چکیده:
One of the important issues in the problem of dynamic customer segmentation is the transfer of customers to different segments over time and discovering the patterns governing these movements. Accordingly, this article focuses on customer dynamics and attempts to extract customer behavior groups, dominant characteristics of these groups, and general patterns governing the displacement and migration of customers to different segments over time. To do this, a new combined method based on the K-means algorithm, hierarchical clustering methods, and association rules is presented and applied to real data from a telecommunications company. Based on the obtained results, there are seven different behavioral groups in the transfer of customers to different segments. Also, in a novel approach, an attempt has been made to explain the impact of dynamic customer behavior on changes in segments over time. In this regard, by presenting new approaches and concepts regarding customer behavior dynamics and its impact on structural and content changes in segments, a new grouping of customers is presented in the form of structure-building and stabilizing customers, adaptable structure-dynamic customers, and structure-breaking dynamic customers.
خلاصه ماشینی:
Based on this, this article focuses on customer dynamics and attempts to extract customer behavior groups, the dominant characteristics of these groups, and general patterns governing the displacement and migration of customers to different segments over time.
In previous limited studies, systematic and quantitative methods have not been used to extract dominant patterns, and these patterns have been extracted by counting existing sequences and selecting their maximum frequency, but in this research, data mining methods2 are used and an effort is made to extract customer behavioral groups and general patterns governing the movement and migration of customers to different segments over time, and the dominant characteristics of each of these groups are also investigated and analyzed.
The proposed method includes extracting individual customer sequences in membership to different segments over time and clustering these sequences and discovering the dominant characteristics of each group using association rules.
In fact, this stage of the research seeks to answer the question of how the dynamics of the customer and different customer behavioral groups affect the structural and content changes of segments over time?
To analyze the data and extract patterns governing the movement of customers between different parts over time, data mining methods are used, and a combined method based on the K-means algorithm, hierarchical clustering methods, and association rules is presented.
Conclusion and Suggestions In this article, a new combined method based on K-means clustering, hierarchical methods, and association rules was presented to identify behavioral groups of customers in membership to different segments over time and also to analyze the dominant characteristics of these groups.
(
, 
, 
)
