Application of machine learning in predicting consumer behavior and precision marketing.

Area of Science:

• Machine Learning
• Consumer Behavior Analysis
• Precision Marketing

Background:

• Increasing market competition and complex consumer behavior necessitate advanced customer identification and conversion rate optimization strategies.
• Accurate prediction of consumer purchasing intentions is crucial for effective marketing campaigns.

Purpose of the Study:

• To investigate the application of machine learning models for predicting consumer purchase intention.
• To compare the performance of Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and Backpropagation Artificial Neural Network (BPANN) in consumer behavior prediction.

Main Methods:

• Utilized four machine learning models: SVM, XGBoost, CatBoost, and BPANN.
• Evaluated model performance through experiments focusing on prediction accuracy, F1 scores, and ROC AUC.
• Conducted feature importance analysis to identify key drivers of purchasing behavior.

Main Results:

• CatBoost and XGBoost demonstrated superior prediction performance on complex features and large datasets, achieving F1 scores of 0.93 and 0.92, respectively.
• CatBoost achieved the highest ROC AUC of 0.985.
• SVM showed high accuracy but underperformed with large-scale data.

Conclusions:

• Machine learning models, particularly CatBoost and XGBoost, are effective tools for predicting consumer purchase intention and enabling precision marketing.
• Key features like page views and residence time significantly influence purchasing behavior.
• Model predictions can inform optimization strategies, including recommendation systems, dynamic pricing, and personalized advertising.