A Dataset on Corn Silage in China Used to Establish a Prediction Model Showing Variation in Nutrient Composition

It is important to assess the nutritional concentrations of forage before it can be used for tremendous improvements in dairy industry. This improvement requires a rapid, accurate, and portable method detecting nutrient values, instead traditional laboratory analysis. Fourier-transform infrared (ATR-FTIR) spectroscopy technology was applied, partial least squares regression (PLSR) backpropagation artificial neural network (BP-ANN) algorithms were current study. The objective this study estimate discrepancy content rumen degradation WPCS grown various regions propose novel analytical predicting whole plant corn silage (WPCS). Zhengdan 958 cultivar selected from eight different sites compare discrepancies nutrients degradation. A total 974 samples 235 farms scattered across five Chinese collected, indicators modeled. As result, substantial found when they cultivated growing regions. Wuxi showed 1.14% higher dry matter (DM) than that Jinan. Lanzhou had 11.57% 8.25% lower neutral detergent fiber (NDF) acid (ADF) Jinan, respectively. DM degradability planted Bayannur considerably Jinan (6 h degradability: vs. = 49.85% 33.96%), starch (71.79%) also highest after 6 rumen. results indicated contents NDF, ADF, estimated precisely based on ATR-FTIR combined with PLSR or BP-ANN algorithm (R2 ≥ 0.91). followed by crude protein (CP), (0.82 ≤ R2 0.90), ether extract (EE), ash (0.66 0.81). prediction performance (R2PLSR R2BP-ANN; RMSEPLSR RMSEBP-ANN). same could effectively evaluate CP, WPCS.