A cognitive bias approach to feature selection and weighting for case-based learners

Research in psychology, psycholinguistics, and cognitive science has discov ered and examined numerous psychological constraints on human information p rocessing. Short term memory limitations, a focus of attention bias, and a preference for the use of temporally recent information are three examples. This paper shows that psychological constraints such as these can be used effectively as domain-independent sources of bias to guide feature set sele ction and weighting for case-based learning algorithms. We first show that cognitive biases can be automatically and explicitly enc oded into the baseline instance representation: each bias modifies the repr esentation by changing features, deleting features, or modifying feature we ights. Next, we investigate the related problems of cognitive bias selectio n and cognitive bias interaction for the feature weighting approach. In par ticular, we compare two cross-validation algorithms for bias selection that make different assumptions about the independence of individual component biases. In evaluations on four natural language learning tasks, we show tha t the bias selection algorithms can determine which cognitive bias or biase s are relevant for each learning task and that the accuracy of the case-bas ed learning algorithm improves significantly when the selected bias(es) are incorporated into the baseline instance representation.