Abstract:
Consumer purchasing is increasingly influenced by digital interactions and
data-driven marketing strategies. This thesis investigates the efficacy of four
classification techniques: Decision Tree, Naïve Bayes, Support Vector Machine
(SVM) and Random Forest Algorithm in customer purchase intentions using the
“Shopping Trends and Customer Behavior” dataset. These algorithms learn the
relationship between input attributes and target classes during the training phase and
then apply this model to new data to predict its class.
The comparative analysis showed that the Random Forest model achieved the
highest performance across all metrics, confirming its robustness in handling mixedtype data. Although SVM also performed well and slightly less accurate when data
was standardized; Naïve Bayes was faster. Decision Tree provided good
interpretability but was more prone to overfitting. Predicting consumer purchasing
behavior is crucial for businesses to optimize marketing strategies, manage inventory,
and enhance overall efficiency in the competitive e-commerce landscape. This study
conducts a comparative analysis of four prominent machine learning (ML)
algorithms—Decision Tree, Naïve Bayes, Support Vector Machines (SVM), and
Random Forest—to determine their efficacy in modeling and forecasting consumer
purchasing trends, using a comprehensive dataset of shopping behaviors. Each
algorithm is applied to the dataset, and their performance is rigorously assessed, using
key metrics including accuracy, precision, recall, and F1-score.
This analysis highlights the critical role of advanced ML technologies in
refining strategic decision-making and offers practical insights into selecting the most
appropriate model for predicting consumer purchasing patterns based on specific
business needs and data characteristics.