Alzennyr Da Silva

A clustering approach for sampling data streams in sensor networks

The growing usage of embedded devices and sensors in our daily lives has been profoundly reshaping the way we interact with our environment and our peers. As more and more sensors will pervade our future cities, increasingly efficient infrastructures to collect, process and store massive amounts of data streams from a wide variety of sources will be required. Despite the different application-specific features and hardware platforms, sensor network applications share a common goal: periodically sample and store data collected from different sensors in a common persistent memory. In this article, we present a clustering approach for rapidly and efficiently computing the best sampling rate which minimizes the Sum of Square Error for each particular sensor in a network. In order to evaluate the efficiency of the proposed approach, we carried out experiments on real electric power consumption data streams provided by EDF (Électricité de France).

CLUSMASTER: A Clustering Approach for Sampling Data Streams in Sensor Networks

The growing usage of embedded devices and sensors in our daily lives has been profoundly reshaping the way we interact with our environment and our peers. As more and more sensors will pervade our future cities, increasingly efficient infrastructures to collect, process, and store massive amounts of data streams from a wide variety of sources will be required. Despite the different application-specific features and hardware platforms, sensor network applications share a common goal: periodically sample and store data collected from different sensors in a common persistent memory. In this article we present a clustering approach for rapidly and efficiently computing the best sampling rate which minimizes the SSE (Sum of Square Errors) for each particular sensor in a network. In order to evaluate the efficiency of the proposed approach, we carried out experiments on real electric power consumption data streams produced by a 1-thousand sensor network provided by the French energy group – EDF (Electricite de France).

CAMEUD: clustering approach for mining evolving usage data

The growing number of traces left behind user transactions on the Internet (e.g. customer purchases, user navigations, etc.) has increased the importance of Web usage data analysis. A notable challenge of this analysis is the fact that the way in which a website is visited can evolve over time. As a result, the usage models must be continuously updated in order to reflect the current behaviour of the visitors. In this article, we introduce CAMEUD, a clustering approach to mine and detect changes in evolving usage data. The proposed approach is totally independent from the clustering algorithm applied in the classification problem and is able to detect and determine the nature of changes undergone by the usage groups (appearance, disappearance, fusion and split) at subsequent time intervals. Experiments on synthetic and real usage data sets evaluate the efficiency of CAMEUD.

Comparing Clustering on Symbolic Data

Although various dissimilarity functions for symbolic data clustering are available in the literature, little attention has thus far been paid to making a comparison between such different distance measures. This paper presents a comparative study of some well known dissimilarity functions treating symbolic data. A version of the fuzzy c-means clustering algorithm is used to create groups of individuals characterized by symbolic variables of mixed types. The proposed approach provides a fuzzy partition and a prototype for each cluster by optimizing a criterion dependent on the dissimilarity function chosen. Experiments involving benchmark data sets are carried out in order to compare the accuracy of each function. To analyse the results, we apply an external criterion that compares different partitions of a same data set.