Performance of Minkowski-type Distances in Similarity Search - A Geometrical Approach

Singh, Arpan; Jayaram, Balasubramaniam

doi:10.1109/iccca49541.2020.9250751

Cited by 2 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The contributions of this study are the incorporation of weights to each dimension of the dataset based on the principal component analysis (PCA) and the utilization of the Minkowski distance as the distance metric. Minkowski distance has been used in various studies for optimization such as in Chouikhi et al (2017), Xu et al (2019), Khaldi et al (2020, and Singh and Jayaram (2020).…”

Section: Introductionmentioning

confidence: 99%

An Improved K-Power Means Technique Using Minkowski Distance Metric and Dimension Weights for Clustering Wireless Multipaths in Indoor Channel Scenarios

Materum

Teologo

2021

Journal of Information and Communication Technology

View full text Add to dashboard Cite

Wireless multipath clustering is an important area in channel modeling, and an accurate channel model can lead to a reliable wireless environment. Finding the best technique in clustering wireless multipath is still challenging due to the radio channels’ time-variant characteristics. Several clustering techniques have been developed that offer an improved performance but only consider one or two parameters of the multipath components. This study improved the K-PowerMeans technique by incorporating weights or loads based on the principal component analysis and utilizing the Minkowski distance metric to replace the Euclidean distance. K-PowerMeans is one of the several methods in clustering wireless propagation multipaths and has been widely studied. This improved clustering technique was applied to the indoor datasets generated from the COST 2100 channel Model and considered the multipath components’ angular domains and their delay. The Jaccard index was used to determine the new method’s accuracy performance. The results showed a significant improvement in the clustering of the developed algorithm than the standard K-PowerMeans.

show abstract

Section: Introductionmentioning

confidence: 99%