the intrinsic contradictions of IoT, pointing out that the manifestation diversities of
things and the inadequate understanding of the users are the key factors, leading to
affecting the further development of IoT, and then put forward a strategy for it—SWoT.
Based on these, machine learning is introduced and combined with semantic
technology in the paper. Aiming at this research, we design the hierarchical architecture
framework of SWoT. And a more specific and rigorous SWoT structural relationship
model is constructed; the Agents are extended to the abstract substitutes of all entities,
and the interoperability mechanism of the entities is realized equivalently through the
dynamic relationships between the Agents. Herein, Agents are divided into two categories: individual-Agent and object-Agent. At the same time, combining with composite ontology, dynamic relationship reasoning models of two kinds of Agents are
constructed.
2 Extension of SWoT Model
Construct a corresponding avatar (Agent) for the things (network entities, hereinafter
referred to as entities) in SWoT. In this way, SWoT can be considered from two
perspectives: (1) at the horizontal level, Agents constitute the network layer, and
physical entities constitute the physical layer; (2) at the vertical level, the same category
of hierarchy is constituted by each kind of entity and the corresponding Agents.
Human-centered construction-based IoT [7] regards serving people as the main purpose, and therefore, in order to be more general, entities are divided into two categories,
including the individuals and the objects.
Although layers can be independent each other, they are also interrelated and have
loose coupling characteristics. The hierarchies constructed by individuals and corresponding Agents can exist independently. Agent reasoning is used to predict human
daily activities and discover interpersonal relationships. Similar mechanisms can also
be applied to objects and corresponding agent levels. At the same time, all entities in
the real world are connected by events [8, 9], and therefore, the two levels above can
influence each other through various events between them.
So, the evolution process of the formal representation model of SWoT is as follows:
Entity Set Pattern ! Entity Network Pattern ! Agent Network Pattern.
Entity set pattern. SWoT= (X, N, E, T). X = (O, I) represents the entity set in
SWoT. Here, O stands for “object,” referring to the entity distinguished from “individual” in SWoT; I stands for “individual;” N stands for “network,” which is the
various storage sets of entity information in SWoT; E stands for “event.” T stands for
time, which reflects the dynamic signs of SWoT.
Entity Network Pattern. SWoT = (N
X , E, R, T). N
X = (IoO, IoI) = IoO ⊕
IoI = N
O
[ N
I represents the fusion network consisted of entity sets X in SWoT. IoO
(Internet of Objects) = O ⊕ N = N
O , which is networks of existing objects (e.g.,
sensor networks, etc.). IoI (Internet of Individuals) = I ⊕ N = N
I , which is social
networks (e.g., Wechat, QQ, Blog, etc.). Here, binary operator ⊕ represents fusion;
E = E O⊗O [ E I⊗I [ E O⊗I [ I⊗O is a set of events occurring on O [ I. E O⊗O is a set
of events between objects, E I⊗I is a set of events between individuals, and E O⊗I [ I⊗O is
a set of events between individuals and objects; R = R O⊗O [ R I⊗I [ R O⊗I [ I⊗O
Re-sculpturing Semantic Web of Things …
555
things and the inadequate understanding of the users are the key factors, leading to
affecting the further development of IoT, and then put forward a strategy for it—SWoT.
Based on these, machine learning is introduced and combined with semantic
technology in the paper. Aiming at this research, we design the hierarchical architecture
framework of SWoT. And a more specific and rigorous SWoT structural relationship
model is constructed; the Agents are extended to the abstract substitutes of all entities,
and the interoperability mechanism of the entities is realized equivalently through the
dynamic relationships between the Agents. Herein, Agents are divided into two categories: individual-Agent and object-Agent. At the same time, combining with composite ontology, dynamic relationship reasoning models of two kinds of Agents are
constructed.
2 Extension of SWoT Model
Construct a corresponding avatar (Agent) for the things (network entities, hereinafter
referred to as entities) in SWoT. In this way, SWoT can be considered from two
perspectives: (1) at the horizontal level, Agents constitute the network layer, and
physical entities constitute the physical layer; (2) at the vertical level, the same category
of hierarchy is constituted by each kind of entity and the corresponding Agents.
Human-centered construction-based IoT [7] regards serving people as the main purpose, and therefore, in order to be more general, entities are divided into two categories,
including the individuals and the objects.
Although layers can be independent each other, they are also interrelated and have
loose coupling characteristics. The hierarchies constructed by individuals and corresponding Agents can exist independently. Agent reasoning is used to predict human
daily activities and discover interpersonal relationships. Similar mechanisms can also
be applied to objects and corresponding agent levels. At the same time, all entities in
the real world are connected by events [8, 9], and therefore, the two levels above can
influence each other through various events between them.
So, the evolution process of the formal representation model of SWoT is as follows:
Entity Set Pattern ! Entity Network Pattern ! Agent Network Pattern.
Entity set pattern. SWoT= (X, N, E, T). X = (O, I) represents the entity set in
SWoT. Here, O stands for “object,” referring to the entity distinguished from “individual” in SWoT; I stands for “individual;” N stands for “network,” which is the
various storage sets of entity information in SWoT; E stands for “event.” T stands for
time, which reflects the dynamic signs of SWoT.
Entity Network Pattern. SWoT = (N
X , E, R, T). N
X = (IoO, IoI) = IoO ⊕
IoI = N
O
[ N
I represents the fusion network consisted of entity sets X in SWoT. IoO
(Internet of Objects) = O ⊕ N = N
O , which is networks of existing objects (e.g.,
sensor networks, etc.). IoI (Internet of Individuals) = I ⊕ N = N
I , which is social
networks (e.g., Wechat, QQ, Blog, etc.). Here, binary operator ⊕ represents fusion;
E = E O⊗O [ E I⊗I [ E O⊗I [ I⊗O is a set of events occurring on O [ I. E O⊗O is a set
of events between objects, E I⊗I is a set of events between individuals, and E O⊗I [ I⊗O is
a set of events between individuals and objects; R = R O⊗O [ R I⊗I [ R O⊗I [ I⊗O
Re-sculpturing Semantic Web of Things …
555