Cost-sensitive, Scalable and Adaptive Learning Using Ensemble Methods
 |
|
Описание:
Prior work in inductive learning focused on generic algorithms that sought to reduce complexity. Thus, simplifying assumptions were made. 1. All data resides on a single processor, and resides entirely in main memory; Clearly in modern organizations today, most data resides in a distributed architecture with only small portions being resident in main memory at each moment. 2. Each datum is considered equally important and uniform costs are assumed. In real world contexts, different exemplars frequently have varying costs. 3. All features are freely acquired with no computational or monetary costs. This is unrealistic for many applications, such as medical diagnosis. Usually, the test for each feature consumes different costs and cannot be ignored, i.e., accurate models that only take advantage of the most expensive features are not acceptable. 4. Model is computed on the basis of complete knowledge. A learned hypothesis will be applied to scenarios that are completely represented in...Похожие книги
 | Managing Customers Profitably Автор: Lynette Ryals Год: 2009 |
 | Hazardous Substances and Human Health: Exposure, Impact and External Cost Assessment at the European Scale Автор: Bachmann T.M. Год: 2006 |
 | Facilities Maintenance & Repair Cost Data Автор: Charlie Wing Год: 2007 |
 | Confronting the Costs of War Автор: Michael N. Barnett Год: 1992 |
| Survey Errors and Survey Costs Автор: Robert M. Groves Год: 2004 |
 | Micro Instrumentation Автор: Melvin V. Koch Год: 2007 |