PAFTA - Forecasting sales and economic demand for businesses with a time series approach

In forecasting time series, scientific forecasts should be made based on historical data. It involves building a model through historical analysis and using it to observe the future. It is not always an accurate forecast and the accuracy of forecasts can vary greatly. Especially when dealing with factors outside of our control. Often, the more comprehensive our data, the more accurate the predictions can be. Forecasting is used in various industries, including: weather forecasting, economic forecasting, engineering forecasting, health care forecasting, financial forecasting, retail forecasting, business forecasting, life sciences forecasting Environmental, social studies forecasting and more. Basically anyone who has fixed historical data can analyze that data with time series analysis methods and then model and forecast. The project is currently developed by game script libraries such as Plotly, Pandas, NeuralProphet, Streamlit, and PIL. This project is called PAFTA and runs on all operating systems in the web browser and does not require any special hardware. In this project, the NeuralProphet algorithm was used to predict the sales demand. The purpose of this project is to provide a user-friendly and easy-to-use cross-platform software to empower businesses to chart their future. Due to the accuracy achieved, it can be said that PAFTA can perform well in real world data and meet the forecasting needs of a small or large business with an intuitive and easy-to-use user interface.

کلیدواژه ها:

Sales forecasting ، time series ، deep learning ، neural networks