Quantitative preference models are used to predict customer choices among design alternatives by collecting prior purchase data or survey answers. This paper examines how to improve the prediction accuracy of such models without collecting more data or changing the model. We propose to use features as an intermediary between the original customer-linked design variables and the preference model, transforming the original variables into a feature representation that captures the underlying design preference task more effectively. We apply this idea to automobile purchase decisions using three feature learning methods (principal component analysis (PCA), low rank and sparse matrix decomposition (LSD), and exponential sparse restricted Boltzmann machine (RBM)) and show that the use of features offers improvement in prediction accuracy using over 1 million real passenger vehicle purchase data. We then show that the interpretation and visualization of these feature representations may be used to help augment data-driven design decisions.
Improving Design Preference Prediction Accuracy Using Feature Learning
Journal of mechanical design ; 138 , 7
2016
Article (Journal)
English
BKL: | 50.15 Konstruktionslehre / 52.20 Antriebstechnik, Getriebelehre / 52.15 Maschinenelemente, Maschinenbaugruppen | |
Local classification TIB: | 770/5315/5330 |
Improving Prediction Accuracy in Mechanistic–Empirical Pavement Design Guide
Transportation Research Record | 2006
|Improving Prediction Accuracy in Mechanistic-Empirical Pavement Design Guide
Transportation Research Record | 2006
|Improving RTDGPS accuracy using hybrid PSOSVM prediction model
Online Contents | 2014
|Improving RTDGPS accuracy using hybrid PSOSVM prediction model
Elsevier | 2014
|Improving Robust Design with Preference Aggregation Methods
British Library Conference Proceedings | 2004
|