在社群影響力下限期收益最大化之智能迭代訂價策略

Intelligent Iterative Pricing Strategy for Time-constrained Revenue Maximization with Social Influences

陳方擎(Fang-Ching Chen)

指導教授 ： 戴志華

國立臺北大學/電機資訊學院/資訊工程學系/碩士(2016年)

摘要

研究表明，利用社群影響力（也被稱為口碑效應），可以在社交網絡帶來更智能的營業，如新事物的推廣和病毒式營銷。從以前的研究中，我們注意到，社群影響力的效果不僅有助於推廣產品給更多的人，同時也增加了人們對產品的估值。然後在這篇論文中，我們感興趣的是一個問題：當社群影響力傳播時，是否可以透過仔細調整產品訂價以賺取更高的收益"，並研究在一個時間區段內從社群網路帶來最大收益的迭代訂價策略。繼之前的作品後，我們研究了這個問題基於合併貨幣概念的Linear Threshold（M-LT）擴散模型。為了解決這個問題，我們導出一個基於M-LT 擴散的理論型模型，稱之為LAW model，並提出了兩種可行的演算法|AP-CDS 和Special AP-CAP|來在每個時間情況找最佳產品訂價。藉由在真實的社群網絡進行模擬，並與傳統的定價方法如固定訂價和貪婪訂價比較，其結果證明了AP-CDS 和Special AP-CAP 在選擇導致的最大收益的一個連續訂價的效用。

關鍵字

社群影響力 ； 社群網路 ； 影響力最大化 ； 收益最大化

並列摘要

It has been shown that utilizing social influences, also known as the word-of-mouth effects, can bring smarter business over social networks such as innovation promotion and viral marketing. From previous works, we note that the effect of social influences not only helps the promotion of a product to more people, but also increases people's valuations towards the product. In this paper, we are then interested in asking a question "whether higher revenue can be earned by carefully adjusting the product pricing as the social influence spreads" and studying the iterative pricing problem for bringing the maximum revenue from social networks within a time period. Following previous works, we study this problem based on the monetary-concept incorporated Linear Threshold (M-LT) diffusion model. To solve the problem, we derive a theoretical model, refer to as the LAW model, based on the M-LT diffusion, and propose two feasible algorithms, AP-CDS and the special AP-CAP, to find the best product pricing at each time instance. By running simulations on real
social networks and comparing with traditional pricing approaches such as fixed pricing and greedy pricing, the effectiveness of AP-CDS and the special AP-CAP at choosing a series of pricing leading to the maximal revenue is demonstrated.

並列關鍵字

Social Influence ； Social Network ； Influence Maximization ； Revenue Maximization

參考文獻

[25] J. Pei, B. Jiang, X. Lin, and Y. Yuan. Probabilistic skylines on

[2] S. E. Asch. Effects of Group Pressure on the Modification and Distortion

influence in nearly optimal time. In SIAM SODA, 2014.

pricing in social networks with incomplete information. In WINE,

for prevalent viral marketing in large-scale social networks. In