Sample Vector Research Articles

Missing Data Imputation for Multisite Rainfall Networks: A Comparison between Geostatistical Interpolation and Pattern-Based Estimation on Different Terrain Types

AbstractMissing rainfall data are a major limitation for distributed hydrological modeling and climate studies. Practitioners need reliable approaches that can be employed on a daily basis, often with too limited data in space to feed complex predictive models. In this study we compare different automatic approaches for missing data imputation, including geostatistical interpolation and pattern-based estimation algorithms. We introduce two pattern-based approaches based on the analysis of historical data patterns: (i) an iterative version of K-nearest neighbor (IKNN) and (ii) a new algorithm called vector sampling (VS) that combines concepts of multiple-point statistics and resampling. Both algorithms can draw estimations from variably incomplete data patterns, allowing the target dataset to be at the same time the training dataset. Tested on five case studies from Denmark, Australia, and Switzerland, the algorithms show a different performance that seems to be related to the terrain type: on flat terrains with spatially homogeneous rain events, geostatistical interpolation tends to minimize the average error, while in mountainous regions with nonstationary rainfall statistics, data mining can recover better the rainfall patterns. The VS algorithm, requiring minimal parameterization, turns out to be a convenient option for routine application on complex and poorly gauged terrains.

CLUSTERING PANEL DATA VIA PERTURBED ADAPTIVE SIMULATED ANNEALING AND GENETIC ALGORITHMS

Non-hierarchical cluster analysis for panel data is known to be hampered by structural preservation, computational complexity and efficiency, and dependency problems. Resolving these issues becomes increasingly important as efficient collection and maintenance of panel data make application more conducive. To address some computational issues and structural preservation, Bonzo [3] presented a stochastic version of Kosmelj and Batagelj's approach [16] to clustering panel data. The method used a probability link function (instead of the usual distance functions) in defining cluster inertias with the aim of preserving the clusters' probabilistic structure. Formulating clustering as an optimization problem, the objective function allows the application of heuristic and stochastic optimization techniques. In this paper, we present a modified heuristic for adaptive simulated annealing (ASA) by perturbing the state vector's sampling distribution, specifically, by perturbing the drift of a diffusion process. Such an approach has been used to hasten convergence towards global optimum at equilibrium for diversely complex, combinatorial, and large-scale systems. The perturbed ASA (PASA) heuristic is then embedded in a genetic algorithm (GA) procedure to hasten and improve the stochastic local search process. The PASA-GA hybrid can be further modified and improved such as by explicit parallel implementation.