الفهرس Only 14 pages are availabe for public view
 |  | from | 161 |  |  |
| | | from | 161 | | |
Abstract
The main objective of this thesis is the studying of the cognitive phenomena of concept
categorization and knowledge representation aiming to build computational cognitive models for both
phenomena.
In pursuing this aim, a new computational cognitive model is proposed for binary concept
categorization (CCMBC) based on studying psychological theories of human concept categorization. The
core of the proposed model is deriving the categorization rules from the training data in order to classify
new instances based on rules gained. This model is formulated by using a mathematical entity known as
concept algebra which is one of the paradigms of denotational mathematics. Concept algebra is the
mathematical base of human’s logic thinking. Also, a new semantic model based on cognitive aspects is
proposed to enable a cognitive computer to process the knowledge as the human mind and find the
representation rules. This thesis preforms a comparison between the proposed models with several
machine learning and statistical classification techniques. The proposed models have been tested with
several data sets belonging to the UCI machine learning repository and the results prove that the
performance of the proposed models outperform several machine learning and statistical classification
techniques.
chapters: threeThis thesis consists of
Chapter one:
This chapter presents the background about the mathematical and computational cognitive modeling.
Then, it emphasizes on computational cognitive modeling that plays a central role in the computational
understanding of the human mind to predict human behavior. Moreover, cognitive informatics (CI) is
presented to study of cognitive and information science. Finally, denotational mathematics (DM) is
illustrated to deal with complex mathematical entities emerged in cognitive science, which are able to
describe software and intelligent architectures and behaviors rigorously, precisely, and expressively.
Chapter two:
This chapter presents the background about categorization as a basic cognitive process of arranging
objects into categories, the purpose of categorization, and the four basic types of category learning
models. Then, some supervised machine learning approaches to category learning are presented. Finally,
a new computational cognitive model for binary categorization (CCMBC) is proposed based on studying
psychological theories of human concept categorization that consists of four phases. New compositional
operations on formal concepts in concept algebra are defined to be used in CCMBC to simulate the
human concept categorization. The experimental results and discussions are presented and prove that
CCMBC outperform other categorization algorithms.
Chapter three:
This chapter presents the background about cognitive computing (CC) and cognitive computers
(CogCs) that are a novel form of intelligent computers, which embody major natural intelligence
behaviors of the mind. Then, the three major stages for knowledge processing in CogC are presented.
Finally, the proposed semantic model is formulated to find a suitable representation of knowledge and
making this representation available in a usable form when it is needed in the future. Experimental results
and discussions are presented. It proves that the proposed semantic model has the highest performance in
representing the knowledge by symbolic rules.