Database Reference
In-Depth Information
Basic uncertain data models
Probabilistic database model
Uncertain object model
Elements
Elements
A set of uncertain tuples
A set of generation rules
A set of uncertain objects
Each object contains a set of instances
Equivalent
Properties
Properties
Data type: static
Dependency: independency among tuples in
different generation rules
Data type: static
Dependency: independency among objects
Uncertain data stream model
Probabilistic linkage model
Elements
Uncertain road network model
Elements
Elements
A set of uncertain data streams
Each stream contains a series of instances
Two sets of tuples (objects)
A set of linkages (instances) between tuples
A simple graph
Each edge weight isan uncertain object
Properties
Properties
Properties
dynamic
Data type:
Dependency: independency among objects
Data type: static
Dependency:
2 tulpes share 0 or 1 linkage
Data type: static
Dependency:
Complex structure:
correlations among weights
a simple graph
Extended uncertain data models
Fig. 2.7
The basic and extended uncertain data models adopted in this topic.
2.3.3.3 Connections with the Uncertain Object Model
In the uncertain road network model, the weight of each edge can be considered
as an uncertain object with a set of instances. The weight of a path is an aggregate
sum of the uncertain objects corresponding to the edges in the path. Therefore, a
probabilistic query essentially ranks a set of aggregate sums of uncertain objects.
2.4 Summary
In this chapter, we reviewed two basic uncertain data models, the uncertain object
model and the probabilistic database model, as well as the possible worlds semantics
that have been extensively adopted in other research on uncertain data.
Moreover, we proposed three extended uncertain data models that extend the
uncertain object model from different aspects.
•
The
uncertain data stream model
captures the dynamic nature of uncertain data.
Each uncertain data stream is a series of (potentially) unlimited instances. Given
a sliding window that selects the time span of interest, the instances of each
uncertain data stream in the sliding window can be considered as an uncertain
object. The uncertain data stream model suits the needs of applications that in-
