2019/07/26 by Nicholas James Rowe, Rowe, Nicholas
Computer Science · Earth and Planetary Sciences · Engineering · #Electrical and Bioimpedance Tomography #FOS: Computer and information sciences #Geophysical and Geoelectrical Methods #Image and Signal Denoising Methods #Methodology (stat.ME)
paper · pdf · doi:10.48550/arxiv.1907.11508
openalex publication_date 2019/07/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This paper considers application of the SALSA algorithm as a method of forecasting and applies it to simulated electrical signal, temperature recording from the Australian Bureau of Meteorology and stock prices from the Australian stock exchange. It compares it to basic linear extrapolation and casual smoothing extrapolation, in all cases SALSA extrapolation proves to be a better method of forecasting than linear extrapolation. However, it cannot be imperially stated that it is superior to Causal smoothing extrapolation in complex systems as it has a higher L2 euclidean in these experiments. while usually retaining more shape and statistical elements of the original function than Causal smoothing extrapolation. Leading to the conclusion the Causal Smoothing extrapolation can provide a more conservative forecast for complex systems while the SALSA algorithm more accurately predicts the range of possible events as well as being the superior forecasting method for electrical signals, the physical process it is designed to forecast.