Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities
The paper is devoted to defining the scope of research activities aimed at involving business stakeholders in a software process in a form of assessing the perceived quality of the service-oriented system in its usage context when the initial specification of the system is available in natural langu...
Збережено в:
| Дата: | 2010 |
|---|---|
| Автори: | , |
| Формат: | Стаття |
| Мова: | English |
| Опубліковано: |
Інститут проблем штучного інтелекту МОН України та НАН України
2010
|
| Назва видання: | Штучний інтелект |
| Теми: | |
| Онлайн доступ: | http://dspace.nbuv.gov.ua/handle/123456789/56270 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities / V.A. Shekhovtsov, N.A. Bazhenov // Штучний інтелект. — 2010. — № 3. — С. 161-169. — Бібліогр.: 12 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-56270 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-562702014-02-16T03:13:13Z Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities Shekhovtsov, V.A. Bazhenov, N.A. Интеллектуальные речевые технологии. Компьютерная обработка естественно-языковых текстов и семантический поиск The paper is devoted to defining the scope of research activities aimed at involving business stakeholders in a software process in a form of assessing the perceived quality of the service-oriented system in its usage context when the initial specification of the system is available in natural language form. We propose to use NLP techniques to extract the scope from this specification and to represent it in the format of specific predesign models compatible with the rest of the simulation solution. Стаття присвячена визначенню області проведення досліджень, пов’язаних з підключенням зацікавлених осіб до процесу розробки програмного забезпечення через оцінювання сприйманої якості сервіс-оріентованих систем в контексті їхнього використання, коли початкова специфікація системи задана природною мовою. Пропонується використання технології аналізу природної мови для отримання інформації про область застосування з цієї специфікації у форматі спеціальних моделей предпроектування, які є сумісними з основними модулями імітаційного рішення. Статья посвящена определению области проведения исследований, связанных с подключением заинтересованных лиц к процессу разработки программного обеспечения через оценивание воспринимаемого качества сервис-ориентированных систем в контексте их использования, в случае, если начальная спецификация системы задана естественным языком. Предлагается использование технологии анализа естественного языка для получения информации об области применения из этой спецификации в формате специальных моделей предпроектирования, которые являются совместимыми с основными модулями имитационного решения. 2010 Article Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities / V.A. Shekhovtsov, N.A. Bazhenov // Штучний інтелект. — 2010. — № 3. — С. 161-169. — Бібліогр.: 12 назв. — англ. 1561-5359 http://dspace.nbuv.gov.ua/handle/123456789/56270 681.03 en Штучний інтелект Інститут проблем штучного інтелекту МОН України та НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| topic |
Интеллектуальные речевые технологии. Компьютерная обработка естественно-языковых текстов и семантический поиск Интеллектуальные речевые технологии. Компьютерная обработка естественно-языковых текстов и семантический поиск |
| spellingShingle |
Интеллектуальные речевые технологии. Компьютерная обработка естественно-языковых текстов и семантический поиск Интеллектуальные речевые технологии. Компьютерная обработка естественно-языковых текстов и семантический поиск Shekhovtsov, V.A. Bazhenov, N.A. Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities Штучний інтелект |
| description |
The paper is devoted to defining the scope of research activities aimed at involving business stakeholders in a software process in a form of assessing the perceived quality of the service-oriented system in its usage context when the initial specification of the system is available in natural language form. We propose to use NLP techniques to extract the scope from this specification and to represent it in the format of specific predesign models compatible with the rest of the simulation solution. |
| format |
Article |
| author |
Shekhovtsov, V.A. Bazhenov, N.A. |
| author_facet |
Shekhovtsov, V.A. Bazhenov, N.A. |
| author_sort |
Shekhovtsov, V.A. |
| title |
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities |
| title_short |
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities |
| title_full |
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities |
| title_fullStr |
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities |
| title_full_unstemmed |
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities |
| title_sort |
using nlp to define the scope for stakeholder assessment of simulated service qualities |
| publisher |
Інститут проблем штучного інтелекту МОН України та НАН України |
| publishDate |
2010 |
| topic_facet |
Интеллектуальные речевые технологии. Компьютерная обработка естественно-языковых текстов и семантический поиск |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/56270 |
| citation_txt |
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities / V.A. Shekhovtsov, N.A. Bazhenov // Штучний інтелект. — 2010. — № 3. — С. 161-169. — Бібліогр.: 12 назв. — англ. |
| series |
Штучний інтелект |
| work_keys_str_mv |
AT shekhovtsovva usingnlptodefinethescopeforstakeholderassessmentofsimulatedservicequalities AT bazhenovna usingnlptodefinethescopeforstakeholderassessmentofsimulatedservicequalities |
| first_indexed |
2025-07-05T07:32:52Z |
| last_indexed |
2025-07-05T07:32:52Z |
| _version_ |
1836791389982031872 |
| fulltext |
«Штучний інтелект» 3’2010 161
3S
UDC 681.03
V.A. Shekhovtsov, N.A. Bazhenov
National Technical University “Kharkov Polytechnic Institute”
shekvl@yahoo.com, nabazhenov@gmail.com
Using NLP to Define the Scope for Stakeholder
Assessment of Simulated Service Qualities
The paper is devoted to defining the scope of research activities aimed at involving business stakeholders in
a software process in a form of assessing the perceived quality of the service-oriented system in its usage
context when the initial specification of the system is available in natural language form. We propose to use
NLP techniques to extract the scope from this specification and to represent it in the format of specific
predesign models compatible with the rest of the simulation solution.
Introduction
Gathering the opinions of business stakeholders (not necessary possessing any IT
background) is considered an important part of the software process by both software
engineering researchers and practitioners. The main subject of such opinions is currently the
prospective system’s functionality; it forms the foundations for widely studied functional
requirements elicitation problem. Less studied but still important kind of opinions are
assessments of the quality of the prospective system. The reason of this importance is that if
stakeholders cannot express such opinions early in the software development lifecycle these
they could be easily lost. As a result of this loss, the understanding of the desired quality of
the system becomes biased towards the view of the IT people. This could lead to the
stakeholder dissatisfaction with the quality of the software under development (SUD) late in
the development lifecycle. This is dangerous, as the problems with satisfying the customer
revealed that late could easily lead to the complete failure of the whole project.
To address this problem, we proposed ISAREAD-S framework (Interactive Simu-
lation-Aided Requirements Engineering and Architectural Design for Services) [1], [2]. It is
aimed at investigating ways of supporting stakeholder involvement in the software process
by allowing business stakeholders to assess the perceived qualities of the prospective system
in its usage context. To implement such support we plan to elaborate a set of simulation-
based methods aimed at making QoS (quality of service) assessment procedures accessible to
the business stakeholders and using their assessments as a driving force for activities related
to requirements engineering and architectural design.
The research goal is to define a scope for these simulation activities (ISAREAD-S
scope), i.e. the initial set of services and qualities of interest together with possible usage
contexts. We propose to obtain this scope by analyzing the existing description of the problem
domain by means of Natural Language Processing (NLP) techniques.
The structure of the paper is as follows. Section 1 shows the principles of the existing
procedures (mechanisms) for organizing the interaction with stakeholders; these mechanisms
form the target activities for the scope definition solutions proposed in the paper, Section 2 for-
mulates the problem statement, Section 3 describes the models used to define the ISAREAD-S
scope, Section 4 outlines the proposed NLP-based approach for transferring the specification
for the SUD or the problem domain into the ISAREAD-S scope definition, and finally sec-
tion with necessary conclusions.
Shekhovtsov V.A., Bazhenov N.A.
«Искусственный интеллект» 3’2010 162
3S
1. Interactive assessment of simulated service qualities
ISAREAD-S approach includes three kinds of mechanisms.
Service-level mechanisms organize assessments for the simulated qualities of the
particular services. We define two mechanisms of this kind: model composition and model
execution mechanism. Model composition mechanism combines parameterised simulation
models with corresponding interaction models for the particular qualities to form interac-
tive simulation models allowing business stakeholders to participate. The participation
depends on the quality being processed; it involves experiencing the quality and making an
assessment using specific scale. Model execution mechanism is based on model driven
paradigm as it treats interactive simulation model as executable model; it is responsible for
interactive execution of such models to organize assessments of the qualities of interest.
Process-level mechanisms put these service-level assessments into the context of simu-
lated usage processes. We propose to represent such usage processes using Business Process
Modeling (BPM) notations such as BPMN or Petri Nets. Similarly to the service-level case,
process-level model composition and model execution mechanisms are defined. The first me-
chanism embeds service-level interactive simulation models for the services of interest into the
business process model to form interactive simulation model for service usage process. Model
execution mechanism executes this model to establish a user session. In the course of this ses-
sion, the user is asked to perform assessments of the qualities of interest for the services of interest.
An iterator mechanism controls the interaction sessions and gathers all necessary
assessments for the particular SUD (treated as a set of services of interest).
On top of these mechanisms the policies for requirements elicitation, negotiation, and
validation, architecture assessment and lifecycle support are defined. In addition, the
adapters are responsible for the transformation of the external information related to the
project (such as the amount of available resources) into the numerical values for the factors
influencing the simulations.
2. The problem of ISAREAD-S scope definition
The above description of the ISAREAD-S framework misses one important issue: it
is not clear how to define the set of services and qualities of interest and possible usage
context (which form the scope of the future simulation-related activities) prior to the asses-
sment. It was implicitly assumed that this set is known beforehand to software engineers.
Unfortunately, it is not always the case; as a result, defining such set by hand every time
the new project begins could be cost-ineffective.
The ways of handling this problem vary significantly depending on availability of the
description of the prospective system (or at least the problem domain). In this work, we
restrict ourselves with the case when such description is available in natural language form,
which leads to the following research question which determines the problem statement:
How to determine the set of services and qualities of interest together with possible usage
contexts from the existing natural language specification of the prospective SUD and make
it available to the ISAREAD-S framework?
To address this question, we propose to use Natural Language Processing techniques
to process the specification documents with a goal of obtaining structural representation of
the scope of the ISAREAD-S application (Fig. 1). This way, such process could be auto-
mated which reduces the costs of applying the framework.
3. Representing ISAREAD-S scope with predesign models
Before investigating possible ways to process the description of the prospective system or the
problem domain we need to define the target of this processing, namely, the format we are plan-
ning to use to represent the scope of the ISAREAD-S framework to be used by its components.
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities
«Штучний інтелект» 3’2010 163
3S
Figure 1 – Problem definition
As our framework focuses on supporting stakeholder involvement into the software
process activities we should implement the interface which, on the one hand, is suitable for
validation by business stakeholders (i.e. the people without special IT-skills) and, on the other
hand, is suitable for implementing the support for the simulation of service qualities. We use
special requirements models for this purpose. These models propose to the user a tabular
form of communication which is understandable for domain experts.
To implement user-centered view of the SUD functionality (functional ISAREAD-S
scope), i.e. the set of services of interest, we use Klagenfurt Conceptual Predesign approach
developed in Klagenfurt University [3], [4] which defines a metamodel (Fig. 2) for Klagen-
furt Conceptual Predesign Model (KCPM). This model consists of static and dynamic parts,
which include process-invariant and service process information respectively. Thing-type and
connection-type form the static part. Thing-type generalizes the class notion as well as
attribute and value notions. The examples of this concept are persons, objects, resources, at-
tributes, characteristics, abstracts and other entities. Connection-types represent relations bet-
ween thing-types. Operation-type and cooperation-type constitute the dynamic part. Ope-
ration-types are intended to define service operations, their actors and parameters, which are
expressed by thing-types. Cooperation-types are used to model business processes which
orchestrate services. A cooperation-type consists of a triple of sets <Prc, {A,O}, Poc>, where
Prc – the set of pre conditions, Poc – the set of post conditions, {A,O} – the set of pairs con-
sisting of an operation-type and an actor executing this operation.
Figure 2 – KCPM metamodel for a functional ISAREAD-S scope
To implement user-centered view at the qualities of the prospective SUD (non-
functional (quality-related) ISAREAD-S scope), i.e. the set of qualities of interest, we use
Quality-Aware Predesign Model for Services (QAPM-S) [5]. It is connected to KCPM (fun-
ctional model) via common abstract concept ModelingElement and dynamic elements –
operation-type and cooperation type (Fig. 3). This model is based on the notions defined in
ISO/IEC 9126 standard such as quality category, quality characteristic, quality metric, etc.
Shekhovtsov V.A., Bazhenov N.A.
«Искусственный интеллект» 3’2010 164
3S
Figure 3 – QAPM-S metamodel for a quality-related ISAREAD-S scope
4. Using NLP techniques for ISAREAD-S scope definition
Morphosyntactic and semantic analysis. To obtain structural representation of
ISAREAD-S scope we need to bridge the gap between natural language knowledge data
generated by stakeholders and structured modeling data used by designers and developers.
These acts are being solved using different NLP methods [6]. For the moment we use one
of the approaches based on context-free grammars and Chomsky generative syntax. More
detail information is described in previous paper [7]. This approach combines probabilistic
part-of-speech tagging with sophisticated rule-based chunking and produces from free
requirements text structured ontology-oriented tree-based output, available in XML format [4].
Its main benefits are:
− enriching classical part-of-speech (POS) tagging with additional lexical information;
− compound nouns identification related to term identification;
− two-level verb subclass identification (firstly taking into account the number of
noun pseudo-objects (nPO); secondly after chunking procedure correction according to the
verb arguments number);
− post-modifier of nouns;
− flexible settings allowing to manipulate with different lexical features (phrasal
nodes, etc.);
− verb roles disambiguation.
But in spite of these advantages there are some difficulties with the processing of
complex language constructions and English language dependency. To process other lan-
guages the developed engine requires the appropriate language resources both the language
data itself and its structure rules information. For increasing accuracy of text processing
and targeting an extended set of human languages more flexible and semantic-oriented
approaches (like MTT [8], or UNL [9]) can be embedded in ISAREAD-S framework.
The example of this requirements elicitation step is listed below:
- <sentence>
- <sentence type="subordinate">
<con0 derivedPOS="n0" type="subord">If</con0>
- <n3>
<det0 form="general" type="def">the</det0>
<n0 base-form="order" derivedPOS="v0" num="sg" type="common">order</n0>
</n3>
<v0 base-form="come" form="ind" num="sg" phrasalverb="come in" ps="3"
temp="pres" verbclass="iV">comes</v0>
<p0 derivedPOS="pt0" phrasalverb="come in">in</p0>
</sentence>
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities
«Штучний інтелект» 3’2010 165
3S
- <n3>
<det0 form="general" type="def">the</det0>
- <n0 desc="compound" type="common">
<n0 base-form="bookkeeping" corelex="act" derivedPOS="a0" num="sg"
type="common">bookkeeping</n0>
<n0 base-form="department" corelex="grs" num="sg"
type="common">department</n0>
</n0>
</n3>
<v0 base-form="check" derivedPOS="n0" form="ind" temp="pres" num="sg" ps="3"
verbclass="tvag2">checks</v0>
- <n3>
<det0 form="general" type="def">the</det0>
<n0 base-form="payment" corelex="poa" num="sg" type="common">payment</n0>
</n3>
</sentence>
The requirements sentence “If the order comes in, the bookkeeping department checks
the payment” was transformed into structured tree-based head-lexicalized format. The second
part of sentence is governed by verb “check” which has two arguments: subject – noun
phrase “the bookkeeping department” and object – noun phrase “the payment”. This XML
material is ready for interpretation and modeling notions extraction.
Interpretation. After free text processing the obtaining output is ready to be inter-
preted into different modeling concepts and parameters. For modeling context-aware
functional service aspects the user-centered approach is used [3]. As this approach’s pivot
is the stakeholder involvement in development process, its main view is the set of
glossaries (organized into QAPM-S model [5]), which are filled with modeling concept
from processed text. For this purpose we use various interpretation rules, mainly based on
predicate argument structure of verbs and its agentivity. The list of verb classes identified
during the linguistic processing and used for interpretation can be found in [7]. The head of
noun phrases and the parts of compound nouns are transferred into thing-type glossary. The
containing and attributive relations create entries in connection-type glossary. The
transitivity of verbs flags to what element of behavior model part it belongs to. For
example, verbs with categories “tvag2” (monotransitive verb with agent subject), “tv3”
(ditransitive verb), “iV” (intransitive verb) [1] denote an operation-type.
The set of such operation types correspond to the set of services of interest
comprising the ISAREAD-S scope. The noun on subject place is mapped to an executing
actor, whereas on object place to a calling actor or parameter depending on context and
other trigger rules. The verb without agent subject, e.g. “tv2”, denotes the condition in the
cooperation-type glossary [10].
Quality characteristics being modeled are structured into the QAPM-S model and can
also be extracted from the NLP output. The quantor phase contains the quantitative quality
characteristics in this model. The service usage context represented by operation-type and
cooperation type used from its original glossaries is also interpreted using sentence and
phrase relationships.
The main rules of interpretation are listed in tables 1 – 2. The example of interpre-
tation engine applying for the mapping from tree-based structured requirements into
glossaries’ entries is represented in fig. 4. The engine process the requirements sentence
applying step by step the interpretation rules climbing up the rule order.
Consider more detail the list of rules used to form functional requirements model. Rule 1
reflects the basic principle that every noun is thing or abstract notion and therefore mapped
Shekhovtsov V.A., Bazhenov N.A.
«Искусственный интеллект» 3’2010 166
3S
into thing-type entry. The attribute assigned during noun type identification can be trans-
formed to “classification” field in thing type. Rule 2 means that in the case of compound, i.e.
congruence between parts of compound concerning proper or common type. Rule 3 denotes
the relation between the head and dependent member in compound noun structure. Rule 4
identifies the post-modifier in the case of prepositional phrase (p2) [11]. “p2” can play role
both as the argument of verbal phrase and as adjunctive or adverbial modifier of other
sentence argument. This ambiguity is tried to solve by developed approach [6]. And in the
case of modifier it should reproduce the entry in connection type glossary. Rule 5 captures the
abstraction generalization and component/object which are treated as connection types [12].
These cases occur if the verb “to be” in 3rd person functions as the main verb in the sentence.
Rule 6 is intended for possession relation between subject and object. It occurs then the verb
“to have” functions as the main verb. Other default cases for the verb with more than one
argument are applied by Rule 7. These cases present the relation between verb arguments
with respect to its valency. Rule 8 provides the operation-type identification. The noun phra-
se, which is agent subject, becomes an “actor” whereas other noun phrases, which have other
argument roles, are mapped into parameters of operation-type. Rule 9 is executed in the case
of past participle of transitive and intransitive verbs and denotes the pre-condition in
cooperation-type glossary. Rule 10 specifies the condition for operation-type if the verb is
not an agentive verb and all the arguments of the verb become candidates for involved thing-
types (according to the procedure described in [11]). Rule 11 indicates the conditions of
property/state type; it means then adjective or adverbial phrase describe the properties of
thing-type. Rule 12 decodes the event type of condition; it means then act is carried out, but
the subject does not act. Rule 13 allows us to suppose the passive voice construction as the
candidates for conditions of cooperation-types. Rule 14 emphasizes the if/when constructions
as the basic valid sentence patters of cooperation-type.
Consider the list of rules for quality requirements. Rule 1 shows how the quantitative
value of quality characteristic can be handled. We presuppose that quantor modifier of
noun phrase can serve such value. The subject in that sentence becomes the QualityMetric
element. Rule 2 specifies the qualitative parameters of quality requirement. This part of in-
vestigations is still in progress and requires the additional intelligence methods to distin-
guish elements of QAPM-S among the whole variety of modeling concepts, such as ap-
plying specified ontology patterns, which are able to detect the traditional quality
characteristics (e.g. defined in ISO standards: availability, accessibility, performance, etc.),
or word class identification via WordNet-based systems, which could provide the assign-
ment of terms to needed cluster.
Table 1 – Part of interpretation rules for functional requirements model
№ Rule Description Glossary Example
1. n0 → thing type
n0.corelex →
thing type.classificatio
n
Thing type Order
2. n0(desc=compound) → thing
type Compound noun Thing type Order processing
service
3. n0(desc=compound) → attribute Attribute Connection
type Order item
4. n3, n2.child=p2 → post-modifier Attribute Connection
type
The man with
the hat
5. v0(verbclass=copV) →
generalization/aggregation
Abstract
generalization “is_a”
or component
“is_part_of”
Connection
type A truck is a car
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities
«Штучний інтелект» 3’2010 167
3S
Table 1 (Cont.)
№ Rule Description Glossary Example
6. v0(verbclass=possV) →
possession Possession Connection
type
Each hard drive
has a capacity
7.
v0(verbclass=psychV | tvag2 |
locV | tv3 | sentV | tv2) →
argument relation
Relation between
verb arguments with
respect to its valency
Connection
type
The order
department
relates the item
to the order
8.
v0(verbclass=iV | tvag2 | tv3 |
sentV | psychV) → operation
type; n3subject →actor, n3object |
p2 | sentence(type=inf)
→parameters
Activity/Action is
executed by agent-
subject.
Operation
type
The order
department for
each ordered
item checks its
availability on
stock
9.
v0(verbclass=iV | tvag2 | tv3 |
sentV | psychV, temp=perf) →
condition; n3 | p2 |
sentence(type=inf)
→involvedTypes
Completion of
activity
Cooperation
type
If department
has checked the
order,…
10. v0(verbclass=tv2) → condition,
n3 →involvedTypes Post or precondition Cooperation
type
Payment is
needed
11. v0(verbclass=locV | possV |
copV) → condition Property/state Cooperation
type
All articles of
order are in
stock
12. v0(verbclass=eV) → condition Event Cooperation
type
The window
broke
13. v0(mode=pass) → condition Passive construction Cooperation
typt
The article is
ordered
14.
<con0> <n3> <v0> [<n3> |
<p2> | <sentence>], <adv2>
<n3> <v0> [<n3> | <p2> |
<sentence>]
If – then construction Cooperation
type
If each ordered
item is on
stock, then
order
department
relates that item
to the order.
Table 2 – Part of interpretation rules for quality requirements model
№ Rule Description Element Example
1.
<n3subject> <v0(verbclass=copV)>
<n3object>, n3object.child=q2 → n3subject
-QualityMetric,
QualityInContext.value=q2
The value of
quality parameter
Quality
Metric
The response
time to the stock
items
replenishment is
below 2
seconds.
2.
<n3subject> <v0.verbclass=copV> a2
→ n3subject -QualityMetric,
QualityInContext.description=a2
The characteristic
of quality
parameter
Quality
Metric
Accessibility is
highest possible
Shekhovtsov V.A., Bazhenov N.A.
«Искусственный интеллект» 3’2010 168
3S
Id# Name classification Quantity est. examples … Source
D1 Order processing services service 1 S1, S2
D2 Authorization services service 2 S3, S4
D3 Order thing 1000 S1, S8, S10, S13-S15
D4 Order date attribute 365 2008-01-04 …
D5 Order item thing 100000 1 ton of grain …
D6 Order department organization 1 …
D7 Stock clerk person 5 John Doe …
D8 Payment thing 1000 S3, S6, S13, S14,S17
D9 Bookkeeping department organization 1 …
c-id# name …
Perspective requi-rements
sourcep-id# involved thing-type name
C001 containment p001a D3, order contains S1, S2, S6
p001b D5, order item is contained in
C002 attribute
possessing
p002a D3, Order has an attribute
p002b D4, order date is an attribute of
id# Name classification involved thing-types … requi-rements
sourceid# name type
O1 check items service
operation
D1 order processing service executing S1
D6 order department calling
D5 order item parameter
O4 relate items
to order
service
operation
D1 order processing service executing S2
D6 order department calling
D3 order parameter
D5 order item parameter
O7 authorize
payment
service
operation
D2 security service executing S7
D9 bookkeeping department calling
D8 payment parameter
O8 authenticate
user
service
operation
D1 order processing service executing S8
D2 security service calling
D10 user parameter
Id#
Pre-condition Operation Post-condition sour
ceid# Name involved
types id# name involved
types id# Name involved
types
E1 C1 order
comes in
D3, Order O7 check
payment
D9,
bookkeeping
department
D8, Payment
C2 payment is
authorized
D8,
Payment
S1,
S5
Alternate
C3 payment is
not
authorized
D8,
Payment
in parallel
O1 check all
items
D6, order
department
D5, Order
C4 all articles
are in stock
D3, Order S1,
S2
Alternate
C5 not all
articles are
in stock
D3, Order
E2 C3 payment is
not
authorized
D8,
Payment
O5 Reject order D9,
bookkeeping
department
D3, Order
C6 order is
rejected
D3, Order S8,
S16
Figure 4 – Interpretation process
Conclusions
As a result of applying the proposed technique, QAPM-S representation of the set of
services and qualities of interest (and, optionally, some of their usage contexts) could be
obtained. This representation serves as a scope for subsequent quality simulation and
assessment activities. Automating the task of defining the ISAREAD-S scope reduces the
Using NLP to Define the Scope for Stakeholder Assessment of Simulated Service Qualities
«Штучний інтелект» 3’2010 169
3S
up-front costs for applying ISAREAD-S framework. Reducing these costs can be conside-
red an important step in the direction of increasing the feasibility of its deployment.
Literature
1. Using Linguistic Knowledge for Fine-tuning Ontologies in the Context of Requirements Engineering /
J. Vöhringer, D. Gälle, G. Fliedl, C. Kop, M. Bazhenov // International Journal of Computational Lin-
guistics and Applications. Bahri Publications. – 2010. – Vol. 1. – P. 249-267.
2. Towards Simulation-Based Quality Requirements Elicitation: A Position Paper / R. Kaschek, C. Kop,
V.A. Shekhovtsov, H.C. Mayr // REFSQ 2008. – LNCS, Vol. 5025. – Springer. – 2008. – P. 135-140.
3. Kop C. Mapping Functional Requirements: From Natural Language to Conceptual Schemata / C. Kop,
H.C. Mayr // Proc. SEA'02. – 2002. – P. 82-87.
4. Mayr H.C. Conceptual Predesign – Bridging the Gap between Requirements and Conceptual Design /
H.C. Mayr, C. Kop // Proc. ICRE '98. – IEEE CS Press, 1998. – P. 90-100.
5. Shekhovtsov V.A. Capturing the semantics of quality requirements into an intermediate predesign model /
V.A. Shekhovtsov, C. Kop, H.C. Mayr // Proc. SIGSAND-EUROPE'2008 Symposium, Lecture Notes in
Informatics (LNI) P-129. – Bonn : GI-Edition, 2008. – P. 25-37.
6. Shekhovtsov V.A. Interactive assessment of simulated service qualities by business stakeholders: princip-
les and research issues / V.A. Shekhovtsov // Проблеми програмування. – 2010. – № 2-3. – С. 288-298.
7. Bazhenov N.A. Combining probabilistic tagging with rule-based multilevel chunking for requirements
elicitation / N.A. Bazhenov // Искусственный интеллект. – 2010. – № 2. – С. 6-14.
8. Мельчук И. Опыт теории лингвистических моделей Смысл-Текст / Мельчук И. – М. : Наука,1974.
9. UNDL Foundation [Электронный ресурс]. – Режим доступа : http://undl.org.
10. Extended Tagging and Interpretation Tools for Mapping Requirements Texts to Conceptual (Predesign)
Models / G. Fliedl, C. Kop, H.C. Mayr [et al.] // Proc. of 10th Int. Conf. on Applications of Natural
Language to Information Systems NLDB 2005. – Springer, Heidelberg Lecture Notes in Computer
Science, 2005. – Vol. 3515. – P. 173-180.
11. Deriving static and dynamic concepts from software requirements using sophisticated tagging / G. Fliedl,
C. Kop, H.C. Mayr [et al.] // Data & Knowledge Engineering, 2007. – Vol. 61. – P. 433-448.
12. Linguistically based requirements engineering – The NIBA-project / G. Fliedl, C. Kop, H.C. Mayr [et al.] // Data &
Knowledge Engineering, Special issue on NLDB '99: applications of natural language to information
systems, 2000. – Vol. 35. – P. 111-120.
В.А. Шеховцов, М.О. Баженов
Використання засобів обробки природної мови для визначення області застосування
користувацького оцінювання змодельованої якості обслуговування
Стаття присвячена визначенню області проведення досліджень, пов’язаних з підключенням зацікавлених
осіб до процесу розробки програмного забезпечення через оцінювання сприйманої якості сервіс-
оріентованих систем в контексті їхнього використання, коли початкова специфікація системи задана
природною мовою. Пропонується використання технології аналізу природної мови для отримання інформації
про область застосування з цієї специфікації у форматі спеціальних моделей предпроектування, які є
сумісними з основними модулями імітаційного рішення.
В.А. Шеховцов, Н.А. Баженов
Использование средств обработки естественного языка для определения области применения
пользовательского оценивания смоделированного качества обслуживания
Статья посвящена определению области проведения исследований, связанных с подключением
заинтересованных лиц к процессу разработки программного обеспечения через оценивание
воспринимаемого качества сервис-ориентированных систем в контексте их использования, в случае,
если начальная спецификация системы задана естественным языком. Предлагается использование
технологии анализа естественного языка для получения информации об области применения из этой
спецификации в формате специальных моделей предпроектирования, которые являются совместимыми с
основными модулями имитационного решения.
Статья поступила в редакцию 02.07.2010.
|