OPTIMIZING ADABOOST FOR CROP YIELD FORECASTING: A PCA-BASED DIMENSIONALITY REDUCTION APPROACH

Abstract

Agriculture is the essential human function that makes sure human existence helps various nations’ economies, confrontation troubles because of the weather conditions are increasingly unpredictable moreover the soil experiences composite procedures. Through precision farming and predictive analytics, machine learning in agriculture improves crop yields and resource efficiency. Unfailingly predicting upcoming crop yield is essential for cleanse agronomic creation and guaranteeing nutrients protection. Our study deliberates an innovative MLM (machine learning model) employing PCA (Principal Component Analysis) with AdaBoost classifier. The dataset, which focuses on characteristics like soil nitrogen, phosphorus, potassium, temperature, humidity, pH, and rainfall, comes from an open-source repository (Kaggle) and includes 2200 entities of 22 different crops. The MLM is employed to examine the diverse dimensional statistics and ensure precise crop choice sanctions. The PCA is used in preprocessing for the removal of noise data and feature extraction, AdaBoost algorithm was used for classification. The results of the experiment demonstrate a high accuracy of 99.2% and consistent precision, recall, and F1-scores of 0.992.  All of these results suggest that the MLM (Machine Learning Model) can distinguish between different crop varieties with a low error rate.  Based on these results, the researchers can conclude that Principal Component Analysis (PCA) and the ensemble machine learning algorithm AdaBoost can improve the forecasting capabilities of DM (data mining) processes in agriculture and provide information to agronomists and former for sustainable crop production.