Browsing by Author "Groves, William"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    A regression model for predicting optimal purchase timing for airline tickets
    (2011-10-18) Groves, William; Gini, Maria
    Optimal timing for airline ticket purchasing from the consumer's perspective is challenging principally because buyers have insufficient information for reasoning about future price movements. This paper presents a model for computing expected future prices and reasoning about the risk of price changes. The proposed model is used to predict the future expected minimum price of all available flights on specific routes and dates based on a corpus of historical price quotes. Also, we apply our model to predict prices of flights with specific desirable properties such as flights from a specific airline, non-stop only flights, or multi-segment flights. By comparing models with different target properties, buyers can determine the likely cost of their preferences. We present the expected costs of various preferences for two high-volume routes. Performance of the prediction models presented is achieved by including instances of time-delayed features, by imposing a class hierarchy among the raw features based on feature similarity, and by pruning the classes of features used in prediction based on in-situ performance. Our results show that purchase policy guidance using these models can lower the average cost of purchases in the 2 month period prior to a desired departure. The proposed method compares favorably with a deployed commercial web site providing similar purchase policy recommendations.
  • Thumbnail Image
    Item
    Toward Automating and Systematizing the Use of Domain Knowledge in Feature Selection
    (2015-08) Groves, William
    Constructing prediction models for real-world domains often involves practical complexities that must be addressed to achieve good prediction results. Often, there are too many sources of data (features). Limiting the set of features in the prediction model is essential for good performance, but prediction accuracy may be degraded by the inadvertent removal of relevant features. The problem is even more acute in situations where the number of training instances is limited, as limited sample size and domain complexity are often attributes of real-world problems. This thesis explores the practical challenges of building regression models in large multivariate time-series domains with known relationships between variables. Further, we explore the conventional wisdom related to preparing datasets for model calibration in machine learning, and discuss best practices for learning time-varying concepts from data. The core contribution of this work is a novel wrapper-based feature selection framework called Developer-Guided Feature Selection (DGFS). It systematically incorporates domain knowledge for domains characterized by a large number of observable features. The observable features may be related to each other by logical, temporal, or spatial relationships, some of which are known to the model developer a priori. The approach relies on limited domain-specific knowledge but can replace or improve upon more elaborate domain specific models and on fully automated feature selection for many applications. As a wrapper-based approach, DGFS can augment existing multivariate techniques used in high-dimensional domains to produce improved modeling results particularly in situations where the volume of training data is limited. We demonstrate the viability of our method in several complex domains (natural and synthetic) that have significant temporal aspects and many observable features.