Nonparametric Pricing Analytics with Customer Covariates
针对新企业缺乏历史数据的问题，提出一种无需先验假设的自适应定价策略，通过聚类客户画像和偏好来最大化利润，并证明其遗憾值无法被其他策略改进。
Personalized pricing analytics is becoming an essential tool in retailing. Upon observing the profile of each arriving customer, the firm needs to set a price accordingly based on the observed personalized information, such as income, education background, and past purchasing history, to extract more revenue. For new entrants of the business, the lack of historical data may severely limit the power and profitability of personalized pricing. We recommend a pricing policy to firms that simultaneously learns the preference of customers based on the profiles and maximizes the profit. The pricing policy doesn't depend on any prior assumptions on how the personalized information affects consumers' preferences. Instead, it adaptively clusters customers based on their profiles and preferences, offering similar prices for customers who belong to the same cluster trading off granularity and accuracy. We prove that the regret of the proposed policy cannot be improved by any other policy.