Jaccard Coefficients based Clustering of XML Web Messages for Network Traffic Aggregation

Abstract

This paper provides static efficient clustering model based simple Jaccard coefficients that supports XML messages aggregator in order to potentially reduce network traffic. The proposed model works by grouping only highly similar messages with the aim to provide messages with high redundancy for web aggregators. Web messages aggregation has become a significant solution to overcome network bottlenecks and congestions by efficiently reducing network volume by aggregating messages together removing their redundant information. The proposed model performance is compared to both K-Means and Principle Component Analysis (PCA) combined with K-Means. Jaccard based clustering model has shown potential performance as it only consumes around %32 and %25 processing time in comparison with K-Means and PCA combined with K-Means respectively. Quality measure (Aggregator Compression Ratio) has overcome both benchmark models