can be measured. For example, the above
cited information needs refer to the measurable entity John and to its measurable elements
cholesterol and blood pressure. As shown in Figure 2, after being created, Measurement
Planning Items are updated in each activity with new information defined in that activity
and thus it serves as an input for the succeeding activity. Thereby, after the Identify
Information Needs activity, the Measurement Planning Items contain the Measurement
Goals previously established and the related Information Needs.
Once defined the goals to be achieved and the information needs to be met, the
next activity is Identify Measures, in which the Measures to be used to provide the
information needs are identified. For example, the measures “cholesterol in mg/dl” and
“blood pressure in mmHg” could be used to meet the information needs previously cited.
In some circumstances, it may be necessary to collect and prepare samples to be
used in the next activities of the measurement process. For example, to measure the
cholesterol of a person, a blood sample must be collected and prepared. When this is the
case, the next activity in Plan the Measurement Process is Plan Sampling. This activity
refers to planning how samples will be collected and, if necessary, how they will be
prepared for measurement. In this activity, the planner defines Sampling and Sample
Preparation Responsible, Sampling and Sample Preparation Procedures to be adopted,
and Sampling and Sample Preparation Device Types (or Devices) to be used. This
information is added to the respective Measurement Planning Item, updating it.
Once sampling is planned, or if sampling is not needed, the next activities are
Plan Measurement and Plan Measurement Analysis. The former refers to planning
how a measure should be measured. That is, it defines the Measurement Responsible, the
Measurement Procedure to be adopted and the Measurement Device Types (or Devices)
to be used. Analogously, the later concerns planning how collected data should be
analyzed to get conclusions. It defines the Measurement Analysis Responsible, the
Measurement Analysis Procedure to be adopted and the Measurement Analysis Device
Types (or Devices) to be used. After each of these activities, the respective Measurement
Planning Item is updated with the information accordingly.
It is possible to notice that, when planning a measurement-related activity, there
is a pattern of information that must be defined: the responsible for performing the
activity, the procedures to be adopted and the device types (or devices) to be used. The
responsible refers to the person (e.g., Mary), organization (e.g., Lab A) or role (e.g.,
Doctor) responsible for the activity. Procedures describe how the activity should be
performed and they can require the use of device types (e.g., Weight Balance) or specific
devices (e.g., Mary’s weight balance). For example, the measurement procedure to
measure the blood pressure in mmHg using the device type Blood Pressure Aneroid
Monitor could be “put the stethoscope earpieces into your ears; place the stethoscope
disk on the inside of the crease of the patience elbow; inflate the cuff at a rapid rate by
squeezing the rubber bulb; slightly loosen the valve and slowly let some air out of the
cuff; read the systolic pressure looking at the pointer on the dial when you listen the
heartbeat; read the diastolic pressure looking at the pointer on the dial when you stop
listening the heartbeat”.
After performing the Plan the Measurement Process activity, Measurement
Planning Items are fully set, i.e. the measurement process is planned, and the planned
activities can be now performed. Thus, if sampling is needed, the Sampling Performer
Performs Sampling (see Figure 1). Although not shown in the figure, this activity is
composed of three activities. First, a Measurement Planning Item is selected (Select
Measurement Planning Item for Sampling). Thus, samples are collected (Collect
Sample) and, if necessary, the Sample Preparation Performer prepares them (Prepare
Sample) according to the sample planning defined in the Measurement Planning Item.
Collect Sample refers to obtain a sample from which data will be collected (e.g., obtain
a blood sample of a person aiming to collect data to measure her cholesterol). Prepare
Sample consists of performing specific procedures to make a sample ready for
measurement. For example, it may be needed to freeze a sample before collecting data
from it. The main result of this activity are Samples, when it is necessary only to collect
samples, or Prepared Samples, when it is also necessary to prepare them.
Following Perform Sampling (or Plan the Measurement Process, if samples
are not required), the Measurement Performer must Perform Measurement, which
consists in, first, Selecting Measurement Planning Item for Measurement. Next, if
data will be collected from a sample, the Measurement Performer Selects Sample for
Measurement. Finally, she Collects Data, according to the measurement plan contained
in the selected Measurement Planning Item. The main result of this activity is a set of
Measured Values, i.e., the collected data.
Once data is collected, the Measurement Analysis Performer Analyzes
Measurements. Like Perform Sampling and Perform Measurement, this activity is
decomposed in others. Its main purpose is to provide information to get conclusions and
support decision making. Therefore, analysis should be carried out driven by goals.
Hence, its first activity is Identify Goal for Analysis, when the goal to which information
will be provided is identified. Thus, it is necessary to Select Measures for Analysis,
when the measures able to provide information related to the goal are selected, and Select
Data for Analysis, when data collected for the selected measures (i.e., measured values)
are selected. Finally, the Measurement Analysis Performer Analyzes Measurements,
when data regarding the selected measures are analyzed, producing Analysis Results.
After analyzing data, the Measurement Analysis Performer may need further analysis,
which can involve the need to: (i) analyze the same dataset using different analysis
procedures, (ii) include other measures and data in the analysis or (iii) establish new goals
to be analyzed. All these situations are addressed by returning to different activities of the
Measurement Process. At the end, the Analysis Results are reported by the Report
Analysis Results Performer in the Report Analysis Results activity.
3.2 MTO Structural Model
The structural model of MTO is an extension of COM [Barcellos et al., 2014] by
including mainly concepts to address aspects related to sampling, which were not
considered in the first version of COM. Figure 3 presents a fragment of the structural
model of the MTO, focusing on more central concepts. After the figure, we briefly
describe the concepts. In the text, the concepts are written in bold in the first time they
are cited. Concepts from UFO that are used to ground MTO concepts are written in italics.
Deeper discussions about the use of UFO to ground some concepts can be found in
[Barcellos et al., 2014].
Figure 3. Fragment of the Structural Model of MTO
A Measurable Entity is anything that can be measured, such as a person, a river
or an artifact. Given its very general nature, Measurable Entity corresponds to Individual
in UFO. Measurable Entity is instance of Measurable Entity Type (e.g., Person, River,
Artifact), a First Order Universal in UFO. Measurable Entity Types are characterized by
at least one Measurable Quality Universal, here so-called Measurable Element (e.g.,
Person can be characterized by weight, high, blood pressure, etc.). Measures are used to
quantify Measurable Elements and to characterize Measurable Entity Types. For instance,
the measure weight in kilograms can be used to quantify the measurable element weight
of measurable entities of the type Person. Measure is a Function in UFO in the sense that
it maps an instance of Measurable Element to a value (the Measured Value). Measures
have Scales (Measurement Reference Structure) composed of Scale Values
(Measurement Reference Region), which are the possible values to be associated by the
measure to a measurable element. For example, weight in kilograms has a scale composed
of positive real numbers.
As explained in the behavioral model description, Measurement Planning Item
specifies information about the planning of each measurement-related activity (the goal
to be achieved, the information needs to be met, the measurable entity type and its
measurable element to be measured, the measure to be used, the procedures to be adopted
in each measurement-related activity, the device types (or devices) to be used in them,
and the responsible for each activity). For sake of better visualization, from these
concepts, in Figure 2, we show only Measure, Measurable Entity Type and Measurable
Element.
Activities addressed in the behavioral model (namely, Sampling, Sample
Preparation, Measurement and Measurement Analysis) are represented as Events in the
structural model. The colors used in the concepts representing these events in the
structural model (Figure 3) are the same used in the activities in the behavioral model
(Figure 1), showing their correspondence. Sampling, Sample Preparation and
Measurement are performed based on a Measurement Planning Item. Sampling results
in a Sample, which is a Measurable Entity that after Sample Preparation is turned into
a set of Prepared Samples. A Sample is a proxy for a Measurable Entity, said a Sample
Represented Measurable Entity. In other words, a Sample represents a Sample
Represented Measurable Entity, which is the Measurable Entity characterized by the
sample. For example, a blood sample represents a person, since by measuring the blood
sample it is possible to characterize that person.
Measurement measured a Measurable Element of the Measured Entity (the
role played by the Measurable Entity when it was measured) and resulted in a Measured
Value. Finally, Measurement Analysis analyzed Measured Values and resulted in an
Analysis Result that characterizes the Measured Entity.
4. MTO Evaluation
To evaluate MTO, we used it to describe a real-world case scenario regarding a
Laboratory Test. In this section, we show how MTO can be used to represent a real case
of a HDL Cholesterol Test in which a doctor asked for a HDL Cholesterol Test of Ricardo
and used its results to make decisions about Ricardo’s health. In the following, we
describe the measurement process followed in this scenario. In the text, we use bold to
identify the performed activities of the measurement process and italics to identify their
performers. After the description, we present the objects (i.e., concepts instances)
involved in the process.
Ricardo’s doctor had the goal of “verifying Ricardo’s cardiological-related
conditions” and for that he needed to know, among others, Ricardo’s HDL Cholesterol
level. This started the measurement process. Based on that goal and information need, the
doctor asked for a test to measure HDL Cholesterol in mg/dL using the Bichromatic
Enzymatic method, which establishes the procedures to be adopted to collect sample,
prepare it, measure it and analyze collected data according to reference values. By doing
that, the Ricardo’s doctor (Measurement Process Planner) performed the Plan
Measurement Process activity.
After the medical appointment, Ricardo went to the laboratory Lab A and a
nursing technician (Sampling Performer) Performed Sampling (more specifically, she
Collected Sample) by collecting his blood sample using the established procedure. Later,
a biomedical technician (Sample Preparation Performer) Prepared Sample, by adopting
the established procedure to extract blood serum from the blood sample. After that,
another biomedical technician (Measurement Performer) Performed Measurement by
collecting data about Ricardo’s cholesterol from the blood serum, also according to the
established procedure. The laboratory’s responsible doctor (Measurement Analysis
Performer) evaluated the collected data and compared them with reference values (i.e.,
she performed the Analyze Measurement activity). Finally, the Lab A (Report Analysis
Results Performer) Reported Analysis Results through a document handed in to
Ricardo. Ricardo showed the test results to his doctor, who got information about
Ricardo’s HDL Cholesterol level and made decisions about Ricardo’s health. Figure 4
shows a fragment of Ricardo’s cholesterol test. Next, Table1 presents the instantiation of
MTO’s structural model, focusing on the objects involved in the described scenario.
Measurement-related activities and performers are not shown in the table because they
were indicated in the text.
Figure 4. Fragment of the Ricardo’s HDL Cholesterol Test
Table 1. MTO Instantiation
MTO Laboratory Test Scenario
Measurement Goal Verifying Ricardo’s cardiological-related conditions
Information Need What is Ricardo’s HDL Cholesterol level?
Measurable Entity Type Person; Blood
Measurable Entity Ricardo;Ricardo’s Blood Serum
Measurable Element HDL Cholesterol
Measure HDL Cholesterol in mg/dL
Scale Scale made up of positive integer numbers
Scale Value Positive Integer Numbers
Table 1. MTO Instantiation (cont.)
MTO Laboratory Test Scenario
Sampling/ Sample Preparation / Measurement/ Bichromatic Enzymatic method
Measurement Analysis Procedure
Measurable Planning Item Combination of information presented in the previous
concepts
Sample Ricardo’s Blood Sample
Prepared Sample Ricardo’s Blood Serum Sample
Sample Represented Measurable Entity Ricardo
Measured Value 51 mg/dL
Measured Entity Ricardo; Ricardo’s Blood Serum after measurement
Analysis Result Acceptable
5. Related Works
In the literature there are works proposing ontologies and conceptual models to the
measurement domain. For example, the TOVE Measurement Ontology (TMO) [Kim et
al., 2007] is a measurement core ontology for Semantic Web applications. TMO
addresses concepts related to: (i) measurement system, which deals with attributes that
can be measured, samples, and quality requirements; (ii) measurement activities, which
deals with data collection, inspection and test; and (iii) measurement points, addressing
measured values and their conformance to the quality requirements. There is some
equivalence between TMO and MTO concepts (e.g., measure and measurable element),
although different terms are used in some cases. However, TMO does not address some
aspects covered by MTO, such as measurement goal, scale, procedures, among others.
Moreover, TMO does not explore some relations between concepts. For instance, in TMO
there is no relation between measure and measured attribute (equivalent to Measurable
Element in MTO).
ISO 19156 (2011) defines a conceptual schema for observations and
measurements. It focuses on measurement in the context of environmental sampling,
defining a common set of sampling feature types classified primarily by topological
dimension, as well as samples for observations away from its natural surroundings. Like
MTO, it defines sampling-related concepts such as Sampling Method (Sampling
Procedure in MTO), Sampled Feature (Sample in MTO) and Sampling Time. Even
though the last concept is not explicitly mentioned in MTO models, it is also covered
because Sampling is an event in UFO, thus it also brings information regarding time
within it. However, although ISO 19156 addresses sampling aspects, it does not properly
cover other core aspects of measurement and it is applied to a specific domain.
Olsina and Martin (2003) proposed an ontology for software metrics and
indicators. Although focusing on software measurement aspects, the ontology is quite
general and includes some core concepts defined in MTO (sometimes using different
terms) such as Measurement, Measurable Entity, Measurable Element and Measure.
Later, Becker et al. (2015) used a generic process ontology to semantically enrich the
terms of the Olsina and Martin’s ontology by means of stereotypes. As a result, Becker
et al. categorized measurement-related concepts in the process context. For example,
Measurement was categorized as Task and Metric (equivalent to Measure in MTO) as
Method. However, this is not enough to describe the behavioral aspect of the
measurement process.
In summary, apart from TMO and MTO, none of the cited works is concerned
with a comprehensive and common conceptualization about measurement. Moreover, the
cited works focus only on structural aspects, while MTO addresses both structural
(domain) and task-related (behavioral) aspects of the measurement process.
6. Final Considerations
Measurement occurs in many application domains. There are various standards and
references about measurement and in some of them it is possible to identify common
knowledge, although the terminology used is distinct. Since ontologies are acknowledged
as being quite appropriate to solve conceptual ambiguities and knowledge vagueness, we
proposed MTO (Measurement Task Ontology) aiming to provide a conceptualization to
enable to reach a semantic agreement in a broader sense, i.e., by achieving a common
understanding regarding both domain (structural) and task-related (behavioral) aspects of
measurement.
Being a task ontology, MTO defines behavioral and structural models. The
structural model is an extension of the Core Ontology on Measurement (COM) [Barcellos
et al., 2014]. Since COM does not cover aspects that are common in more complex
measurements, such as sampling and measurement analysis through successive data
analysis and also does not address behavioral aspects describing the measurement
process, there was a need to represent the measurement process under a behavioral
perspective and to cover concepts involved in this process and not addressed in COM
[Barcellos et al., 2014].
MTO has been evaluated through instantiation of real-world case scenarios. In
this paper, we showed the instantiation involving a HDL cholesterol test. The scenario
was properly instantiated, providing initial evidence that the ontology is able to represent
real word situations.
MTO provides a conceptualization about measurement that can be used in several
domains. Moreover, MTO can be specialized to deal with particularities of measurement
applied to specific domains. MTO can be used for knowledge workers and can also serve
as a reference model to solve interoperability issues, such as standards harmonization and
systems integration. As future work, we intend to use MTO as a reference model to
integrate data from different sources in an Environmental Quality Research Project.
Acknowledgment
This research is funded by the Brazilian Research Funding Agency CNPq 407235/2017-
5, CAPES (23038.028816/2016-41), and FAPES (69382549/2014).
References
Barcellos, M.P., Falbo R. A. and Frauches, V. "Towards a Measurement Ontology Pattern
Language." ONTO. COM/ODISE@ FOIS. 2014.
Barcellos, M. P. and Falbo, R. A. "A software measurement task ontology." Proceedings
of the 28th Annual ACM Symposium on Applied Computing. ACM, 2013.
Becker, P., Papa, F., and Olsina, L. (2015). Process ontology specification for enhancing
the process compliance of a measurement and evaluation strategy. CLEI Electronic
Journal, 18(1), 3-3.
Cox, S. J. D., Simons B. A. and Yu J., "A harmonized vocabulary for water quality."
(2014).
Gkoutos, G. V., Schofield P. N. and Hoehndorf R. "The Units Ontology: a tool for
integrating units of measurement in science." Database 2012 (2012).
Guarino, N. “Formal Ontology and Information Systems” in: Formal Ontologies in
Information Systems, IOS Press, 1998, 3 -15.
Guizzardi, G. “On Ontology, ontologies, Conceptualizations, Modeling Languages, and
(Meta)Models, In Frontiers in Artificial Intelligence and Applications, Databases and
Information Systems IV, IOS Press, Amsterdam, 2007
Guizzardi, G. “Ontological Foundations for Structural Conceptual Models”, Universal
Press, The Netherlands, 2005.
Guizzardi, G., Falbo, R. A. and Guizzardi, R.S.S.” Grounding software domain ontologies
in the Unified Foundational Ontology (UFO): the case of the ODE software process
ontology.” In: Proceedings of the XI Iberoamerican Workshop on Requirements
Engineering and Software Environments (Brazil, 2008), 244-251.
ISO, TC. "211 (2011). ISO 19156: 2011-Geographic information--Observations and
measurements-International Standard. Geneva, Switzerland." International
Organization for Standardization.
ISO/IEC:ISO/IEC15939(E) Software Engineering - Software Measurement Process.
(2007)
Jasper, R. and Uschold M. "A framework for understanding and classifying ontology
applications." Proceedings 12th Int. Workshop on Knowledge Acquisition, Modelling,
and Management KAW. Vol. 99. 1999.
JCGM: VIM - International Vocabulary of Metrology – Basic and General Concepts and
Associated Terms. Joint Committee for Guides in Metrology, France (2012)
Kim, H.M., Sengupta, A., Fox, M.S., Dalkilic, M., “A Measurement Ontology
Generalizable for Emerging Domain Applications on the Semantic Web.” Journal of
Database Management 18, 20-42 (2007)
Olsina, L. and Martin, M. D. L. A. (2004). Ontology for software metrics and indicators.
J. Web Eng., 2(4), 262-281.
Martins, A. F. and Falbo, R. A. “Models for representing task ontologies”. In: Proceedings
of the 3rd Workshop on Ontologies and Their Applications, Brazil, 2008.
Pedhazur, E. J. and Schmelkin, L. P. (1991). “Measurement, Design, and Analysis: An
Integrated Approach” (1st ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. pp. 15–
29Wache, Holger, et al. "Ontology-Based Integration of Information-A Survey of
Existing
Scherp, A. et al. "Designing core ontologies." Applied Ontology 6.3 (2011): 177-221.
Wache, H., Voegele, T., Visser, U., Stuckenschmidt, H., Schuster, G., Neumann, H., and
Hübner, S. (2001, August). Ontology-Based Integration of Information-A Survey of
Existing Approaches. In OIS@ IJCAI.