Smart Agriculture Economics and Engineering: Unveiling the Innovation Behind AI-Enhanced Rice Farming

• Main Authors: Zun Liang, Chuan, Tham, Ren Sheng, Tan, Chek Cheng, Abraham Lim, Bing Sern, David Lau, King Luen, Chong, Yeh Sai
• Format: Conference or Workshop Item
• Language: English
• Published: 2024
• Subjects: HA Statistics; QA Mathematics; SB Plant culture; T Technology (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/42277/1/IIICe2024%20%28E1-2%29-Smart%20Agriculture%20Economics%20and%20Engineering%20%28Poster%29.pdf; http://umpir.ump.edu.my/id/eprint/42277/

The occurrence of climate change, exponential population growth, global food inflation, technological advancements, and national and international social-environmental stressors have increased the uncertainty of food security globally, including in Southeast Asia. These factors have directly and indirectly caused challenges to Sustainable Development Goals (SDGs), including no poverty (SDG1), zero hunger (SDG2), and good health and well-being (SDG3). Moreover, rice cultivation provides employment and livelihoods for millions of individuals, directly relating to decent work and economic growth (SDG8). Specifically, the Food and Agriculture Organization (FAO) of the United Nations (2023) reported a consistent upward trajectory in the 3-year average prevalence of severe and moderate food insecurity (in total population) across low- and upper-middle-income nations in the Southeast Asia region. To address these challenges, an innovative Artificial Intelligence-based (AI-based) predictive algorithm has been proposed, leveraging the Cross Industry Standard Process for Data Mining (CRISP-DM) data science framework. This predictive algorithm is designed specifically for short-term rice production prediction and stood out by seamlessly integrating features that addressed both availability, accessibility, and stability dimensions of food security, further identifying key determinants impacting rice production, categorized into three clusters: atmospheric, socioeconomic, and farming practices. The analysis utilized a novel modified stacked Multiple Linear Regression- -Support Vector Regression (MLR- -SVR) algorithm, and a novel modified stacked MLR- -Support Vector Regression (MLR- -SVR) algorithm, demonstrating high predictive capability, especially in a limited dataset environment, which the algorithms’ superiority ranked utilizing modified Taguchi-based VIseKriterijumska Optimizacija I Kompromisno Resenje (Taguchi-based VIKOR) multi-criteria decision-making algorithm. Subsequently, the selected superior modified stacked ensemble MLR-SVR-based algorithms are utilized to forecast the 5-year future rice production for each low-middle and upper-middle Southeast Asia nation. The results generally showed reasonable forecasting outcomes across the region, with the exception of Cambodia (KHM). This article holds significant potential for academics and industries, particularly agriculture, food production, data analytics, and technology. The algorithm's deployment can improve efficiency and productivity in agricultural operations, inform decision-making processes across industries, and spur innovation. Furthermore, the insights generated from this research have implications for shaping policies and regulations related to food security and agricultural sustainability.