A Self-Adaptive OPTICS Clustering Algorithm Based on the Lightning Distribution
| Autor: | Yaoling Zhi, Junchi Zhou, Huali Wu |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Computer science
0206 medical engineering 0202 electrical engineering electronic engineering information engineering Thunderstorm 020201 artificial intelligence & image processing OPTICS algorithm Self adaptive 02 engineering and technology Cluster analysis Algorithm 020602 bioinformatics Universality (dynamical systems) |
| Zdroj: | ICNC-FSKD |
| DOI: | 10.1109/fskd.2018.8686947 |
| Popis: | The clustering algorithm is a popular lightning data processing method. Traditional density-based algorithm can't function without input of initial parameter Min_ρ or neighborhood radius e. Besides, in most regarding studies, the initial parameter is set by experience, but not on the basis of selective and sensitive analysis. Due to poor universality, they can hardly be used in daily business. This work analyzed a thunderstorm in 31/7/2017 and found that the characteristic parameter ρ_0, extracted from the lightning data distribution, could help designate the initial parameter of the clustering algorithm. So the improved OPTICS algorithm could dynamically retrieve its initial parameter from the lightning data that is to be processed, enjoying good self-adaption. Our calculation proved the extracted characteristic parameter ρ_0 the optimum initial input which has satisfying clustering effect. As a result, the improved OPTICS algorithm applies well to the lighting data with different spatial and temporal distribution and this could be used in daily business. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|