This paper examines several analogies employed in computational data analy- sis techniques: the analogy to the brain for artificial neural networks, the anal- ogy to statistical mechanics for simulated annealing and the analogy to evolu- tion for genetic algorithms. After exploring these analogies, we compare them to analogies in scientific models and highlight that scientific models address specific empirical phenomena, whereas data analysis models are application- neutral: they can be used whenever a set of data meets certain formal require- ments, regardless of what phenomenon these data pertain to. Through the analogy, computational data analysis techniques inherit a conceptual idea from which the principle of the technique is developed. In all cases of computational data analysis techniques, the analogies used - and the metaphors generated by them - help us to understand the technique by providing a more concrete framework for understanding what is otherwise an abstract method. In the dif- ferent examples, however, the significance of the analogles varies. Analogy can, though need not, be indispensable for a technique.
CITATION STYLE
Bailer-Jones, D. M., & Bailer-Jones, C. A. L. (2002). Modeling Data: Analogies in Neural Networks, Simulated Annealing and Genetic Algorithms. In Model-Based Reasoning (pp. 147–165). Springer US. https://doi.org/10.1007/978-1-4615-0605-8_9
Mendeley helps you to discover research relevant for your work.