SALES PREDICTION OF VEGETABLE SEED PRODUCTS USING SIMPLE LINEAR REGRESSION

Abstract

The growth of the modern agricultural sector drives the need for an accurate sales prediction system, especially for vegetable seed products that are highly dependent on the season and market demand. An imbalance between stock and demand can cause losses, either in the form of overstock or undersupply. This condition requires a data-based planning strategy to ensure stock availability according to actual needs in the field. A historical data-based sales prediction approach is a relevant solution to optimize the distribution and procurement process. This study aims to apply a simple linear regression method in predicting vegetable seed sales based on historical data for one year. The prediction model is built using the time variable (month) as the independent variable and the number of seed requests as the dependent variable. This technique was chosen because of its ability to identify linear relationship patterns between time and sales trends in a simple but effective way. The data used comes from internal records of farmers and distributors, which are then classified into two main categories: leafy vegetable seeds (spinach, kale, mustard greens) and fruit vegetable seeds (tomatoes, chilies, eggplants). The results of the study showed that simple linear regression was able to provide fairly accurate predictive results. This model can be used as a basis for decision making in production planning, supply chain management, and seed inventory management, thus supporting the efficiency of farming businesses and reducing potential losses due to mismatches between demand and supply.