Predictive Analysis for Market Sales using Polynomial Regression Models

Document Type : Research Paper

Authors

Department of Statistics and Informatics, College of Administration & Economics, University of Salahaddin, Iraq.

Abstract

In this paper, we analyze sales data for two selected items from a supermarket and apply polynomial regression models to uncover insights into their sales trends. By fitting a polynomial curve to the data, we aim to understand how such pricing affects the number of sales of these items. This analysis compares the effectiveness of various regression models in predicting the price of products in different contexts. The linear regression model is the most effective for predicting the price of tomato paste, as it gives the data a better fit, with almost identical MSE and SEE, indicating similar prediction accuracy and errors. The linear regression model is the most effective for predicting the dependent variable, which is the number of sales (tomato paste) items, highlighting that the price of the item (independent variable) is statistically significant and provides a clear relationship with the dependent variable. In the context of sales, the number of oil items as the dependent variable prediction, the quadratic model is the most effective for explaining the price of oil items with both independent variables (price of power one and price of power two) being statistically significant, suggesting that the quadratic model best captures the connection between the two variables.

Keywords

Main Subjects

Kirkuk Journal of Science
Volume 20, Issue 3
September 2025
Page 78-86

Files

History

  • Receive Date: 09 April 2025
  • Revise Date: 03 August 2025
  • Accept Date: 04 August 2025

Share

Export Citation

Statistics