A novel prediction method of complex univariate time series based on k-means clustering.
Yunxin LiuShifei DingWeikuan JiaPublished in: Soft Comput. (2020)
Keyphrases
- significant improvement
- cost function
- objective function
- high precision
- pairwise
- experimental evaluation
- detection method
- exponential smoothing
- prediction accuracy
- dynamic programming
- computationally efficient
- support vector machine svm
- non stationary
- spectral clustering
- prediction algorithm
- high accuracy
- clustering method
- computational cost
- synthetic data
- cross validation
- preprocessing
- prediction error
- bayesian networks