Refine
Document Type
- Doctoral thesis (3)
Institute
- Wirtschaftswissenschaftliche Fakultät (3) (remove)
Language
- English (3)
Keywords
- Prognoseverfahren (3) (remove)
In forecast combination, multiple predictions are linearly combined through the assignment of weights to individual forecast models or forecasters. Various approaches exist for defining the weights, which typically involve determining the number of models to be used for combination (selection), choosing an appropriate weighting function for the forecast scenario (weighting), and using regularization techniques to adjust the calculated weights (shrinkage). The papers listed address the integration of the three approaches into holistic data analytical models.
The first paper develops a two-stage model in which weights are first calculated based on the in-sample error covariances to minimize the error on the available data by combining the individual models. Based on the selection status of a model, the weight of an individual model is then linearly shrunk either toward the mean or toward zero. The selection status is thereby derived apriori from information criteria, where Contribution 1 introduces the selection based on the model’s in-sample accuracy and performance robustness under uncertainty.
Contribution 2 modifies the two-stage model to shrink the forecasters’ weights non-proportionally to the mean or zero. Further, a new information criterion based on forward feature selection is proposed that iteratively selects the forecaster that is expected to achieve the largest increase in accuracy when combined.
Contribution 3 extends the iteration-based information criterion presented in the second contribution to consider diversity gains in addition to accuracy gains when selecting and combining forecasters.
A one-stage model for simultaneous weighting, shrinking, and selection is finally built and evaluated on simulated data in Contribution 4. Instead of requiring a prior selection criterion, the model itself learns which forecasters to shrink to the mean or to zero, while relying on a new sampling procedure to tune the model.
This thesis develops methodological approaches to extend the service spectrum of price comparison websites for electronic consumer goods with respect to the determination of the optimal purchase time point. The central decision criterion of price comparison sites’ customers is the (expected) minimum price.
This paper therefore focuses on the prediction of minimum prices and therefrom derived events and answers three questions that set the stage for extending services of price comparison sites:
1. How long do is the expected waiting time to buy our desired product for a set budget or desired price?
2. When should a customer buy so that s/he pays the lowest price within a short, predefined decision horizon?
3. How can a price comparison site make price predictions for a large, heterogeneous set of products over different horizons, and what are the best methods for doing so?
The database used for the analysis results from the specific requirements of the study context. While the first two papers focus on a subset of electronic consumer goods - smartphones - to demonstrate the application of the developed methodology, the third paper addresses the challenges of price prediction for the heterogeneous product landscape on price comparison sites and therefore includes multiple product categories in the analysis.
This cumulative dissertation consists of five research contributions considering different questions for forecasting price decline events for consumer goods. The main goal of the dissertation is the development of a methodological framework to derive economically beneficial buying recommendations on the basis of historic minimum price time series that support price-sensitive customers in scheduling their buying decisions when purchasing homogeneous goods. The central steps for the development, configuration and application of a suitable forecasting methodology are described in the five research papers. The first article develops a probabilistic forecasting methodology and shows opportunities and challenges when evaluating decision recommendations. Article 2 exemplifies the developed statistical procedure and shows that generated decision recommendations are economically viable using a real sample from the German e-commerce market. The third article discusses the economic and statistical implications when setting a price decline threshold in more detail and shows that the economically optimal threshold can be determined based on the historic minimum price time series. Additionally, an approach is presented that produces precise estimates based on the incomplete time series 90 days after the product’s market entry. Article 4 develops an additional approach to improve target price alarms on price comparison sites. The fifth article extends the methodological spectrum for forecasting price change events and shows that identified dynamics are generalizable over time and product groups. The exploitation of these dynamics leads to more precise and economically relevant buying recommendations.