K-NN FOREST: a software for the non-parametric prediction and mapping of environmental variables by the<i>k</i>-Nearest Neighbors algorithm

Chirici, Gherardo; Corona, Piermaria; Marchetti, Marco; Mastronardi, Alessandro; Maselli, Fabio; Bottai, Lorenzo; Travaglini, Davide

doi:10.5721/eujrs20124536

Cited by 17 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Variable selection was beyond the scope of this study, but it remains a complex challenge (McRoberts 2009;Chirici et al 2012;Packalén et al 2012) due to the curse of dimensionality, which is of particular relevance to the kNN technique (Hastie et al 2005, ch. 2.5).…”

Section: Discussionmentioning

confidence: 98%

See 1 more Smart Citation

Thek-nearest neighbor technique with local linear regression

Magnussen

Tomppo

2014

Scandinavian Journal of Forest Research

View full text Add to dashboard Cite

In a standard k-nearest neighbor (kNN) technique, imputations of unit-level values in the variables of interest (Y) are based on the k-nearest neighbors in a set of reference units. Nearest is defined with respect to a distance metric in the space of auxiliary variables (X). This study evaluates kNN imputations of Y with a selection, by the same distance metric, of k-nearest locally weighted regression models. Imputations are obtained as predictions using the X values of the k-nearest neighbors in the population. In simulated random sampling from three artificial multivariate populations and two actual univariate populations and sampling units composed of a single population element or a cluster of four elements, the new kNN technique: (1) improved the correlation between an imputation and its actual value; (2) lowered the root mean square error (RMSE) of imputations; (3) increased the slope in regressions of actual y values regressed against their imputed values; (4) performed relatively best with k values of 4 and sample sizes of 200 or greater; (5) compared favorably with a recently proposed kNN calibration procedure; and (6) had a higher (15-28%) RMSE than with a simple local linear regression. Distribution matching had a consistent negative effect (+10%) on RMSE.

show abstract

Section: Discussionmentioning

confidence: 98%

“…The k-nearest neighbor technique (kNN) has become a popular and easy-to-implement method for multivariate mapping (Chirici et al 2012). In kNN, the values of one or more target variables (Y) are imputed for elements (e.g.…”

Section: Introductionmentioning

confidence: 99%

Thek-nearest neighbor technique with local linear regression

Magnussen

Tomppo

2014

Scandinavian Journal of Forest Research

View full text Add to dashboard Cite

show abstract

“…Here, the target set was made of the IRS LISS III pixels belonging to the even-aged forest area (128 402 ha corresponding to 3 210 050 pixels), and the reference was 304 pixels, belonging to the field plots of the local forest inventory. The estimates were calculated using the K-NN FOREST free software (Chirici et al 2012b). We tested three different distance measures implemented within the K-NN FOREST software (Euclidean, Mahalanobis and Fuzzy), with k values ranging from 1 to 10 based on the averaged spectral values of a 3 × 3 pixels area surrounding the field plots .…”

Section: Iforest -Biogeosciences and Forestrymentioning

confidence: 99%

“…We tested three different distance measures implemented within the K-NN FOREST software (Euclidean, Mahalanobis and Fuzzy), with k values ranging from 1 to 10 based on the averaged spectral values of a 3 × 3 pixels area surrounding the field plots . The Leave-One-Out (LOO) approach (Fazakas et al 1999, Chirici et al 2012b) was used to test several k-NN configuration, achieving the most accurate estimation using the Euclidean distance with k= 6. For a more detailed description of the k-NN algorithm and the assumptions and the implications related to its use, we refer to the vast bibliography available , Baffetta et al 2009, McRoberts 2009).…”

Section: Iforest -Biogeosciences and Forestrymentioning

confidence: 99%

Spatially explicit estimation of forest age by integrating remotely sensed data and inverse yield modeling techniques

Frate¹,

Carranza²,

Garfì³

et al. 2016

iForest

View full text Add to dashboard Cite

We present an innovative method based on the application of inverse yield models for producing spatially explicit estimations of forest age. Firstly, a raster growing stock volume map was produced using the non-parametric kNearest Neighbors estimation method on the basis of IRS LISS-III remotely sensed imagery and field data collected in the framework of a local forest inventory. Secondly, species-specific inverted yield equations were applied to estimate forest age as a function of growing stock volume. The method was tested in 128 000 ha of even-aged forests in central Italy (Molise region). The accuracy of the method was assessed using an independent dataset of 305 units from a local standwise forest inventory. The results demonstrated that the forest age map was accurate, with a root mean square error of 15.8 years (30% of the mean of field values), and useful for supporting forest management purposes, such as the assessment of harvesting potential and ecosystem services. Thanks to the use of remotely sensed data and spatial modeling, the proposed approach is cost-effective and easily replicable for vast regions.

show abstract

“…Classes were defined through a set of rules and organized in hierarchical groups, so that a child class inherited properties from the parent one. The classification process was carried out either by specifying thresholds for each rule (crisp classification) [Comber et al, 2012], by specifying a set of probability density functions (fuzzy classification) [HongLei et al, 2013] or through a k-Nearest Neighbour (NN) approach [Chirici et al, 2012]. Figure 5 shows the workflow.…”

Section: Image Classificationmentioning

confidence: 99%

Hierarchical classification of complex landscape with VHR pan-sharpened satellite data and OBIA techniques

Gianinetto

Rusmini²,

Candiani

et al. 2014

European Journal of Remote Sensing

View full text Add to dashboard Cite

Land-cover/land-use thematic maps are a major need in urban and country planning. This paper demonstrates the capabilities of Object Based Image Analysis in multi-scale thematic classification of a complex sub-urban landscape with simultaneous presence of agricultural, residential and industrial areas using pan-sharpened very high resolution satellite imagery. The classification process was carried out step by step through the creation of different hierarchical segmentation levels and exploiting spectral, geometric and relational features. The framework returned a detailed land-cover/land-use map with a Cohen's kappa coefficient of 0.84 and an overall accuracy of 85%.

show abstract

K-NN FOREST: a software for the non-parametric prediction and mapping of environmental variables by thek-Nearest Neighbors algorithm

Cited by 17 publications

References 27 publications

Thek-nearest neighbor technique with local linear regression

Thek-nearest neighbor technique with local linear regression

Spatially explicit estimation of forest age by integrating remotely sensed data and inverse yield modeling techniques

Hierarchical classification of complex landscape with VHR pan-sharpened satellite data and OBIA techniques

Contact Info

Product

Resources

About