The derived demand for advertising expenses and implications on sustainability: a comparative study using deep...

Area of Science:

• Marketing Analytics
• Data Science
• Business Intelligence

Background:

• Machine learning models are increasingly used in marketing for data-driven insights and strategic decision-making.
• Existing research on demand forecasting often overlooks the impact of marketing strategies like advertising on demand.
• Accurate demand forecasting is crucial for business sustainability and strategic planning.

Purpose of the Study:

• To evaluate various machine learning methods for accurate demand forecasting.
• To specifically assess the impact of advertising expenses on demand prediction.
• To compare the performance of different regression and deep learning techniques.

Main Methods:

• Utilized machine learning techniques: Support Vector Regression (SVR), Random Forest Regression (RFR), and Decision Tree Regressor (DTR).
• Employed deep learning techniques: Artificial Neural Network (ANN) and Long Short Term Memory (LSTM).
• Analyzed a real-world dataset from a television manufacturer, including advertising expenditures and sales data.

Main Results:

• Compared the accuracy of demand forecasting across all tested machine learning and deep learning models.
• Identified Long Short Term Memory (LSTM) as the most accurate model for demand forecasting.
• Demonstrated the effectiveness of deep learning in solving complex marketing regression problems.

Conclusions:

• Long Short Term Memory (LSTM) significantly outperforms other models in demand forecasting accuracy when considering advertising expenses.
• The study highlights the potential of advanced machine learning techniques to improve marketing strategy and business outcomes.
• Accurate demand forecasting driven by marketing data can lead to enhanced business sustainability.