Research on PCA-Kmeans++ clustering algorithm considering Spatiotemporal dimension

Aiming at the problem that traditional clustering algorithms cannot adapt to spatiotemporal data mining, this paper proposes a new clustering algorithm PCA-Kmeans++. First, in order to reduce the interference of data dimension, an improved PCA (principal component analysis) dimensionality reduction...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2023 2nd International Conference on 3D Immersion, Interaction and Multi-sensory Experiences (ICDIIME) s. 195 - 201
Hlavní autoři: Huang, Jiale, Dai, Jingtong, Li, Yanjin
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.06.2023
Témata:
AFC Data
Classification algorithms
Clustering algorithms
Data mining
Dimensionality reduction
Interference
K-means
PCA
Spatiotemporal clustering algorithm
Spatiotemporal phenomena
Visualization
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Aiming at the problem that traditional clustering algorithms cannot adapt to spatiotemporal data mining, this paper proposes a new clustering algorithm PCA-Kmeans++. First, in order to reduce the interference of data dimension, an improved PCA (principal component analysis) dimensionality reduction algorithm is built. On this basis, a K-means++algorithm considering space-time dimension is proposed to cluster the reduced factors. Finally, 100000 AFC data are collected for validity verification. The results show that: (1) The improved PCA algorithm has better dimensionality reduction effect. (2) The spatiotemporal clustering algorithm based on K-means++can effectively enhance the efficiency of classification decision-making. This study provides relevant basis and methodology for proposing a generic clustering algorithm.
DOI:10.1109/ICDIIME59043.2023.00042