Relative Reference Measure for Hierarchical Document Clustering

Provided by: Creative Commons
Topic: Big Data
Format: PDF
Clustering is a foremost concept in data mining. Clustering usually require a measure that needs to be computed among the clustering objects this measure could be either a similarity or a dissimilarity measure, here a multiple relative reference based similarity measure is used which could give more informative assessment of the similarity moreover this measure is derived from the traditional cosine similarity measure. In order to evaluate the performance of proposed measure an incremental clustering algorithm-master only algorithm is used initially with cosine similarity and then with Relative Reference similarity measures for document clustering.

Find By Topic