Assessing the spatial sensitivity of a random forest model: Application in gridded population modeling

Sinha, Parmanand; Gaughan, Andrea E.; Stevens, Forrest R.; Nieves, Jeremiah J.; Sorichetta, Alessandro; Tatem, Andrew J.

doi:10.1016/j.compenvurbsys.2019.01.006

Cited by 85 publications

(56 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…RF is an ensemble learning method that has been widely used in different applications [59][60][61]. It is constructed by a large set of decision trees, with each tree being built using a random set of features and samples.…”

Section: Machine Learning Algorithms For Yield Predictionmentioning

confidence: 99%

Combining Multi-Source Data and Machine Learning Approaches to Predict Winter Wheat Yield in the Conterminous United States

et al. 2020

View full text Add to dashboard Cite

Winter wheat (Triticum aestivum L.) is one of the most important cereal crops, supplying essential food for the world population. Because the United States is a major producer and exporter of wheat to the world market, accurate and timely forecasting of wheat yield in the United States (U.S.) is fundamental to national crop management as well as global food security. Previous studies mainly have focused on developing empirical models using only satellite remote sensing images, while other yield determinants have not yet been adequately explored. In addition, these models are based on traditional statistical regression algorithms, while more advanced machine learning approaches have not been explored. This study used advanced machine learning algorithms to establish within-season yield prediction models for winter wheat using multi-source data to address these issues. Specifically, yield driving factors were extracted from four different data sources, including satellite images, climate data, soil maps, and historical yield records. Subsequently, two linear regression methods, including ordinary least square (OLS) and least absolute shrinkage and selection operator (LASSO), and four well-known machine learning methods, including support vector machine (SVM), random forest (RF), Adaptive Boosting (AdaBoost), and deep neural network (DNN), were applied and compared for estimating the county-level winter wheat yield in the Conterminous United States (CONUS) within the growing season. Our models were trained on data from 2008 to 2016 and evaluated on data from 2017 and 2018, with the results demonstrating that the machine learning approaches performed better than the linear regression models, with the best performance being achieved using the AdaBoost model (R2 = 0.86, RMSE = 0.51 t/ha, MAE = 0.39 t/ha). Additionally, the results showed that combining data from multiple sources outperformed single source satellite data, with the highest accuracy being obtained when the four data sources were all considered in the model development. Finally, the prediction accuracy was also evaluated against timeliness within the growing season, with reliable predictions (R2 > 0.84) being able to be achieved 2.5 months before the harvest when the multi-source data were combined.

show abstract

Section: Machine Learning Algorithms For Yield Predictionmentioning

confidence: 99%

Combining Multi-Source Data and Machine Learning Approaches to Predict Winter Wheat Yield in the Conterminous United States

et al. 2020

View full text Add to dashboard Cite

show abstract

“…In most Sub-Saharan African (SSA) cities, census data is hard to access or outdated. For regional or national assessments a lot of survey indicators have been produced at coarser resolutions [6,[43][44][45][46][47][48][49][50][51] but this was the first DHS fine-scale indicator production derived from VHR earth observation information directed specifically for intra-urban policy making and decision support. An additional highlight of this work is that not only were VHR variables able to train robust models based on DHS surveys, but also their predictions were in relative agreement with exhaustive census data at various geographical resolutions.…”

Section: Discussionmentioning

confidence: 99%

Modelling the Wealth Index of Demographic and Health Surveys within Cities Using Very High-Resolution Remotely Sensed Information

Georganos

Gadiaga²,

Linard³

et al. 2019

Remote Sensing

View full text Add to dashboard Cite

A systematic and precise understanding of urban socio-economic spatial inequalities in developing regions is needed to address global sustainability goals. At the intra-urban scale, access to detailed databases (i.e., a census) is often a difficult exercise. Geolocated surveys such as the Demographic and Health Surveys (DHS) are a rich alternative source of such information but can be challenging to interpolate at such a fine scale due to their spatial displacement, survey design and the lack of very high-resolution (VHR) predictor variables in these regions. In this paper, we employ satellite-derived VHR land-use/land-cover (LULC) datasets and couple them with the DHS Wealth Index (WI), a robust household wealth indicator, in order to provide city-scale wealth maps. We undertake several modelling approaches using a random forest regressor as the underlying algorithm and predict in several geographic administrative scales. We validate against an exhaustive census database available for the city of Dakar, Senegal. Our results show that the WI was modelled to a satisfactory degree when compared against census data even at very fine resolutions. These findings might assist local authorities and stakeholders in rigorous evidence-based decision making and facilitate the allocation of resources towards the most disadvantaged populations. Good practices for further developments are discussed with the aim of upscaling these findings at the global scale.

show abstract

“…Thus, the population model has two levels of additional information for the redistribution of non-point source estimates-(1) a pixel-level population density weighting from the RF model and (2) a subnational constraint from administrative unit totals. Ultimately, the reliability of the gridded population data is a function of the input population counts, which will vary on a country basis, the reliance of the use of ancillary variables with inherent error, and spatial grain of the subnational constraint applied in the model (Sinha et al 2019). There are multiple methods in the literature for modeling gridded population, including those datasets used in other gridded emissions studies (Leyk et al 2019).…”

Section: Discussionmentioning

confidence: 99%

Evaluating nighttime lights and population distribution as proxies for mapping anthropogenic CO₂emission in Vietnam, Cambodia and Laos

Gaughan

Oda

Sorichetta

et al. 2019

Environ. Res. Commun.

Self Cite

View full text Add to dashboard Cite

Tracking spatiotemporal changes in GHG emissions is key to successful implementation of the United Nations Framework Convention on Climate Change (UNFCCC). And while emission inventories often provide a robust tool to track emission trends at the country level, subnational emission estimates are often not reported or reports vary in robustness as the estimates are often dependent on the spatial modeling approach and ancillary data used to disaggregate the emission inventories. Assessing the errors and uncertainties of the subnational emission estimates is fundamentally challenging due to the lack of physical measurements at the subnational level. To begin addressing the current performance of modeled gridded CO 2 emissions, this study compares two common proxies used to disaggregate CO 2 emission estimates. We use a known gridded CO 2 model based on satelliteobserved nighttime light (NTL) data (Open Source Data Inventory for Anthropogenic CO 2 , ODIAC) and a gridded population dataset driven by a set of ancillary geospatial data. We examine the association at multiple spatial scales of these two datasets for three countries in Southeast Asia: Vietnam, Cambodia and Laos and characterize the spatiotemporal similarities and differences for 2000, 2005, and 2010. We specifically highlight areas of potential uncertainty in the ODIAC model, which relies on the single use of NTL data for disaggregation of the non-point emissions estimates. Results show, over time, how a NTL-based emissions disaggregation tends to concentrate CO 2 estimates in different ways than population-based estimates at the subnational level. We discuss important considerations in the disconnect between the two modeled datasets and argue that the spatial differences between data products can be useful to identify areas affected by the errors and uncertainties associated with the NTL-based downscaling in a region with uneven urbanization rates.

show abstract

Assessing the spatial sensitivity of a random forest model: Application in gridded population modeling

Cited by 85 publications

References 46 publications

Combining Multi-Source Data and Machine Learning Approaches to Predict Winter Wheat Yield in the Conterminous United States

Combining Multi-Source Data and Machine Learning Approaches to Predict Winter Wheat Yield in the Conterminous United States

Modelling the Wealth Index of Demographic and Health Surveys within Cities Using Very High-Resolution Remotely Sensed Information

Evaluating nighttime lights and population distribution as proxies for mapping anthropogenic CO₂emission in Vietnam, Cambodia and Laos

Contact Info

Product

Resources

About

Assessing the spatial sensitivity of a random forest model: Application in gridded population modeling

Cited by 85 publications

References 46 publications

Combining Multi-Source Data and Machine Learning Approaches to Predict Winter Wheat Yield in the Conterminous United States

Combining Multi-Source Data and Machine Learning Approaches to Predict Winter Wheat Yield in the Conterminous United States

Modelling the Wealth Index of Demographic and Health Surveys within Cities Using Very High-Resolution Remotely Sensed Information

Evaluating nighttime lights and population distribution as proxies for mapping anthropogenic CO2emission in Vietnam, Cambodia and Laos

Contact Info

Product

Resources

About

Evaluating nighttime lights and population distribution as proxies for mapping anthropogenic CO₂emission in Vietnam, Cambodia and Laos