Support Vector Machines for Classification and Mapping of Reservoir Data

Kanevski, Mikhaïl; Pozdnukhov, A.; Canu, Stéphane; Maignan, M.; Wong, P.M.; Shibli, S.A.R.

doi:10.1007/978-3-7908-1807-9_21

“…Agriculture-related applications include Holmes et al (1998) for apple bruising, Yeates and Thomson (1996) for bull castration and venison analysis, and the Michalski and Chilausky's soybean disease diagnosis work (Michalski and Chilausky 1980), which is a classic benchmark problem in machine learning. Considerable efforts are recorded in the water-related fields, using rule-based reasoning (Zhu and Simpson, 1996;Dzeroski et al, 1997;Comas et al, 2003;Spate, 2005;Ramos-Martínez et al, 2014), decisiontrees (Kokotos et al, 2011), regression-trees (Dseroski et al, 2003), Support Vector Machines (SVM) (Kanevski et al, 2002), case-based reasoning (Martínez et al 2006 ;Wong et al, 2007), regression trees (Dzeroski and Drumm, 2003) or hybrid techniques (Cortés et al, 2002, Yang et al 2012. In the study of air quality, classification has been used for air quality data assurance issues (Athanasiadis and Mitkas, 2004) and the operational estimation of pollutant concentrations (Athanasiadis et al, 2003;Stebel et al, 2013;Yeganeh et al, 2012).…”

Section: Non-restrictive Propertiesmentioning

confidence: 99%

Which method to use? An assessment of data mining methods in Environmental Data Science

Gibert

¹

,

Izquierdo

²

,

Sànchez–Marrè

³

et al. 2018

Environmental Modelling & Software

View full text Add to dashboard Cite

Data Mining (DM) is a fundamental component of the Data Science process. Over recent years a huge library of DM algorithms has been developed to tackle a variety of problems in fields such as medical imaging and traffic analysis. Many DM techniques are far more flexible than more classical numerial simulation or statistical modelling approaches. These could be usefully applied to data-rich environmental problems. Certain techniques such as artificial neural networks, clustering, case-based reasoning or Bayesian networks have been applied in environmental modelling, while other methods, like support vector machines among others, have yet to be taken up on a wide scale. There is greater scope for many lesser known techniques to be applied in environmental research, with the potential to contribute to addressing some of the current open environmental challenges. However, selecting the best DM technique for a given environmental problem is not a simple decision, and there is a lack of guidelines and criteria that helps the data scientist and environmental scientists to ensure effective knowledge extraction from data. This paper provides a broad introduction to the use of DM in Data Science processes for environmental researchers. Data Science contains three main steps (pre-processing, data mining and post-processing). This paper provides a conceptualization of Environmental Systems and a conceptualization of DM methods, which are in the core step of the Data Science process. These two elements define a conceptual framework that is on the basis of a new methodology proposed for relating the characteristics of a given environmental problem with a family of Data Mining methods. The paper provides a general overview and guidelines of DM techniques to a non-expert user, who can decide with this support which is the more suitable technique to solve their problem at hand. The decision is related to the bidimensional relationship between the type of environmental system and the type of DM method. An illustrative two way table containing references for each pair Environmental System-Data Mining method is presented and discussed. Some examples of how the proposed methodology is used to support DM method selection are also presented, and challenges and future trends are identified.

show abstract

“…Better results were obtained by using mean values of σ kopt and C opt , for all N c classifiers. Using the same values of the hyperparameters for all the N c classifiers has been shown to give satisfactory results [43]. Furthermore, we used the same hyperparameters for the ten realizations, since they were all derived from a uniform thinning of the same rainfall data set.…”

Section: Applications To Real Data a Remotely Sensed Datamentioning

confidence: 99%

Classification of missing values in spatial data using spin models

Žukovič,

Hristopulos

2013

Preprint

0

View full text Add to dashboard Cite

A problem of current interest is the estimation of spatially distributed processes at locations where measurements are missing. Linear interpolation methods rely on the Gaussian assumption, which is often unrealistic in practice, or normalizing transformations, which are successful only for mild deviations from the Gaussian behavior. We propose to address the problem of missing values estimation on two-dimensional grids by means of spatial classification methods based on spin (Ising, Potts, clock) models. The "spin" variables provide an interval discretization of the process values, and the spatial correlations are captured in terms of interactions between the spins. The spins at the unmeasured locations are classified by means of the "energy matching" principle: the correlation energy of the entire grid (including prediction sites) is estimated from the sample-based correlations. We investigate the performance of the spin classifiers in terms of computational speed, misclassification rate, class histogram and spatial correlations reproduction using simulated realizations of spatial random fields, real rainfall data, and a digital test image. We also compare the spin-based methods with standard classifiers such as the k-nearest neighbor, the fuzzy k-nearest neighbor, and the Support Vector Machine. We find that the spin-based classifiers provide competitive choices.

show abstract

“…The values are in millimeters and some summary statistics are as follows: z min = 7.1, shown to give satisfactory results [43]. Furthermore, we used the same hyperparameters for the ten realizations, since they were all derived from a uniform thinning of the same rainfall data set.…”

Section: Applications To Real Data a Remotely Sensed Datamentioning

confidence: 99%

Classification of missing values in spatial data using spin models

Žukovič¹,

Hristopulos²

2009

View full text Add to dashboard Cite

A problem of current interest is the estimation of spatially distributed processes at locations where measurements are missing. Linear interpolation methods rely on the Gaussian assumption, which is often unrealistic in practice, or normalizing transformations, which are successful only for mild deviations from the Gaussian behavior. We propose to address the problem of missing values estimation on two-dimensional grids by means of spatial classification methods based on spin (Ising, Potts, clock) models. The "spin" variables provide an interval discretization of the process values, and the spatial correlations are captured in terms of interactions between the spins. The spins at the unmeasured locations are classified by means of the "energy matching" principle: the correlation energy of the entire grid (including prediction sites) is estimated from the sample-based correlations. We investigate the performance of the spin classifiers in terms of computational speed, misclassification rate, class histogram and spatial correlations reproduction using simulated realizations of spatial random fields, real rainfall data, and a digital test image. We also compare the spin-based methods with standard classifiers such as the k-nearest neighbor, the fuzzy k-nearest neighbor, and the Support Vector Machine. We find that the spin-based classifiers provide competitive choices.

show abstract

Support Vector Machines for Classification and Mapping of Reservoir Data

Cited by 8 publications

References 4 publications

Which method to use? An assessment of data mining methods in Environmental Data Science

Which method to use? An assessment of data mining methods in Environmental Data Science

Classification of missing values in spatial data using spin models

Classification of missing values in spatial data using spin models

Contact Info

Product

Resources

About