Predicting Vegetables And Fruits Through Supply Chain Insights

The prices of vegetables and fruits will change daily depending on many parameters like rainfall in various regions across the country, the demand for the commodity in the region, the amount of water required for the commodity and the amount water available in the area of cultivation, cost of transportation from the farmers to the supplier then to the retail shops then to the consumers, etc. This work explores the possibility of predicting the prices of these commodities using the rainfall data of the place which supplies the particular commodity the most. For example, Nagpur is India's leading supplier of oranges, so the rainfall data from this region is taken along with the price of oranges in a market situated in Goa to predict the price of oranges in the future. This work explores various machine-learning models for classification and regression. For classification KNN, K-means, SVM, decision tree, Random Forest, Naïve Bayes, Gradient Boost, with and without dimensionality reduction, and for regression, the models used are LSTM, ANN, Random forest, Decision tree, KNN, Gradient Boost, SVR, etc. The evaluation metrics used for classification are accuracy, F-1 score, and precision. MSE, MAE, and R-squared error are used as evaluation metrics for regression.