Predictive Analytics in Digital Marketing: A Statistical Modeling Approach for Predicting Consumer Behavior

Introduction: The evolution of predictive analytics in digital marketing is deeply rooted in the development of statistical modeling and data analytics. Aim: The aim of the present research was to analyze the use of advanced statistical models for predicting consumer behavior in digital marketing environments, highlighting the relevance of predictive analytics in data- driven strategic decision making. Methodology: five machine learning, logistic regression, decision tree, random forest, support vector machines (SVM) and neural networks models were evaluated on a synthetic dataset representative of digital consumers belonging to Generation Z. The analysis considered key metrics such as overall accuracy, cross-validation mean and standard deviation, in order to measure both the effectiveness and stability of each model. Results: The results showed that the logistic regression, Random Forest, SVM and neural network models achieved an accuracy of 97% with overall consistency (standard deviation of 0.0), positioning them as reliable tools for predicting consumption trends. In contrast, the decision tree showed lower accuracy (92%) and higher variability, which limits its applicability in complex scenarios. Conclusion: The study concludes that the combination of accuracy and stability is essential for the implementation of effective predictive models in digital marketing and also highlights the importance of integrating these models into campaign automation and personalization systems to anticipate preferences, improve customer experience and optimize resources.