Provided by:
Creative Commons
Topic:
Big Data
Format:
PDF
Theoretical analysis and scientific illustrations display that Multi-Viewpoint based Similarity (MVS) is possibly more suitable for written text records than the well-known cosine similarity. Depending on MVS, two requirements features, IR and IV, and their specific clustering techniques, MVSC-IR and MVSC-IV, have been presented. In contrast to other state-of-the-art clustering techniques that use different types of likeness evaluate, on a huge variety of papers, data sets and under different assessment analytics, the proposed algorithms display that they could offer significantly improved clustering efficiency.
