Dataset Information

Random forest as a generic framework for predictive modeling of spatial and spatio-temporal variables.

ABSTRACT: Random forest and similar Machine Learning techniques are already used to generate spatial predictions, but spatial location of points (geography) is often ignored in the modeling process. Spatial auto-correlation, especially if still existent in the cross-validation residuals, indicates that the predictions are maybe biased, and this is suboptimal. This paper presents a random forest for spatial predictions framework (RFsp) where buffer distances from observation points are used as explanatory variables, thus incorporating geographical proximity effects into the prediction process. The RFsp framework is illustrated with examples that use textbook datasets and apply spatial and spatio-temporal prediction to numeric, binary, categorical, multivariate and spatiotemporal variables. Performance of the RFsp framework is compared with the state-of-the-art kriging techniques using fivefold cross-validation with refitting. The results show that RFsp can obtain equally accurate and unbiased predictions as different versions of kriging. Advantages of using RFsp over kriging are that it needs no rigid statistical assumptions about the distribution and stationarity of the target variable, it is more flexible towards incorporating, combining and extending covariates of different types, and it possibly yields more informative maps characterizing the prediction error. RFsp appears to be especially attractive for building multivariate spatial prediction models that can be used as "knowledge engines" in various geoscience fields. Some disadvantages of RFsp are the exponentially growing computational intensity with increase of calibration data and covariates and the high sensitivity of predictions to input data quality. The key to the success of the RFsp framework might be the training data quality-especially quality of spatial sampling (to minimize extrapolation problems and any type of bias in data), and quality of model validation (to ensure that accuracy is not effected by overfitting). For many data sets, especially those with lower number of points and covariates and close-to-linear relationships, model-based geostatistics can still lead to more accurate predictions than RFsp.

SUBMITTER: Hengl T

PROVIDER: S-EPMC6119462 | biostudies-literature | 2018

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Random forest as a generic framework for predictive modeling of spatial and spatio-temporal variables.

Hengl Tomislav T Nussbaum Madlene M Wright Marvin N MN Heuvelink Gerard B M GBM Gräler Benedikt B

PeerJ 20180829

Random forest and similar Machine Learning techniques are already used to generate spatial predictions, but spatial location of points (geography) is often ignored in the modeling process. Spatial auto-correlation, especially if still existent in the cross-validation residuals, indicates that the predictions are maybe biased, and this is suboptimal. This paper presents a random forest for spatial predictions framework (RFsp) where buffer distances from observation points are used as explanatory ...[more]

PMID: 30186691

Similar Datasets

Project description:Spatially continuous predictions of seabed hardness are important baseline environmental information for sustainable management of Australia's marine jurisdiction. Seabed hardness is often inferred from multibeam backscatter data with unknown accuracy and can be inferred from underwater video footage at limited locations. In this study, we classified the seabed into four classes based on two new seabed hardness classification schemes (i.e., hard90 and hard70). We developed optimal predictive models to predict seabed hardness using random forest (RF) based on the point data of hardness classes and spatially continuous multibeam data. Five feature selection (FS) methods that are variable importance (VI), averaged variable importance (AVI), knowledge informed AVI (KIAVI), Boruta and regularized RF (RRF) were tested based on predictive accuracy. Effects of highly correlated, important and unimportant predictors on the accuracy of RF predictive models were examined. Finally, spatial predictions generated using the most accurate models were visually examined and analysed. This study confirmed that: 1) hard90 and hard70 are effective seabed hardness classification schemes; 2) seabed hardness of four classes can be predicted with a high degree of accuracy; 3) the typical approach used to pre-select predictive variables by excluding highly correlated variables needs to be re-examined; 4) the identification of the important and unimportant predictors provides useful guidelines for further improving predictive models; 5) FS methods select the most accurate predictive model(s) instead of the most parsimonious ones, and AVI and Boruta are recommended for future studies; and 6) RF is an effective modelling method with high predictive accuracy for multi-level categorical data and can be applied to 'small p and large n' problems in environmental sciences. Additionally, automated computational programs for AVI need to be developed to increase its computational efficiency and caution should be taken when applying filter FS methods in selecting predictive models.

Project description:A key challenge in island biogeography is to quantity the role of dispersal in shaping biodiversity patterns among the islands of a given archipelago. Here, we propose such a framework. Dispersal within oceanic archipelagos may be conceptualized as a spatio-temporal process dependent on: (1) the spatial distribution of islands, because the probability of successful dispersal is inversely related to the spatial distance between islands and (2) the chronological sequence of island formation that determines the directional asymmetry of dispersal (hypothesized to be predominantly from older to younger islands). From these premises, directional network models may be constructed, representing putative connections among islands. These models may be translated to eigenfunctions in order to be incorporated into statistical analysis. The framework was tested with 12 datasets from the Hawaii, Azores, and Canaries. The explanatory power of directional network models for explaining species composition patterns, assessed by the Jaccard dissimilarity index, was compared with simpler time-isolation models. The amount of variation explained by the network models ranged from 5.5% (for Coleoptera in Hawaii) to 60.2% (for Pteridophytes in Canary Islands). In relation to the four studied taxa, the variation explained by network models was higher for Pteridophytes in the three archipelagos. By the contrary, small fractions of explained variation were observed for Coleoptera (5.5%) and Araneae (8.6%) in Hawaii. Time-isolation models were, in general, not statistical significant and explained less variation than the equivalent directional network models for all the datasets. Directional network models provide a way for evaluating the spatio-temporal signature of species dispersal. The method allows building scenarios against which hypotheses about dispersal within archipelagos may be tested. The new framework may help to uncover the pathways via which species have colonized the islands of a given archipelago and to understand the origins of insular biodiversity.

Dataset Information

Random forest as a generic framework for predictive modeling of spatial and spatio-temporal variables.

Publications

Random forest as a generic framework for predictive modeling of spatial and spatio-temporal variables.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure