Design Method For Knowledge Base Systems in Education Using COKB-ONT
Design Method For Knowledge Base Systems in Education Using COKB-ONT
Design Method For Knowledge Base Systems in Education Using COKB-ONT
162
World Academy of Science, Engineering and Technology 48 2008
concepts in the set C; H represents these special relations on domains of knowledge. The kinds of facts are as follows:
C. This relation is an ordered relation on the set C, and H - Fact of kind 1: information about object kind.
can be considered as the Hasse diagram for that relation. - Fact of kind 2: a determination of an object or an
R is a set of other relations on C, and in case a relation r is attribute of an object.
a binary relation it may have properties such as reflexivity, - Fact of kind 3: a determination of an object or an
symmetry, etc. In plane geometry and analytic geometry, attribute of an object by a value or a constant
there are many such relations: relation “belongs to” of a expression.
point and a line, relation “parallel” between two line - Fact of kind 4: equality on objects or attributes of
segments, relation “perpendicular” between two line objects. This kind of facts is also popular, and there
segments, the equality relation between triangles, etc. are many problems related to it on the knowledge
base.
The set Ops consists of operators on C. This component
- Fact of kind 5: a dependence of an object on other
represents a part of knowledge about operations on the
objects by a general equation.
objects. Almost knowledge domains have a component
- Fact of kind 6: a relation on objects or attributes of
consisting of operators. the objects. In almost problems there are facts of
The set Funcs consists of functions on C-Objects. kind 6 such as the parallel of two lines, a line is
Knowledge about functions is also a popular kind of perpendicular to a plane, a point belongs to a line
knowledge in almost knowledge domains in practice, segment.
especially fields of natural sciences such as fields of - Fact of kind 7: a determination of a function.
mathematics, fields of physics. - Fact of kind 8: a determination of a function by a
The set Rules represents for deductive rules. The rules value or a constant expression.
represent for statements, theorems, principles, formulas, and - Fact of kind 9: equality between an object and a
so forth. Almost rules can be written as the form “if <facts> function.
then <facts>”. In the structure of a deductive rule, <facts> is - Fact of kind 10: equality between a function and
a set of facts with certain classification. Therefore, we use another function.
deductive rules in the COKB model. Facts must be classified - Fact of kind 11: a dependence of a function on
so that the component Rules can be specified and processed other functions or other objects by an equation.
in the inference engine of knowledge base system or The above models and kinds of facts can be used to
intelligent systems. represent knowledge in practical applications. Unification
Base on the COKB model, the knowledge base can be algorithms of facts were designed and used in different
organized by the following components: applications.
1. The dictionary of concepts about kinds of objects, C. Specification Language for COKB model
attributes, operators, functions, relations and
The language for the COKB model is constructed to
related concepts.
2. The table of descriptions for structures and features represent the knowledge of the form COKB model. This
of objects. For example, we can request a triangle language includes the following:
to compute and to give us its attributes. - A set of characters: letter, number, special letter.
3. The tables for representing hierarchical relations of - Vocabulary: keywords, names.
concepts. - Data types: basic types and structured types.
4. The tables for representing other relations of - Expressions and sentences.
concepts. - Statements.
5. The tables for representing knowledge about - Syntax for specifying the components of COKB
operators. model.
6. The tables for representing knowledge about
functions. III. DESIGN METHOD
7. The tables of descriptions for kinds of facts. For In this section, we will present a process to construct a
example, a relational fact consists of kind of the knowledge base system for solving some problems in higher
relation and the list of objects in the relation. mathematics. Besides, techniques in each phase will be
8. The tables of descriptions for rules. For example, a presented also.
deductive rule consists of hypothesis part and
conclusion part. Both of them are lists of facts. A. Structure of System
9. The lists or sets of rules. A system, which supports searching, querying and solving
10. The lists of problem patterns. higher mathematics problem, has the structure of an expert
B. Kinds of facts in COKB model system. We can design the system which consists of six
components:
In the COKB model there are 11 kinds of facts accepted.
- The knowledge base.
These kinds of facts have been proposed from the - The inference engine.
researching on real requirements and problems in different
163
World Academy of Science, Engineering and Technology 48 2008
164
World Academy of Science, Engineering and Technology 48 2008
IV. CASE STUDY: DESIGN PARTERN – A • Derivation: derivation at a point, one-sided derivation at
KNOWLEDGE BASE SYSTEM FOR QUERYING one point, derivation in a range, derivation of inverse
KNOWLEDGE OF HIGHER MATHEMATICS function, high derivative, etc.
The software can support for searching, querying and • Differential: differential invariant, differential at a point,
solving some problems in higher mathematics. The first differential on a range, primary differential, advance
function of the program is "Search for Knowledge". This differential, etc.
function helps users to find out necessary knowledge • Initial function: complete primitive, primitive function
quickly. It can search for concepts, definitions, properties, calculus, etc.
related theorems or formulas, and problem patterns. Users • Integral: integral formula, integral domain,
declare some information base on a simple language. The indeterminate integral, determinate integral, generalized
integral with infinite bound (type 1 generalized integral),
information can consist of objects, relations between objects
generalized integral of unbound function (type 2
or between attributes. The program can automatically give
generalized integral), integrate by parts, etc.
instructions that help them to know more about the problem.
• Theory of series: series, convergent series, positive
The second function of the program is “Query for
series, negative series, alternate series, random series,
Knowledge”. This function helps users to query knowledge series of functions, domain of convergence of series of
based on the query language which is a very simple functions, convergence of series of functions,
language. exponential series, radius of convergence of exponential
Stage 1: Collecting real knowledge based o n COKB- series, Taylor's series of a function f(x) at adjacent point
ONT model. x0., trigonometric series, Fourier’s series, Fibonacci
The knowledge of higher mathematics consists of the series, etc.
following components. B. List of properties: properties of series of convergence,
A. Concepts: bounded, order, etc.
• Numeric representation: natural numbers, integers, real C. List of formulas: Natural limit of series, Formulas for
numbers, complex numbers, etc. power function, Formulas for Logarithmic function,
• Category of sets: collection, Venn diagram, set Taylor formula, MacLaurin’s theorem, L’Hospital rule,
operatoions, etc. Derivation table, Cauchy-Schwarz’s inequality,
• Mapping: map, injection, surjection, bijection, identity Newton-Leibnitz formula, etc.
mapping, inverse mapping, etc. D. List of computing method: Function exploration,
• Sequence of numbers: sequence, bounded sequence, computing method of undetermined integral: analysis
bounded below sequence, bounded above sequence, method, variable method, partial method, computing
Cauchy’s sequence, monotone sequence, Fibonacci method of determined integral: analysis method,
sequence, pseudo random number sequence, strictly variable method, partial method finding radius
decreasing sequence, convergent sequence of function, convergence of power series, etc.
convergence monotone sequence, double sequence, null E. List of theorems: Infinitesimal and infinitely great
sequence, decreasing sequence, random number theorem, Continuity theorem of function: Weierstrass1
sequence, strictly increasing sequence, limit of a theorem, Weierstrass2 theorem, Theorem of derivation,
sequence, etc. Mean theorem: Fermat, Rolle, Lagrange, Cauchy,
• Real function with one variable: function, domain of a L’Hospital 1 and L’Hospital 2 Theorems for function
function, the range of a function, arithmetic function, exploration, Theorem for initial function and integral,
even function, odd function, doubly periodic function, Series theorem, etc.
monotone function, polynomial, rational function, F. Some applications of simply analytic function:
composite function, inverse function, reversible geometric meaning of derivation, application of
function, complex function, power function, determinate integral, etc.
exponential, Logarithmic function, trigonometric Stage 2: Classifying the k n o w l e d g e in the Stage 1, to
functions, inverse trigonometric functions, limit of a analyze requirements.
function, one-sides limit, limit of composite function, • Basic concept: numeric representation, collection.
limit of monotone function, infinitesimal, infinitely • Concepts are constructed from other concepts. For
great, continuity of function, continuous function at one example: Map: collection; Function: map; Limit of
point, one-sided continuous function, continuous series: series; Limit of function: function, limit of series;
function in a range, piecewise continuous function, point Chart of function: function, point; Derivation: function,
of discontinuity of function, function chart, maximum of limit of function; derivatives of higher order: derivation;
a function, minimum of a function, concave of function, Differential: function, derivation; Initial function:
convexity of function, bending point of function chart, function, derivation; Integral: function, initial function;
approximations of function chart, tangent line of Series: array of number; Positive series: series;
function chart, etc. Alternative series: series; Series of functions: series,
functions; Power series: series of functions; etc.
165
World Academy of Science, Engineering and Technology 48 2008
166
World Academy of Science, Engineering and Technology 48 2008
Return f: FUNCTION #f is the inverse function of y - The file SOMEOBJECTS.txt stores certain objects.
Begin_description Stage 4: Modeling for problems and designing
\Contains_Object\INVERSE_FUNCTION
End_description algorithms. Problems are represented using a model that is
End_Function called networks of C-Objects.
Begin_Function LIMIT_OF_ARRAY(d) The design of deductive algorithms for solving
d: ARRAY
Return a: REAL problems and the design of interface of the system can
Begin_description be developed by three steps for modeling:
\Contains_Object\ LIMIT_OF_ARRAY Step 1: Classify problems such as problems as
End_description frames, problems of a determination or a proof of a
End_Function
etc. fact, problems of finding objects or facts, etc.
End_Functions Step 2: Classify facts and representing them based
- The file FACT_KINDS.txt stores the definition of kinds on the kinds of facts of COKB model.
of facts. Step 3: Modeling kinds of problems from
Begin_Methods classifying in step 1 and 2. From models of each
Begin_Method < name of method> kind, we can construct a general model for
Begin_description
<path to the file which describes method> problems, which are given to the system for solving
End_description them.
End_Method The basic technique for designing deductive algorithms
Begin_Method <name of method> is the unification of facts. Based on the kinds of facts
Begin_description
<path to the file which describes method> and their structures, there will be criteria for unification
End_description proposed. Then it produces algorithms to check the
End_Method unification of two facts.
... The next important work is doing research on strategies
End_Methods
Example: for deduction to solve problems on computer. The most
Begin_Methods difficult thing is modeling for experience, sensible
Begin_Method INDETERMINATE_FORM_ELIMINATION reaction and intuitional human to find the heuristics
Begin_description rules, which were able to imitate the human thinking
\Contains_Object\
INDETERMINATE_FORM_ELIMINATION for solving problems.
End_description Stage 5: Creating a query language for the COKB-ONT.
End_Method There are eight of simple query kinds and one of combine
Begin_Method FUNCTION_EXPLORATION_PROCESS
query kind:
Begin_description
\Contains_Object\ FUNCTION_EXPLORATION_PROCESS • Kind 1: ?concepts|define <concept>
End_description Example:
End_Method ?concepts COLLECTION
...
End_Methods
?define FUNCTION
- The file RULES.txt stores deductive rules. • Kind 2: ?attributes <concept>
Begin_Rules Example: ?attributes COLLECTION
Begin_Rule <Rule name>:<kind of rule> • Kind 3: ?properties <concept>
Variables:
<object>:<kind> Example: ?properties CONVERGENCE_ARRAY
<object>:<kind> • Kind 4: ?formula <concept>
... Example: ?formula LIMIT_OF_ARRAY
Begin_description
< path to the file which describes rules> • Kind 5:
End_description a)?theorems <concept>
Goal: Example: ?theorems DERIVATE
<result of using rule> b) ?theorems <concept1, concept2,...>
End_Goal
End_Rule Example: ?theorems FUNCTION, DERIVATE
Begin_Rule <Rule name>:<kind of rule> c) ?content <name of theorem>
Variables: Example: ?content WEIERSTRASS_THEOREM
<object>:<kind> • Kind 6: ?methods <concept>
<object>:<kind>
... Example: ?methods LIMIT_OF_FUNCTION
Begin_description • Kind 7: ? concepts_related <concept>
< path to the file which describes rules> Example: ? concept_object FUNCTION
End_description
Goal:
Kind 8: querying of many kind (kind 2 to kind 7)
<result of using rule> q → kind 1..7
End_Goal kind 8 → q1 and q2
End_Rule | kind 81 and kind 82
...
End_Rules Example:
167
World Academy of Science, Engineering and Technology 48 2008
168