MINING SENSOR DATA STREAMS - Managing and Mining Sensor Data - page 165

extremely important issue since the data may evolve considerably over
time. Other important topics which need to be explored are as follows:
Streams are often collected by devices such as sensors in which the
data is often noisy and error-driven. Therefore, a key challenge is
to effectively clean the data. This may involve either imputing or
modeling the underlying uncertain data. This can be challenge,
since any modeling needs to be done in real time, as large volumes
of the data stream arrive.
A related area of research is in using the modeled data for data
mining tasks. Since the underlying data is uncertain, the uncer-
tainty should be used in order to improve the quality of the under-
lying results. Some recent research addresses the issue of clustering
uncertain data streams [7].
Many recent applications such as privacy-preserving data mining
have not been studied effectively in the context of data streams. It
is often a challenge to perform privacy-transformations of contin-
uously arriving data, since newly arriving data may compromise
the integrity of earlier data. The data stream domain provides a
number of unique challenges in the context of the privacy problem.
References
[1] Aggarwal C., Xie Y., Yu P. (2011) On Dynamic Data-driven Selection
of Sensor Streams, ACM KDD Conference .
[2] Aggarwal C., Bar-Noy A., Shamoun S. (2011) On Sensor Selection
in Linked Information Networks, DCOSS Conference .
[3] Abadi D., Madden S., Lindner W. (2005) REED: robust, ecient
filtering and online event detection in sensor networks, VLDB Con-
ference .
[4] Aggarwal C. (2007) Data Streams: Models and Algorithms, Springer .
[5] Aggarwal C., Procopiuc C, Wolf J. Yu P., Park J.-S. (1999) Fast
Algorithms for Projected Clustering. ACM SIGMOD Conference .
[6] Aggarwal C. (2006) On Biased Reservoir Sampling in the presence
of Stream Evolution. VLDB Conference .
[7] Aggarwal C., Yu P. (2008) A Framework for Clustering Uncertain
Data Streams. ICDE Conference .
[8] Aggarwal C. (2003) A Framework for Diagnosing Changes in Evolv-
ing Data Streams. ACM SIGMOD Conference .