Collaborative data stream mining in ibiquitous environment using dynamic classifier selection

TY - JOUR

T1 - Collaborative data stream mining in ibiquitous environment using dynamic classifier selection

AU - DEE Group Author

AU - Sousa, Pedro Alexandre da Costa

PY - 2013/1/1

Y1 - 2013/1/1

N2 - In ubiquitous data stream mining, different devices often aim to learn concepts that are similar to some extent. In many applications, such as spam filtering or news recommendation, the data stream underlying concept (e.g., interesting mail/news) is likely to change over time. Therefore, the resultant model must be continuously adapted to such changes. This paper presents a novel Collaborative Data Stream Mining (Coll-Stream) approach that explores the similarities in the knowledge available from other devices to improve local classification accuracy.Coll-Streamintegrates the community knowledge using an ensemble method where the classifiers are selected and weighted based on their local accuracy for different partitions of the feature space. We evaluateColl-Streamclassification accuracy in situations with concept drift, noise, partition granularity and concept similarity in relation to the local underlying concept. The experimental results show thatColl-Streamresultant model achieves stability and accuracy in a variety of situations using both synthetic and real-world datasets.Read More:http://www.worldscientific.com/doi/abs/10.1142/S0219622013500375

AB - In ubiquitous data stream mining, different devices often aim to learn concepts that are similar to some extent. In many applications, such as spam filtering or news recommendation, the data stream underlying concept (e.g., interesting mail/news) is likely to change over time. Therefore, the resultant model must be continuously adapted to such changes. This paper presents a novel Collaborative Data Stream Mining (Coll-Stream) approach that explores the similarities in the knowledge available from other devices to improve local classification accuracy.Coll-Streamintegrates the community knowledge using an ensemble method where the classifiers are selected and weighted based on their local accuracy for different partitions of the feature space. We evaluateColl-Streamclassification accuracy in situations with concept drift, noise, partition granularity and concept similarity in relation to the local underlying concept. The experimental results show thatColl-Streamresultant model achieves stability and accuracy in a variety of situations using both synthetic and real-world datasets.Read More:http://www.worldscientific.com/doi/abs/10.1142/S0219622013500375

KW - Collaborative data stream mining

KW - ubiquitous knowledge discovery

KW - performance evaluation

KW - concept drift

U2 - 10.1142/S0219622013500375

DO - 10.1142/S0219622013500375

M3 - Article

VL - 12

SP - 1287

EP - 1308

JO - International Journal Of Information Technology & Decision Making

JF - International Journal Of Information Technology & Decision Making

SN - 0219-6220

IS - 6

ER -