The Implementation of a Prediction System for Sugarcane’s Destruction Rate From Sugarcane Stem Borer via Hybrid Machine Learning

Sugarcane stem borer poses a significant threat to sugarcane production, causing extensive damage and economic losses in the agricultural sector. Accurately predicting the destruction rate caused by this pest is crucial for implementing timely and effective management strategies. This paper presents the development and implementation of a time series prediction system designed to estimate the short-time destruction rate of sugarcane in the near future caused by stem borer infestations. The proposed Hybrid system combines Adaptive and Batch machine learning techniques, allowing for continuous model updates and adaptation to new data while maintaining initial accuracy, thereby improving predictive accuracy over time. A dataset comprising historical pest infestation data, manually recorded from the fields, and environmental factors, retrieved online, was used to train the model to capture complex temporal dependencies, with features such as temperature, humidity, seasonal patterns, and pest density being key predictors. The performance of the model was evaluated using the MAE, AIC, BIC, and HQIC, which demonstrated its robustness and reliability in predicting the destruction rate. This proposed Hybrid prediction system offers a valuable tool for sugarcane farmers and agronomists, enabling proactive decision-making to mitigate the impact of stem borer infestations. The findings of this study contribute to the broader application of machine learning in agricultural pest management, highlighting the potential for Hybrid learning approaches to enhance prediction accuracy in dynamic environments.