The present study aims to identify the effective components for improving customer relations and increasing loyalty by analyzing customer purchasing behavior in electronic businesses. This research was conducted in two phases. In the first phase, we identified the influencing components for improving customer relationships and developed a conceptual framework consisting of structures, components, and sub-components that have a significant impact on enhancing customer relationships. The proposed model was then subjected to investigation and statistical analysis through a survey conducted by e-business experts. The results of the survey highlighted the considerable importance of "advertisements" in instant messengers as an effective means of communication and acquiring new customers. Moving on to the second phase of the research, we selected three electronic businesses (productive, wholesaler, and service) and calculated the scores of the components within the conceptual framework based on the average opinions of an eva‎luator group. By employing analytics and customer segmentation techniques, we identified the most influential component of customer loyalty using machine learning models. At each step, we identified the best structures with the highest F-measure in supervised machine learning models. Subsequently, we determined the best component within the selected structure and the sub-component within the best component. The implementation of the second phase demonstrated that the sub-components of "information analysis," "personalization of advertisements," and "quality of service after purchase" had the highest level of influence on customer loyalty, with acceptable F-measure scores, respectively, in data mining models. Additionally, we developed a novel segmentation clustering-based Model in the second scenario of the second phase. Through the utilization of various data mining methods, we achieved accurate predictions of the next customer visit time. Comparing the results with the conventional RFM segmentation model, our proposed prediction model exhibited an acceptable improvement in F-measure. Notably, for customers with inaccurate RFM predictions, the LP-FARM+ model demonstrated a remarkable enhancement. From a financial perspective, we also assessed the proposed model, revealing a Monetary Improvement Rate (MIR). To validate the model, we examined two benchmark datasets, further confirming its efficacy and reliability. Overall, this study provides valuable insights into improving customer relations and increasing loyalty in electronic businesses, offering practical implications for organizations aiming to gain a competitive advantage in the dynamic digital marketplace.

صبا صارمي نيا

علي زينل همداني , محسن شهرياري