<![CDATA[In an era of fierce competition within the electronics industry, the ability to project market demand accurately has become a critical factor for companies when formulating marketing strategies, managing inventory, and making timely business decisions. This study develops a sales forecasting model for electronic equipment based on multiple linear regression, using historical sales data collected over several periods. The dataset comprises a dependent variable—sales realization—and two key independent variables: sales targets and the number of interested customers. The initial phase of the study involves data collection and cleansing to address missing values, duplicates, and anomalies. Numeric variables are normalized to ensure a uniform scale, and new features—such as squared terms and interaction variables—are created to enrich the model’s information. The dataset is then split into training and testing subsets in an 80:20 ratio to ensure the model’s generalizability. Regression coefficients are estimated using the ordinary least squares method, and model fit is evaluated using the coefficient of determination (R²), mean absolute error (MAE), and root mean squared error (RMSE). Analysis results indicate that the multiple linear regression model captures the relationships among variables effectively, as evidenced by an R² of 0.92—meaning 92% of the variance in sales realization can be explained by sales targets and customer interest. The MAE of 3.15 and RMSE of 4.27 suggest that prediction errors remain within acceptable bounds for business planning purposes. A t‑test on each coefficient yields p‑values below 0.05, confirming that both independent variables contribute significantly to sales realization. The final model is integrated into the company’s analytical framework as a quantitative tool for generating quarterly and annual sales forecasts. By using this predictive equation, management can simulate various scenarios involving changes in targets and levels of customer interest, allowing for more responsive strategic planning. Practical implications include optimized inventory control, precise scheduling of marketing campaigns, and targeted allocation of logistical resources. In conclusion, multiple linear regression proves to be an effective and reliable method for forecasting sales of electronic equipment and supports both operational and strategic decision‑making. This study opens avenues for further enhancement by incorporating external variables—such as seasonal promotions, competitor pricing, and macroeconomic factors—and by applying more advanced machine learning techniques to improve prediction accuracy in the future.]]>
Copyrights © 2025
