Identifying Field Attributes that Predict Soybean Yield Using Random Forest Analysis

• DOI: 10.2134/agronj2015.0222
• Publication Date: 2016-02-19T19:46:49Z
• Authors: Ethan R. Smidt; Shawn P. Conley; Jun Zhu; Francisco J. Arriaga
• Keywords: Soil test; Precision Agriculture

Until recently, soybean [ Glycine max (L.) Merr.] fields were often seeded at a single rate. Advances in GPS and variable rate technology (VRT) are allowing growers to use planting prescriptions optimize yields input costs. This study was conducted find the key predictors for characterizing seed yield from commonly collected precision agriculture data layers. Research 11 unique both 2013 2014 Wisconsin all 22 site‐years following corn Zea mays (L.)]. Seeding rate, soil sampling, yield, survey gathered each site analysis. A statistical procedure used determination of prediction parameters, random forest analysis, identified map unit as most important when predicting pooled sets 2014. The next factors were, order importance, P, organic matter, available water supply upper 100 cm, K, elevation elevation, 150 cm during Individual field analyses determined predictor on average, followed by pH pH, P Core Ideas Commonly accessed can vary depending scale. Soil K good Wisconsin. Random decision tree useful accurate analysis tools.