CC BY
8A CONCEPTUAL MODEL STRUCTURING AND REPRESENTING KNOWLEDGE
Pugachev A.A.
ORCID:0000-0002-0089-3609, PhD of technical science, head scientific-practical center of innovative solutions and systematic studies of the Baikal branch of Humanitarian Institute (Moscow)
Pugacheva A.A. 0RCID:0000-0001-5588-9212, student of bachelor of civil engineering faculty of East Siberian state University of technologies and management,
Ulan-Ude, Republic of Buryatia.
Pavlutskaya N. M. 0RCID:0000-0002-4105-2080, PhD in pedagogical Sciences, associate Professor of Physics Department East Siberian state University of technologies and management,
Ulan-Ude, Republic of Buryatia
ABSTRACT
The article reveals an algorithm of object-oriented analysis of knowledge (algorithm OOACBS) used in the design, development and operation of computer-based training systems (hereafter CBS), implying a disaggregation of the subject area in general into eight strata. The author considers the minimum set of strata allocations when designing e-learning courses which hosted on the educational portal of Baikal branch of Humanities Institute (Moscow), and is also used in the process of assessing the level of knowledge of students of the East Siberian State University of technologies and management (Ulan-Ude).
Keywords: object-structured analysis of knowledge, stratification of knowledge, diagnostic knowledge, logic with vector semantics, e-learning course.
Conception of continuous education is one of the effective tools for solving problems of conformity of specialists' qualification to the rapidly growing level of knowledge, abilities and skills that are required in the conditions of modern technological progress. It is obvious that only skilled workers will provide a high level of competitiveness of goods, services and will bring our country to the forefront of the world community. That is why the goal of improving the system of training and retraining in all sectors of industry and economy is becoming strategic today. One of the ways of solving this problem may become widespread introduction of new information and educational technologies designed to ensure the high quality and efficiency of education, including in distance form.
Today, the market of educational services presents a large number of distance learning systems (e.g. Moodle, eFront, ATutor, etc.). Raise of the efficiency of their use in the educational process may be achieved through a systematic approach to the design and development of training content, which shall include, in particular:
presentation of educational material in the form of knowledge bases and the realization of sampling algorithms
to get knowledge from it which need to solve specific didactic problems;
realization of models of learners in the subject area; adaptation of educational material, a navigation system and a learning system in accordance with the model of the learner;
flexible diagnostic capabilities not only of knowledge, abilities and skills, allowing them to be an adequate assessment and identify causes of problems associated with their acquisition and consolidation, but also the level of competences' formation.
One of the methods of complex solution of the above problems is proposed in this article. The basis of this method is a conceptual model for structuring knowledge based on the representation of different types of knowledge as objects of stratified information space [4-6]. Within the model the authors have developed a modified algorithm for object-structure analysis [4, 7] (OSActs algorithm) which is designed for detailed structuring of knowledge used in the design, development and operation of computer-based training systems (CTS), and implying disaggregation of the subject area in general on eight strata (Table. 1).
s1 WHY-knowledge (why do we teach) Strategic analysis: the purpose and functions of the system (training objectives and tasks of the CTS).
s2 WHO- knowledge (who teaches) Organizational Analysis: CTS team of developers.
s3 WHAT- knowledge (what to teach) Conceptual analysis: basic concepts, conceptual structure of the subject area (subject knowledge)
s4 WHOM- knowledge (whom to teach) Personal analysis: knowledge about the student (student model)
s5 HOW- knowledge (how to teach) Functional analysis: control of the learning process, the implementation of adaptation (didactic knowledge).
Table 1
Stratification of Knowledge
s6 WHY- knowledge(why it is taught in this way) Casual or Causal analysis: forward and reverse link (diagnostic knowledge).
s7 WHERE- knowledge (where to teach) Spatial analysis: environment, equipment, communication (additional sources of knowledge and skills, for example, other CTS, literature, software, etc.)
s8 WHEN- knowledge (when to teach) Time analysis: temporary settings and restrictions.
Strata isolated on the basis of analysis of the main didactic tasks solved with the help of the CTS [2]:
• initial familiarization with the software, the development of its basic notions and concepts;
• basic training at different levels of depth and specification;
• development of skills for solving typical practical problems in the software;
• development of skills of analysis and decision-making in non-standard (non-typical) problem situations;
• development of skills to certain types of activities;
• conducting of educational research experiments with models of the studied objects, processes and environment
activities;
• restoration of knowledge and skills (for rare situations, tasks and technological operations);
• Monitoring and evaluation of the levels of knowledge and skills.
The algorithm is divided into two stages: the "vertical" and "horizontal" analysis. At the stage of vertical analysis split of methodological knowledge strata is performed. The amount and composition of allocated strata at this stage depends on the type of system being developed. Then, the horizontal analysis of the costs, including the construction of multi-level structures of individual strata is performing (table 2).
Table 2
OSAKOS
Strata Levels
Level of the field u1 Level of the problem u2 Level of the task u3 Level of the subtask u4 n
Strategic analysis s1 E11 E12 E13 E14 E1n
Organizational analysis s2 E21
Conceptual analysis s3 E31
Personal analysis s4 E41
Functional analysis s5 E51
Casual analysis s6 E61
Spatial analysis s7 E71
Interim analysis s8 E81
Eij
sm Em1 Emn
The content and sequence of horizontal analysis' steps depends on the type of CTS and number of stratum, but it is actually reduced to the implementation of the dual concept of structuring relevant knowledge. The number n - levels depends on the current levels of stratum and determined by feature of the field of knowledge and / or features of the organization of the learning process.
Use of this approach allows you to:
• combine different ways of structuring knowledge conceptually, used for the design, development and operation of adaptive computer-based training systems;
• apply various methods of isolation and sequencing of knowledge, depending on the purpose and type of developed training system;
• focus on the various strata which are necessary for the development of specific training systems, paying tribute to the importance of other strata.
Depending on the type of CTS (computer simulator, e-learning course, which controls the program, an electronic
encyclopedia, etc.) a minimum set of strata may differ significantly.
In particular, a minimum set of strata in the e-learning courses, designed in the East Siberian State University of Technology (Ulan-Ude), and management and the Baikal branch of the Institute of Humanities (Moscow) includes the strata s3, s4, s5, s6. Let us consider them in details.
The problem of representation of subject knowledge [5] (WHAT-knowledge) is the most studied issue now.
Subject knowledge is a semantic network that reflects the hierarchy of domain concepts in a form appropriate to the professional presentation of creator's e-learning course and doesn't cause learner's cognitive dissonance. Subject knowledge should be presented in the form of so-called relevant heterostructure HS = {Anc, Rel}, where nodes are concepts of the subject area Аnc allocated as a "reference nodes". Connections or relationships Rel between concepts used for transitions between them. Interpretation of relations is performed by production rules related to the relevant concepts.
Мзмеримм урока
Методические рекомендации...
@ ♦Учебный материал ^
В*Рндеп[ Эеидинни в жон.,. ч, ЯВ^Теиа №1. Предмет и метод ... ■ фПмаНЭ. Рцноин» jkohouh^. ^ В»Раяцвп II. Микроэкономика ^ НН»Тема №. Спрос, предложен... ^ B*TtMl Ns4. Пов»д«ми« noTpti...
. Простейшая модель пове.., »Ï, Потибкчюилнн »»Sop „,
Потр^бнтельс кн н вы&оз.... ^ ÎTTeçiHM ^ 1В*Т«м»№5,Поаминм прошв». В0*тФма Mii. Рынок ресурсов и... ^
Метаданные материала
Наавани« Программа курса Ч; © Создатель: Андрей Длеюс андровнч Пугачев ■ Шт Q
Течж rtem данных "Т&лгэ~ ф
Олнсмиг нет двнны* ¿ли пом ■Олневчл" ф Издатель: Андрей Александрович Пугачев © Автор; Андрей Длександ ровчч Пугачев :л Q Дата: ЯМПЛЙК Ч; О Тип: content ^ Q Формат; dif-чыхс'лр norf* "Фи.рлн^т J ф
ИдИЛМфИиТОр NeirnJsuj.bii doff nttrtft 'иа^ятцфцятср ■ ф Источник Нет данных doff поля "Mcmwwuff " ф
Pic. 1. Presentation of subject knowledge
The analysis of approaches to the learner's modeling leads to the following conclusion: in most existing approaches the learner's model contains a limited set of options, and hard-coded algorithms of their processing, which reduces the possibility of adaptive systems based on these approaches.
Therefore, the learner's model (WHOM-knowledge) is proposed to set the expression:
МОд ={ МОФ, МОП} (1)
Here МОФ - trio, determined by the method laid down the basis of tools for the development of e-learning courses (in this case used as the SDO):
МОф= {G1 (t), G2, V}, (2)
where G1(t) - graph of overlay model of the student's current knowledge: the arcs - the names of the concepts of the subject area; terminal nodes - assessment of the level of students' knowledge of the corresponding concept (can be a Boolean variable (known / unknown), a qualitative assessment (excellent / good / satisfactory) and the quantitative value (for example, the probability of the fact that the student knows the
Тесты разделов и подразделов
concept));
G2 - the graph of learning objectives - reflects the necessary level of student's knowledge of the current subject area (its structure is similar to the graph G1, but the top - of the required assess of knowledge levels for each of the concepts);
V - vector of student's preferences: the way of representing the material (text, graphics, video), system settings (color, size, skins), etc.
The second component of the model МОП contains a system of parameters defined by the maker of a particular course. These parameters are determined by the needs and preferences of the maker, and may differ significantly from course to course. In general terms, it is the vector of numerical and linguistic variables, which extend and complement the built-in model МОФ (from the point of view of the developer).
Along with the learner's model, conceptual model includes didactic knowledge (such as knowledge) for learning process control.
Показать даччь-е по разделу: бее разделы
В
Тесты Вопросы
Добавить гест ИМЯ НС * ОД" Ы Й |>в з до л
Опу&ЛИКОвЭН Вопросы Средний &9ЛЛ
Тест N»1 Учебный материал * Раздел 1. Введение в экономику » Тема №1. Предмет и метод экономической теории © 10 £2.20 <*>
Тест N»2. Учебный материал » Раздел 1. Введение в зкенаынку » Тема №2. Рыночная экономика ю 74,93 Ч
Тест ма
Учебным материал Раздел II. Микроэкономика » Тема 1Ч»3. Спрос, предложений и цена
Pic. 2. Presentation of pedagogical knowledge
10
62.59 «
Didactic knowledge here - it's a set of rules-productions, interpreting the link between the concepts of the subject area, analyzing and modifying model parameters trainee (WHO-knowledge), and manage the learning process, where, Rule is a rule, Condition is condition of the rule, Action is action. Production approach to the presentation of didactic
knowledge ensure ease of modification and replenishment, and consequently, adaptability and thinner "situational" learning system configuration in general. Controlling of learning process cannot be operated without feedback, which is realized by a variety of diagnostic knowledge.
Teiты равделое и подразделов
Показать данные по разд ал у: Все рагдепы Тесты Вопросы
В
ф Добавить вопрос типа ОопрК"
Fl
Р а эдол
Тип вопроса Сложность Время
Бухгалтерские издержки; Уч»6"ии » Раздел II-Мнкрмконрмикэ • Тема Msi, Поведение производителей. Производство и ... -С Г
В краткосрочном периоде фирма производит 500 единиц п родукци и. С ре д... Учебный материал » Раздел II. Микроэкономика - Тема №5. Поведение производителей, Производство и ... -С F
В макроэкономике, в том числе изучают;
Учебный материал »Раздел lit, Макроэкономика я Тома №7. Макроэкономически« проблемы и
ПОКчТЭЗГППН....
Pic. 3. Presentation of diagnostic knowledge
Description and diagnostic of knowledge representation is based on their models [2]. At the level, which is invariant to the subject area and type of activity, question's model (tasks, exercises) Mq is represented as:
Mq = ( A, C, Ms, Msu, V, P, Vu, Mas, Ov, Oa), (3)
Where:
A - the purpose (what is required of the student, which activities you need to perform);
C - limitations that should be considered when performing the task;
Ms - model of the situation (the component model can match the object being studied, professional work environment, etc.);
Msu - information model that describes a way of representing Ms, and the means of operating it as part of the course;
V - the results (answers);
P - setting the weight;
Vu - the description of the method of entering Mas result -the reference model activities;
Ov - function of evaluation results;
Oa - function of evaluation activities
This model allows to describe all types of diagnostic knowledge. In models of specific tasks, not all components can be used, which belong to (3). The most interesting for analysis Mq members are: Ms, P and Ov.
There are three components in a situation model, according to [2]:
Ms = (Str(Ms), Val(Ms), Int(Ms)), where (4)
Str(Ms) - Ms structure; Val(Ms) - Ms values of parameters; Int(Ms) -Interpretation Ms.
Structure Str(Ms) determines the form of expressions (mathematical relationships, rules, etc.) belonging to Ms. Specific settings are fixed in Val(Ms). This decomposition allows you to create a not-generated, and generated questions and tasks system. The latter ones can be obtained by variation of the parameters Val(Ms) at constant structure of model's situation or by the way of providing transformation Str(Ms). Interpretation Int(Ms)- characterizes the meaning (i.e., objective content) of elements Str(Ms) and Val(Ms).
Functions of evaluation Ov and Oa are defined as follows:
Ov:(Vs , V) Bl, (5)
Oa:(Ma , Mas) ^ Bl, where (6)
Vs - the result entered or selected by students;
Bl - the set of points;
Ma - a model of the activity made by the students during the process of response to the question related to handling Ms on the basis of the implementation of Msu.
Functions Ov, Oa and the mechanisms that realize the evaluation, are traditionally fixed in the framework of the development of tools (evaluation mechanism based on the logics of vector semantics which will be discussed below), but the use of the component approach to their design allows you to get rid of this restriction.
Let's consider the evaluation mechanism which is proposed in this article. For example, the task of estimating can be formulated as follows.
T - a set of diagnostic knowledge determined on a given subject area; TA - active sampling of the set T, by the way of accounting of parameters of the student's model; qi - the question of active sampling ( ). Considering that:
1) each question qi is characterized by weight pi, where
P, = aiai + aiPi + aYi - a4; here a ,Pi, Y,
- the characteristics of the question determined in accordance with V.P. Bespalko's indicators system [2],
a = 15 pt = 1,4, Yi = U; a\,ai,a, 6 M-
weight indicators a, ,Y, , so, a4 6 [0;l] - a correction factor;
2) for each question q, 6 T is definitely a lot of answers
with weights k6 [0,l] j = 1, N(i);
3) total score obtained while making the task is determined by the expression:
,(t) =
k p t , t = 0,2,3
1 --
N,
\ Nt
N
T y
^ kj , t = 1, NT * 0
j=o
p,KD, t = 4, 0 < KD < 1
-NT, t = 5
I N (7)
where t - the type of question; NF, NT - number of responses selected properly and correctly to questions, "Multiple Choice", "multiple choice below" and "correspondence". Require:
1. to evaluate the level of knowledge of student's teaching material covered by TA sample;
2. to establish the level of trust to the given mark in order to determine the moment of finishing the test, or, when "preterm exit" - taking to the account this level on the next steps of learning.
To solve this problem estimation technique is proposed, based on the logics with vector semantics [1]. In these logics truth of
a proposition a is described by the vector [fl] = (a+ ; a ^ ,
where a + , a 6 [0,l]. Here a+ - the true measure of a statement a and a - is a measure of its Lies.
In the case of diagnostic knowledge our judgment can be presented as the statement:
a = "The probationer has passed the test TA
(8)
In the simplest case, when the test can be only fulfilled or only not fulfilled, the result of a single answer to the qi issue (task) in the framework of this approach may be represented as a true vector [a]i:
[a].={(p.;0) if the task is done
(0,p.) if the task is not completed ) (9) When testing each response is considered as a sign of "for" or "against" statement (8). For the accumulation of evidence the rule of combination is used, according to it:
[a] = (F1 (a+, a+ ) F2 (a-, a- )) (10)
where F1(i+) and F2(i-) are some functions of the components of the truth, determined by the accumulation of evidence scheme and subject area. In the case of diagnostic knowledge such as functions following expressions are chosen :
F1(a+,a2+) = a+ © a2+ F2(a-,a2) = a- © a2
(11)
which provide a "symmetrical" accumulation of positive and negative test results.
The criterion for termination of testing is achieving of a given
, , ,,, r (a) = a + © a ~
threshold by measure of uncertainty
, which is considered as a level of confidence in the assessment at every step.
As the operation ©, next expression is selected:
X y min( , X y) according to which all the test items are taken into account in the same way, and the moment of reaching the threshold test of reliability is determined by a simple summation of the weights of all the proposed tasks. The components of the vector of estimation satisfy the constraint
a + + a- < 1.
To convert a vector evaluation to "scalar" the reliability
juA (a) = a + - a
measure is used
-1 <vA (a) < 0 0 <^A (a) < 0.5
. For example: «unsatisfactory»
- «satisfactory»
0.5 < uA (a) < 0.3
r A -«good»;
< fia (a) < 1 - «excellent».
The advantage of vectorization is the possibility of to take into account the weight of each question in the test and evaluate not only knowledge, but also ignorance of the probationer, which increases the effectiveness of feedback in teaching process.
This approach is practically realized in e-learning courses that are used in the training process of the East Siberian State University of Technology and Management (Ulan-Ude) and the Baikal branch of the Institute of Humanities (Moscow).
References
1. Arshinskij L.V. Metody obrabotki nestrogih vyskazyvanij. Irkutsk: VSI MVD RF. - 1998. - 40 p.
2. Bashmakov A.I., Bashmakov I A. Razrabotka komp'juternyh uchebnikov i obuchajushhih sistem. - M.: Informacionno-izdatel'skij dom «Filin», 2003. - 616 p.
3. Bespal'ko V.P. Osnovy teorii pedagogicheskih sistem: Problemy i metody psihologo-pedagogicheskogo obespechenija tehnicheskih obuchajushhih sistem. Voronezh: Izd-vo Voronezh. un-ta, 1977. - 304 p.
4. Gavrilova T.A. Objektno-strukturnaja metodologija konceptual'nogo analiza znanij i tehnologija avtomatizirovannogo proektirovanija baz znanij //Trudy Mezhdunar. konf. «Znanija - dialog-reshenie 95», T. 1. Jalta, 1995. - S. 1-9
5. Gavrilova T.A., Chervinskaja K.R. Izvlechenie i strukturirovanie znanij dlja jekspertnyh sistem. - M.: Radio i svjaz', 1992. - 200 p.
6. Informacionnaja tehnologija issledovanij razvitija jenergetiki /L.D. Krivorukcij, L. V. Massel'. - Novosibirsk: «Nauka». Sib. Izdatel'skaja firma RAN,1995.- 160 p.
7. Pugachev A.A., Proektirovanie i razrabotka adaptivnyh jelektronnyh uchebnyh kursov [Tekst]. - Ulan-Udje: Izdatel'sko-poligraficheskij kompleks FGOU VPO VSGAKI, 2007. - 149 s.
Wu, H., De Bra, P. Sufficient Conditions for Well-Behaved Adaptive Hypermedia Systems. Proceedings of the First Asia-Pacific Conference on Web Intelligence: Research and Development, pp. 148-152, 2001.
