ABSTRACT
Living in a sharply competitive telecom market, customers are provided with a great variety of promotions of telecom offers which are excessive and complex in recent years. Not only customers have no idea how to choose the suitable promotions but also front-line sales cannot recommend suitable promotions depending on merely the traditional recommendation method, namely, their personal intuition and experiences. In this way, promotion recommendation lacks mobile data usage pattern and consumption level of customers so that telecom operators hardly provide precision marketing. Conclusively, it may be possibly an inappropriate telecom offer, resulting in low customer satisfaction and loyalty, high churn rate, or in the worst case, erosion of average revenue per user (ARPU). Therefore, this study proposes a promotion recommendation method and system based on random forest to analyze the customer profiles and historical mobile data usage. Then the marketing information can be obtained for front-line sales to help make precise marketing strategies and recommending promotions in accordance with these customer features. Eventually, customer satisfaction, customer loyalty and the income of telecom operators can be increasing. In experimental results, more than 500 thousand mobile data usage records of customers in Chunghwa Telecom from January to March in 2017 were collected and analyzed for the evaluation of the proposed method. The accuracy of the proposed method is 93.36% which is higher than the traditional method. It also gains an advantage over other three popular classification algorithms for recommendation.
- Ho, Tin Kam, "Random Decision Forests", Proceedings of the 3rd International Conference on Document Analysis and Recognition, Montreal, QC, 14-16 August 1995. pp. 278--282. Google ScholarDigital Library
- K. Yoshii, M. Goto, K. Komatani, T. Ogata, H. G. Okuno "Hybrid Collaborative and Content-based Music Recommendation Using Probabilistic Model with Latent User Preferences," ISMIR, 2006.Google Scholar
- Strobl C, Boulesteix AL, Zeileis A, Hothorn T., "Bias in random forest variable importance measures: illustrations, sources and a solution", BMC Bioinformatics. 2007;8:25.Google ScholarCross Ref
- Strobl C, Boulesteix AL, Kneib T, Augustin T, Zeileis, "A. Conditional Variable Importance for Random Forests", BMC Bioinformatics. 2008;9:307.Google ScholarCross Ref
- Rui Jiang, Wanwan Tang, Xuebing Wu, Wenhui Fu, "A random forest approach to the detection of epistatic interactions in case-control studies", BMC Bioinformatics, 2009, vol. 10 pg. S65.Google ScholarCross Ref
- Biau, G., Devroye, L., and Lugosi, G, "Consistency of random forests and other averaging classifiers", The Journal of Machine Learning Research. 2008; 9:2015--2033. Google ScholarDigital Library
- Biau, G., "Analysis of a random forests model", The Journal of Machine Learning Research. 2012; 98888:1063--1095. Google ScholarDigital Library
- W. H. Hu, B. T. Lin, F. S. Lu, J. Y. Jeng, "An Upselling Pricing Model using SDP-based Rating Mechanism with Dynamic Weight", MISNC 2016. Google ScholarDigital Library
Index Terms
- Promotion Recommendation Method and System Based on Random Forest
Recommendations
E-commerce Recommendation with Personalized Promotion
RecSys '15: Proceedings of the 9th ACM Conference on Recommender SystemsMost existing e-commerce recommender systems aim to recommend the right products to a consumer, assuming the properties of each product are fixed. However, some properties, including price discount, can be personalized to respond to each consumer's ...
An Upselling Pricing Model using SDP-based Rating Mechanism with Dynamic Weight
MISNC, SI, DS 2016: Proceedings of the The 3rd Multidisciplinary International Social Networks Conference on SocialInformatics 2016, Data Science 2016With the launching of 4G services and evolving in mobile technology, data-hungry applications are ubiquitous and embraced by more and more customers. It is obvious that customers are more likely to run out of their monthly data allowance when using ...
Note. Optimal Promotion Strategies: A Demand-Sided Characterization
We generalize Narasimhan's Narasimhan, Chakravarthi. 1988. Competitive promotional strategies. J. Bus.614 427-449. model of retail promotion to include multiple products and general demand functions. Doing so allows us to further characterize optimal ...
Comments