Predicting Average Tomato Prices Using Support Vector Machine with Polynomial Features

Every economy depends on agriculture, and research in this area is primarily focused on boosting food supply and quality while minimizing costs and maximizing profit. Numerous factors can have an impact on tomato pricing, which serve as a crucial indicator of market trends for vegetables. Making educated decisions could be facilitated by predicting tomato average prices. To forecast the typical prices of tomatoes, we used the Support Vector Machine (SVM) model together with polynomial feature modification in this study. For our projections, we used the ‘Minimum’ and ‘Maximum’ pricing attributes. To guarantee consistency and improved performance, the dataset was standardized. With crossvalidation, we used a Grid Search technique to refine the parameters of our SVM model. The Mean Squared Error (MSE) and R2 score metrics were used to assess the model's performance. Our findings were shown graphically by scatter plots that contrasted real and forecasted prices and a histogram that showed the prediction errors. The SVM model has the potential to forecast average tomato prices when used in conjunction with polynomial feature transformation and parameter optimization. The agricultural industry's stakeholders may use this technique to gather knowledge of price trends and guide their decisions. More research may be done to increase prediction accuracy by incorporating more features and using other algorithm.