نوع مقاله : مقاله پژوهشی
نویسندگان
1 گروه مهندسی کامپیوتر و فناوری اطلاعات، دانشکده فنی و مهندسی، دانشگاه قم، قم، ایران.
2 گروه فناوری اطلاعات ، دانشکده فناوری اطلاعات و مهندسی کامپیوتر، دانشگاه شهید مدنی آذربایجان، تبریز، ایران.
3 عضو هیات علمی گروه مهندسی کامپیوتر و فناوری اطلاعات، دانشکده فنی و مهندسی، دانشگاه قم، قم، ایران.
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Purpose: The objective of this study is to analyze customer behavior using multivariate time‑series data. After conducting a comprehensive analysis of customer behavior, customers are prioritized using the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) multi‑criteria decision‑making method. The results of this study can assist organizations in developing more effective and targeted marketing strategies.
Method: In this research, customers’ dynamic behavior is analyzed using the RFM (recency, frequency, monetary value) model represented as multivariate time‑series data. This approach is considered one of the most recent and practical methods for analyzing customer behavior over time. First, an ensemble time‑series clustering method is applied to identify customer clusters and analyze their behavioral patterns from different perspectives. Subsequently, key features are extracted from each time series and used as inputs to a classification model. Finally, the classifier model is interpreted using Shapley Additive Explanations (SHAP), which enables the calculation of the importance of each key feature. These key features and their corresponding weights are then used as inputs for the TOPSIS multi‑criteria decision‑making method to prioritize customers.
Findings: The findings indicate that representing customer purchase data as multivariate time‑series data based on RFM variables enables the identification of groups of customers with similar behavioral patterns over time. The proposed approach simultaneously considers RFM variables across time and provides a dynamic analysis of customer behavior. In addition, the application of the SHAP method allows for the interpretation of the classification model and the determination of the importance of critical behavioral features. Based on these weighted features, the TOPSIS method is applied to rank customers according to their importance and priority within marketing strategies. These results can support organizations in designing targeted and effective marketing initiatives.
Conclusion: The results demonstrate that the proposed approach enables the identification of customer behavioral patterns and facilitates more accurate customer prioritization. These analyses help organizations better understand customer dynamics and focus their marketing strategies on high‑value customers. Overall, the findings support organizations in improving marketing effectiveness and efficiency through a deeper understanding and analysis of customer behavior patterns
کلیدواژهها [English]
)
ارسال نظر در مورد این مقاله