Leveraging Machine Learning for Predictive Analysis of Consumer Behavior in Marketing

In traditional statistical analysis, there are limitations in the study of complex consumer behavior often limited by model assumptions and analysis dimensions. In order to make up for the shortcomings of traditional methods, Machine learning (ML) is more capable of processing the characteristics of complex data and can more effectively analyze the potential patterns in business behaviors. This study is based on questionnaire survey data, which includes variables such as demographic characteristics, shopping frequency, and consumer experience dimensions. Through these dimensions and combined with machine learning algorithms, it is used to reveal the potential mechanism between the behavior patterns influenced by influencers and purchase intentions in the social media environment. The performance of Random Forest (RF) Regression is the best, with an Pearson Correlation Coefficient (R) of determination of about 0.7. In terms of error control, the Mean square error (MSE) was 0.696, the Root Mean square error (RMSE) was 0.8344. The final conclusion is that consumer behavior is positively correlated with variables related to customer experience, imitation behavior, information acceptance, and Cold IRQ.