Active Machine Learning for Consideration Heuristics

Dzyabura, Daria; Hauser, John R.
September 2011
Marketing Science;Sep/Oct2011, Vol. 30 Issue 5, p801
Academic Journal
We develop and test an active-machine-learning method to select questions adaptively when consumers use heuristic decision rules. The method tailors priors to each consumer based on a "configurator." Subsequent questions maximize information about the decision heuristics (minimize expected posterior entropy). To update posteriors after each question, we approximate the posterior with a variational distribution and use belief propagation (iterative loops of Bayes updating). The method runs sufficiently fast to select new queries in under a second and provides significantly and substantially more information per question than existing methods based on random, market-based, or orthogonal-design questions. Synthetic data experiments demonstrate that adaptive questions provide close-to-optimal information and outperform existing methods even when there are response errors or "bad" priors. The basic algorithm focuses on conjunctive or disjunctive rules, but we demonstrate generalizations to more complex heuristics and to the use of previous-respondent data to improve consumer-specific priors. We illustrate the algorithm empirically in a Web-based survey conducted by an American automotive manufacturer to study vehicle consideration (872 respondents, 53 feature levels). Adaptive questions outperform market-based questions when estimating heuristic decision rules. Heuristic decision rules predict validation decisions better than compensatory rules.


Related Articles

  • A Bayes Optimal Approach for Partitioning the Values of Categorical Attributes. BoullĂ©, Marc; Ridgeway, Greg // Journal of Machine Learning Research;9/1/2005, Vol. 6 Issue 9, p1431 

    In supervised machine learning, the partitioning of the values (also called grouping) of a categorical attribute aims at constructing a new synthetic attribute which keeps the information of the initial attribute and reduces the number of its values. In this paper, we propose a new grouping...

  • Superior Guarantees for Sequential Prediction and Lossless Compression via Alphabet Decomposition. Begleiter, Ron; El-Yaniv, Ran; Ron, Dana // Journal of Machine Learning Research;2/1/2006, Vol. 7 Issue 2, p379 

    We present worst case bounds for the learning rate of a known prediction method that is based on hierarchical applications of binary context tree weighting (CTW) predictors. A heuristic application of this approach that relies on Huffman's alphabet decomposition is known to achieve state-of-the-...

  • PRFSIDENTIAL ADDRESS: The Value of Theory in Consumer Research. Tybout, Alice M. // Advances in Consumer Research;1995, Vol. 22 Issue 1, p1 

    The article presents a speech titled "The Value of Theory in Consumer Research," by Alice M. Tybout, a marketing professor at Northwestern University, delivered as the presidential address to the Association for Consumer Research (ACR). She assesses the role of theory in predicting consumer...

  • 7 Tips For Avoiding Common Customer Research Mistakes. Giddings, Chad // Multichannel Merchant Exclusive Insight;1/30/2012, p1 

    The author offers tips to businesses on how to maximize their research to better understand the needs, perceptions and attitudes of their customers. He suggests that companies define what they want to know in advance, pointing out that doing research for the sake of doing it is a waste of time....

  • The Author Replies. Liefeld, John P. // Marketing Research;Summer2003, Vol. 15 Issue 2, p50 

    The article presents a reply to comments on the author's approach to consumer research published in the Summer 2003 issue of Marketing Research. Many consumer research colleagues are true corporatists who are loyal to a selected set of mythodologies in consumer research. The author was once a...

  • A Heuristic Search Algorithm for Flow-Shop Scheduling. Fan, Joshua Poh-Onn; Winley, Graham K. // Informatica (03505596);Oct2008, Vol. 32 Issue 4, p453 

    This article describes the development of a new intelligent heuristic search algorithm (IHSA*) which guarantees an optimal solution for flow-shop problems with an arbitrary number of jobs and machines provided the job sequence is constrained to be the same on each machine. The development is...

  • IBA-II: Another Item-based Algorithm of Learning from Examples. YUAN LIN // Applied Mechanics & Materials;2014, Vol. 631-632, p41 

    Another novel algorithm of learning from example is given. A significant deference between a traditional algorithm with the given item-based algorithm is that the traditional algorithm must scan in example space (i.e. scanning the given examples one by one) to obtain the needed heuristic...

  • Inter-operability and reusability: the science of annotation. Stede, Manfred; Huang, Chu-Ren // Language Resources & Evaluation;Mar2012, Vol. 46 Issue 1, p91 

    Annotating linguistic data has become a major field of interest, both for supplying the necessary data for machine learning approaches to NLP applications, and as a research issue in its own right. This comprises issues of technical formats, tools, and methodologies of annotation. We provide a...

  • HEURISTIC SEARCH PROCESSES IN DECISION MAKING. Payne, John W. // Advances in Consumer Research;1976, Vol. 3 Issue 1, p321 

    Research in psychology indicates that individuals employ heuristics in performing a variety of complex tasks. Evidence is presented that consumers also utilise heuristics to reduce the amount of information searched in making a decision. Aspects of theory and methodology of cognitive psychology...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics