Using survival prediction techniques to learn consumer-specific reservation price distributions.

Area of Science:

• Economics
• Machine Learning
• Consumer Behavior

Background:

• Consumer reservation price (RP) is crucial for personalized pricing, auctions, and negotiations.
• RPs are not directly provided by consumers, necessitating predictive modeling.
• Existing methods may not fully capture individual consumer price sensitivities.

Purpose of the Study:

• To introduce a novel framework for predicting consumer-specific reservation prices (RPs).
• To leverage survival analysis techniques for modeling censored transaction data.
• To generate consumer-specific RP distributions based on individual features.

Main Methods:

• Viewed consumer transactions (and non-transactions) as censored observations.
• Applied survival analysis and prediction techniques to model RPs.
• Utilized four distinct survival prediction methods for validation.

Main Results:

• The developed models demonstrated high performance in estimating consumer-specific RP distributions.
• Validation was conducted using realistic datasets and three distinct evaluation criteria.
• The survival analysis approach proved effective in predicting RPs.

Conclusions:

• The proposed reservation price (RP) framework is effective for predicting individual consumer price sensitivities.
• Survival prediction methods offer a robust approach to modeling censored consumer transaction data.
• This framework has significant implications for personalized pricing, auctions, and negotiation strategies.