Forecasting Corn Yield With Machine Learning Ensembles

Abstract

The emergence of new technologies to synthesize and analyze big data with highperformance computing has increased our capacity to more accurately predict crop yields. Recent research has shown that machine learning (ML) can provide reasonable predictions faster and with higher flexibility compared to simulation crop modeling. However, a single machine learning model can be outperformed by a "committee" of models (machine learning ensembles) that can reduce prediction bias, variance, or both and is able to better capture the underlying distribution of the data. Yet, there are many aspects to be investigated with regard to prediction accuracy, time of the prediction, and scale. The earlier the prediction during the growing season the better, but this has not been thoroughly investigated as previous studies considered all data available to predict yields. This paper provides a machine leaning based framework to forecast corn yields in three US Corn Belt states (Illinois, Indiana, and Iowa) considering complete and partial inseason weather knowledge. Several ensemble models are designed using blocked sequential procedure to generate out-of-bag predictions. The forecasts are made in county-level scale and aggregated for agricultural district and state level scales. Results show that the proposed optimized weighted ensemble and the average ensemble are the most precise models with RRMSE of 9.5%. Stacked LASSO makes the least biased predictions (MBE of 53 kg/ha), while other ensemble models also outperformed the base learners in terms of bias. On the contrary, although random k-fold cross-validation is replaced by blocked sequential procedure, it is shown that stacked ensembles perform not as good as weighted ensemble models for time series data sets as they require the data to be non-IID to perform favorably. Comparing our proposed model forecasts with the literature demonstrates the acceptable performance of forecasts made by our proposed ensemble model. Results from the scenario of having partial in-season weather knowledge reveals that decent yield forecasts with RRMSE of 9.2% can be made as early as June 1 st. Moreover, it was shown that the proposed model performed better than individual models and benchmark ensembles at agricultural district and statelevel scales as well as county-level scale. To find the marginal effect of each input feature on the forecasts made by the proposed ensemble model, a methodology is suggested that is the basis for finding feature importance for the ensemble model. The findings

Coupling machine learning and crop modeling improves crop yield prediction in the US Corn Belt

“…Several two-level stacking ensembles, namely stacked regression, stacked LASSO, stacked random forest, and stacked LightGBM, were built, which are expected to demonstrate excellent performance. The details of each model can be found at Shahhosseini et al 61 .…”

A Comprehensive Review of Crop Yield Prediction Using Machine Learning Approaches With Special Emphasis on Palm Oil Yield Prediction

“…Irrigation ratio [46], number of open wells (OW) [26] [41], number of tanks (TK) [ [40], GDD [24], growing degree days [46], killing degree days [46], day length [51] [13] [43], snow water [13], drought index (DI) [ [66], GCVI [50], VOD [46] [62], RVI [63] [39], GNDVI [63] [39], GRVI [63] [39], EVI2 [63], OSAVI [63] [39], WDRVI [63] [39], NDVIre [64], TSAVI [39], IPVI [39], MSAVI [39], GI [39], PVI [39], SAVI [39]], GESAVI [39], GLAI [39], CWSI [39], NDWI [39], GVI [39] LAI [42] [40] [65], FPAR [42], GPP [42], NIRv [56] [47], CDL [42], cropland census [42], satellite images from the landsat thematic mapper (TM) [42], satellite images from advanced wide field sensor (AWIFS) [42], empty-land [45], harrowed land [45], texture conditions [48], PVI [48].…”

“…Weekly cumulative percentage of planted fields [25], fertilizer usage [26], seed quantity [26] [45] [41], vegetative growth [45] [27], flowering [45], maturity [45], crop genotype [51], Plant population [13], planting progress [13] [59], power of agricultural machinery [50], the electricity consumed in rural areas [50]. Historical yield data [42] [42], transmissivity [42], rock layer permeability [42], water conductivity [42], and the number of micronutrients [42] and hydrochemical analysis [42].…”

“…Shahhosseini et al [12] evaluated computer simulation models based on MLR and RF for the yield prediction and nitrate losses. Based on the analysis, Shahhosseini et al [13] suggested that optimized weighted ensemble and the average ensemble are the most reliable and accurate techniques. However, the efficiency of ensemble modelling is affected by the selection of diverse ML models; therefore, it needs to be investigated.…”

“…Previously, farmers depended on their experiences and authentic historical information for crop yield prediction and settle on the significant cultivation decisions according to the prediction. Nevertheless, in recent years, the advancement of new innovations, including crop model simulation and machine learning, has appeared to predict yield more precisely, along with the capacity to analyze a huge amount of data using high-performance computing [10] [11] [12] [13]. Numerous investigations are presently exhibiting relatively higher potential in the use of machine learning algorithms than traditional statistics [4] [14] [15].…”