Database Reference
In-Depth Information
can process a data item at most once. This leads to constraints
on the implementation of the underlying algorithms. Therefore,
stream mining algorithms typically need to be designed so that
the algorithms work with one pass of the data.
In most cases, there is an inherent temporal component to the
stream mining process. This is because the data may evolve over
time. This behavior of data streams is referred to as
temporal lo-
cality
. Therefore, a straightforward adaptation of one-pass mining
algorithms may not be an effective solution to the task. Stream
mining algorithms need to be carefully designed with a clear focus
on the evolution of the underlying data.
Data which is collected from sensors is often uncertain and error
prone, as a result of which it is critical to be able to reduce the
effects of the uncertainty in the mining process.
Another important characteristic of sensor data streams is that they are
often mined in a distributed fashion. In some cases, intermediate sen-
sor nodes may have limited processing power, and it may be desirable
to perform in-network sensor processing for a variety of mining applica-
tions. In such cases, the application algorithms need to be designed with
such criteria in mind [30, 60]. This chapter will provide an overview of
the key challenges in sensor stream mining algorithms which arise from
the unique setup in which these problems are encountered.
This chapter is organized as follows. In the next section, we will
discuss the generic challenges which arise in the context of storage and
processing of sensor data. The next section deals with several issues
which arise in the context of data stream management. In section 3,
we discuss several mining algorithms on the data stream model. Section
4 discusses various scientific applications of data streams. Section 5
discusses the research directions and conclusions.
2. Sensor Stream Mining Issues
Since data streams are processes which create large volumes of in-
coming data, they lead to several challenges in both processing the data
as well as applying traditional database operations. Therefore, new de-
signs of data streaming systems are required for handling sensor data
[23]. The challenging issues in sensor stream mining may arise during
different phases including data collection, transmission, storage and pro-
cessing. Some of the these key issues are as follows:
