Estimation of Harvest Time of Forage Sorghum (Sorghum Bicolor) CV. Samurai-2 Using Decision Tree Algorithm

Abstract

Efforts to improve feed quality by adding additional nutritional supplements can increase production costs due to the increased concentrate prices. Therefore, one option is to combine the main feed with forages containing a high protein source at a low cost, such as Gramineae (e.g., sorghum). This study aims to estimate the harvest time of sorghum when the biomass content, nutrients, and digestibility for livestock are in good condition using a machine learning algorithm, namely a decision tree. The stages of this study include the collection of observation data in the field, preprocessing, modeling, evaluation, and validation. Images and field observations are the primary datasets used. These datasets become the model input for the decision tree algorithm. The results of this study are the classification model for estimating harvest time with an accuracy of 98.86% and the rule that is generated by the decision tree model, the right time to be harvested are in the condition (Day After Planting > 77.5 days AND Day After Planting ≤ 84 days AND Diameter > 26 mm) or (Day After Planting > 84 days AND Height ≤ 138.5 cm AND Leaves > 8.5 pieces) or (Day After Planting > 84 days AND Height > 138.5 cm). In conclusion, the rule generated from the decision tree algorithm can help estimate the fast harvest time of sorghum bicolor cv. Samurai 2.

Influence of Membership Function and Degree on Sorghum Growth Prediction Models in Machine Learning

“…Many further research studies within the sorghum domain utilize machine learning techniques. These encompass efforts to predict sorghum biomass [12], detect and measure sorghum head counts [13], and make estimations of sorghum crop yields through machine learning algorithms [14] [15]. In combination, these investigations underscore the versatility and potential of machine learning in advancing various aspects of sorghum farming and administration, promoting more effective and sustainable agricultural practices.…”