一种博弈论辅助的机器学习算法检测用户流失行为

doi:10.11959/j.issn.1000-0801.2020164

Abstract

Abstract:

At the end of November 2019,China officially implemented the number portability policy (MNP) that has been in trial for 9 years.The policy will strengthen the liquidity and competitiveness of the telecommunication market,making the problem of subscriber churn more prominent.A game theory-assisted machine learning methodology was proposed,verified and commercialized timely,which could help mobile network operator (MNO) actively respond to competition in the MNP market.The proposed methodology provides MNO with a machine learning model to detect subscriber portability and give differentiated treatment.Experimental results show that the proposed methodology can guide MNOs to make a targeted MNP strategy,and precisely identify “abnormal” subscribers who tend to churn-out and potential new subscribers who may churn-in.In addition,the proposed methodology has been successfully commercialized,greatly improving the marketing efficiency of operators,increasing user satisfaction,and reducing the loss of users by about 50% for a tier-1 MNO in China.

Key words: subscriber churn, mobile number portability, game theory, machine learning

CLC Number:

TP181

Ye OUYANG,Aidong YANG,Fanyu MENG. A game theory-assisted machine learning methodology for subscriber churn behaviors detection[J]. Telecommunications Science, 2020, 36(6): 79-89.

Figures/Tables 13

特征	得分≥5	得分≥4	得分<4	得分≥3	得分<3	得分≥2	得分<2	得分≥1	得分<1	得分<0
${\tilde{x}}_{1}$	S₁₀	S₁₁	S₁₂	S₁₃	S₁₄	S₁₅	S₁₆	S₁₇	S₁₈	S₁₉
${\tilde{x}}_{2}$	S₂₀	S₂₁	S₂₂	S₂₃	S₂₄	S₂₅	S₂₆	S₂₇	S₂₈	S₂₉
${\tilde{x}}_{3}$	S₃₀	S₃₁	S₃₂	S₃₃	S₃₄	S₃₅	S₃₆	S₃₇	S₃₈	S₃₉
					#
${\tilde{x}}_{n}$	S_n0	S_n1	S_n2	S_n3	S_n4	S_n5	S_n6	S_n7	S_n8	S_n9

特征			特征				得分a_i	权重w_i
特征	X₁	X₂	X₃	X₄	X₅	X_e	得分a_i	权重w_i
X₁	0	1	1	0	0	$1 - β_{e_{1}}$	2	0.2
X₄	1	1	0	0	0	$1 - β_{e_{4}}$	2	0.2
X₅	1	1	1	1	0	$1 - β_{e_{4}}$	4	0.4
X_e	$β_{e_{1}}$	$β_{e_{2}}$	$β_{e_{3}}$	$β_{e_{4}}$	$β_{e_{5}}$	0	$\sum_{j} β_{e_{j}}$	$\frac{α_{e}}{\sum_{i = 1}^{e} α_{i}}$

营销目标v			特征
营销目标v	${\tilde{x}}_{1}$	${\tilde{x}}_{2}$	${\tilde{x}}_{3}$	${\tilde{x}}_{4}$	${\tilde{x}}_{5}$	${\tilde{x}}_{A}$
max_in	2%	11	57%	-5	66%	NA
min_out	-4%	20	61%	-8	64%	NA
max_in/out	5%	-14	54%	-1	69%	NA

博弈轮数	项目	运营商
博弈轮数	项目	A	B	C
第1轮	营销策略	max_in/out	max_in/out	max_in/out
	综合评分	0.3	-0.3	0
第2轮	营销策略	max_in/out	max_in/out	max_in
	综合评分	1.5	0.9	-2.4
第3轮	营销策略	max_in/out	max_in/out	min_out
	综合评分	-0.1	-0.7	0.8
第4轮	营销策略	max_in/out	max_in	max_in/out
	综合评分	-0.3	0.9	-0.6
第5轮	营销策略	max_in/out	max_in	max_in
	综合评分	0.9	2.1	-3
第6轮	营销策略	max_in/out	max_in	min_out
	综合评分	-0.7	0.5	0.2
第7轮	营销策略	max_in/out	min_out	max_in/out
	综合评分	0	0.3	-0.3
第8轮	营销策略	max_in/out	min_out	max_in
	综合评分	1.2	1.5	-2.7
第9轮	营销策略	max_in/out	min_out	min_out
	综合评分	-0.4	-0.1	0.5
第10轮	营销策略	max_in	max_in/out	max_in/out
	综合评分	-1.3	0.5	0.8
第11轮	营销策略	max_in	max_in/out	max_in
	综合评分	-0.1	1.7	-1.6
第12轮	营销策略	max_in	max_in/out	min_out
	综合评分	-1.7	0.1	1.6
第13轮	营销策略	max_in	max_in	max_in/out
	综合评分	-1.9	1.7	0.2
第14轮	营销策略	max_in	max_in	max_in
	综合评分	-0.7	2.9	-2.2
第15轮	营销策略	max_in	max_in	min_out
	综合评分	-2.3	1.3	1
第16轮	营销策略	max_in	min_out	max_in/out
	综合评分	-1.6	1.1	0.5
第17轮	营销策略	max_in	min_out	max_in
	综合评分	-0.4	2.3	-1.9
第18轮	营销策略	max_in	min_out	min_out
	综合评分	-2	0.7	1.3
第19轮	营销策略	min_out	max_in/out	max_in/out
	综合评分	0.7	-0.5	-0.2
第20轮	营销策略	min_out	max_in/out	max_in
	综合评分	1.9	0.7	-2.6
第21轮	营销策略	min_out	max_in/out	min_out
	综合评分	0.3	-0.9	0.6
第22轮	营销策略	min_out	max_in	max_in/out
	综合评分	0.1	0.7	-0.8
第23轮	营销策略	min_out	max_in	max_in
	综合评分	1.3	1.9	-3.2
第24轮	营销策略	min_out	max_in	min_out
	综合评分	-0.3	0.3	0
第25轮	营销策略	min_out	min_out	max_in/out
	综合评分	0.4	0.1	-0.5
第26轮	营销策略	min_out	min_out	max_in
	综合评分	1.6	1.3	-2.9
第27轮	营销策略	min_out	min_out	min_out
	综合评分	0	-0.3	0.3

References 24

[1]	MAICAS J P , POLO Y , SESE J . Reducing the level of switching costs in mobile communications:the case of mobile number portability[J]. Telecommunications Policy, 2009,33(9): 544-554.
[2]	OTSUKA T , MITOMO H . User benefits and operator costs of mobile number portability in Japan and impact on market competitiveness[J]. Telecommunications Policy, 2013,37(4-5): 345-356.
[3]	SHIN D H . A study of mobile number portability effects in the united states[J]. Telemat Inform, 2007,24(1): 1-14.
[4]	BU?HLER S , DEWENTER R , HAUCAP J . Mobile number portability in Europe[J]. Telecommunications Policy, 2006,30(7): 385-399.
[5]	FCC. Wireless local number portability (WLNP)[EB]. 2016.
[6]	China Mobile. Announcements[EB]. 2020.
[7]	China Telecom. Investor calendar[EB]. 2020.
[8]	China Unicom. Presentations ＆ Webcasts[EB]. 2020.
[9]	SINA. Operators will cancel the unlimited packages.Telecom will no longer reach the speed limit from next month[EB]. 2019.
[10]	BERSON A , SMITH S , THEARLING K . Building data mining applications for CRM[J]. 1999.
[11]	GARC′IA-MARI?NOSO B , SUáREZ D . Switching mobile operators:evidence about consumers behavior from a longitudinal survey[J]. Telecommunications Policy, 2019,43(5): 426-433.
[12]	GERPOTT T J , RAMS W , SCHINDLER A . Customer retention,loyalty,and satisfaction in the German mobile cellular telecommunications market[J]. Telecommunications Policy, 2001,25(4): 249-269.
[13]	LIANG D , MA Z , QI L . Service quality and customer switching behavior in China’s mobile phone service sector[J]. Journal of Business Research, 2013,66(8): 1161-1167.
[14]	BACKIEL A , BAESENS B , CLAESKENS G . Predicting time-to-churn of prepaid mobile telephone customers using social network analysis[J]. Technological Forecasting and Social Change, 2016,67(9): 1135-1145.
[15]	NITZAN I , LIBAI B . Social effects on customer retention[J]. Journal of Marketing, 2011,75(6): 24-38.
[16]	YESHWANTH V , RAJ V V , SARAVANAN M . Evolutionary churn prediction in mobile networks using hybrid learning[C]// Proceedings of Twenty-Fourth International FLAIRS Conference. Palo Alto:AAAI Press, 2011.
[17]	ZHANG R , ZHU Q . A game-theoretic defense against data poisoning attacks in distributed support vector machines[C]// Proceedings of 2017 IEEE 56th Annual Conference on Decision and Control (CDC). Piscataway:IEEE Press, 2017: 4582-4587.
[18]	SLADE M E . Vancouver’s gasoline-price wars:an empirical exercise in uncovering supergame strategies[J]. Rev.Econmic Studies, 1992,59(2): 257-276.
[19]	FERRERO R , SHAHIDEHPOUR S , RAMESH V . Transaction analysis in deregulated power systems using game theory[J]. IEEE Transactions on Power Systems, 1997,12(3): 1340-1347.
[20]	MAGALONI B . The game of electoral fraud and the ousting of authoritarian rule[J]. American Journal Politcal Science, 2010,54(3): 751-765.
[21]	ADEBIYI S O , OYATOYE E O , MOJEKWU J N . Predicting customer churn and retention rates in Nigeria’s mobile telecommunication industry using Markov chain modeling[J]. Acta Universitatis Sapientiae,Economics and Business, 2015,3(1): 67-80.
[22]	CHEN T , GUESTRIN C . XGBoost:a scalable tree boosting system[C]// Proceedings of the 22nd ACM Sigkdd International Conference on Knowledge Discovery and Data Mining. New York:ACM Press, 2016: 785-794.
[23]	DORFFNER G , . Neural networks for time series processing[C]// Proceedings of Neural Network World.[S.l.:s.n]. 1996.
[24]	CZAJKOWSKI M , SOBOLEWSKI M . How much do switching costs and local network effects contribute to consumer lock-in in mobile telephony?[J]. Telecommunications Policy, 2016,40(9): 855-869.

Metrics

Recommended 0

No Suggested Reading articles found!

运营商				特征
运营商	X₁	X₂	X₃	X₄	X₅	X₆	X₇	X₈
A	-5.88%	-70.4	58.50%	-29.7	68.20%	4 701.9	3.093	162.5
B	-0.79%	-14.8	13.90%	1.3	5.60%	389.5	0.467	38.8
C	-3.91%	-15.3	27.60%	1.3	26.20%	1 807.2	1.391	76.6

营销策略v	X₁	X₂	X₃	X₄	X₅	X_A	综合得分Z_v
营销策略v	0.20	0.10	0.10	0.20	0.40	0	综合得分Z_v
max_in	3	4	4	2	3	NA	3.00
min_out	1	5	5	1	2	NA	2.20
max_in/out	5	1	1	2	4	NA	3.20

营销策略v		A			B			C
营销策略v	*Y₁	*Y₂	*Y₃	*Y₁	*Y₂	*Y₃	*Y₁	*Y₂	*Y₃
max_in	0.17	44%	56%	-0.1	33%	56%	0.27	44%	44%
min_out	1.37	0	100%	-2.5	100%	0	-1.33	100%	0
max_in/out	0.77	22%	78%	0.7	0	89%	0.67	11%	78%

A game theory-assisted machine learning methodology for subscriber churn behaviors detection

RichHTML

PDF下载

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 13

References 24

Related Articles 15

Metrics

Recommended 0

[1]	Zehua GUO, Haowen ZHU, Tongwen XU. Network modal innovation for distributed machine learning [J]. Telecommunications Science, 2023, 39(6): 44-51.
[2]	Yaqiong LIU, Zhe LYU, Yafei ZHAO, Guochu SHOU. A survey on AI techniques applied in the satellite communication/satellite Internet field [J]. Telecommunications Science, 2023, 39(2): 10-24.
[3]	Ruihong JIANG, Yizhe FENG, Yaohua SUN, Haina ZHENG. A survey on networking key technologies for LEO satellite network [J]. Telecommunications Science, 2023, 39(2): 37-47.
[4]	Jianwu ZHANG, Yanjun AN, Huangyan DENG. A survey on DNS attack detection and security protection [J]. Telecommunications Science, 2022, 38(9): 1-17.
[5]	Lushan ZOU, Xiaowen HUANG, Jingmin YANG, Yifeng ZHENG, Guanglin ZHANG, Wenjie ZHANG. Review on resources allocation and pricing methods in mobile edge computing [J]. Telecommunications Science, 2022, 38(3): 113-132.
[6]	Zhichao ZHOU, Yi FENG, Xiaohan XIA, Yuyao FENG, Chao CAI, Jiahui QIU, Lihui YANG, Yunxiao WU. Outdoor location scheme with fingerprinting based on machine learning of mobile cellular network [J]. Telecommunications Science, 2021, 37(8): 85-95.
[7]	Su DIAO, Cheng JIANG. Trend of mobile communication market pattern under the reform of industry co-competition [J]. Telecommunications Science, 2021, 37(3): 154-161.
[8]	Xiang LI,Yuan LI,Zifei ZHANG,Zhe YANG. A density clustering-based network performance failure big data analysis algorithm [J]. Telecommunications Science, 2020, 36(9): 51-58.
[9]	Ziyu XIAO. A review of key technologies for the future of the 2030 network——Beyond 5G [J]. Telecommunications Science, 2020, 36(9): 114-121.
[10]	Yaxing QIU,Xidong WANG,Sen BIAN,Lei YUE. Load balancing based on clustering analysis and deep learning for multi-frequency and multi-mode network [J]. Telecommunications Science, 2020, 36(7): 156-162.
[11]	Feng LIN,Liujing XU,Xiaohua CHEN,Weiqiang QI,Ke CHEN,Tiantian ZHU. Method of Webshell detection based on multi-view feature fusion [J]. Telecommunications Science, 2020, 36(6): 125-132.
[12]	Jiantao SHENG,Maofei CHEN,Dongxin LIU,Laifu WANG,Guoshui SHI,Huamin JIN. Detection of malicious domain name based on a classifier combination [J]. Telecommunications Science, 2020, 36(5): 47-55.
[13]	Tao PANG,Haihua QIU,Biying PAN. Research on artificial intelligence key technologies of mobile terminal [J]. Telecommunications Science, 2020, 36(5): 145-151.
[14]	Qizhu ZHONG. Research and application of VoLTE video call quality based on machine learning [J]. Telecommunications Science, 2020, 36(3): 156-165.
[15]	Zishan LIU,Qiang CHENG,Bo LV. A survey on key technologies of privacy protection for machine learning [J]. Telecommunications Science, 2020, 36(11): 18-27.