Applications of Supervised Algorithms for Sales Prediction in Small Business - Santo Domingo, Ecuador

Small and medium businesses in Ecuador contribute greatly to the generation of employment and important contributions to the economy in general of the country. The objective of the following investigation is the application of supervised algorithms for the prediction of sales in a small business minimarket located in the province of Santo Domingo de Los Tsáchilas. The research methodology follows a predictive quantitative approach, through the use of ARIMA time series, with a non-experimental and cross-sectional design, having as available information 1,825 sales dynamics that left the business during the years 2018 to 2022, obtaining a financial sample of 60 observations. The partial results allow to visualize a non-stationary time series with seasonality and became stationary with a difference of moving means. The final predictive model was an ARIMA (0, 1, 0) (1, 0, 1). As a general conclusion of this investigation, it was obtained that the sales forecasts in the small business are simultaneous in the timeline, evidencing seasonality and a sales growth trend of more than 3%.