Autor/Autorin: Fan Wei
Wei Fan received his PhD in Computer Science from ColumbianUniversity in 2001. His researchninterests are in risk analysis, high performancencomputing, skewed distribution, cost-sensitive, streams,nensemble, nonparametric methods, graph mining, featurenconstruction and selection, sample bias, transfernlearning, commercial data mining systems.
Prior work in inductive learning focused on genericalgorithms that sought to reduce complexity. Thus,simplifying assumptions were made. 1. All dataresides on a single processor, and residesentirely in main memory;Clearly in modern organizations today, most dataresides in a distributedarchitecture with only small portions being residentin main memory at each moment. 2. Each datum is considered equally important and uniformcosts are assumed. In real world contexts, differentexemplars frequently havevarying costs.3. All features are freely acquired with nocomputational or monetarycosts. This is unrealistic for many applications,such as medical diagnosis. Usually, the test for eachfeature consumes different costs and cannot beignored, i.e., accurate models that only takeadvantage of the most expensive features are notacceptable. 4. Model is computed on the basis of completeknowledge. A learned hypothesis willbe applied to scenarios that are completelyrepresented in the training set. Thisassumption is more often violated than satisfied.There are usually new and unknownpatterns that traditional hypotheses will eitherignore or misclassify.
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 the training set. This
assumption is more often violated than satisfied.
There are usually new and unknown
patterns that traditional hypotheses will either
ignore or misclassify.
Wei Fan received his PhD in Computer Science from Columbia
University in 2001. His research
interests are in risk analysis, high performance
computing, skewed distribution, cost-sensitive, streams,
ensemble, nonparametric methods, graph mining, feature
construction and selection, sample bias, transfer
learning, commercial data mining systems.
Über den Autor
Wei Fan received his PhD in Computer Science from Columbia
University in 2001. His research
interests are in risk analysis, high performance
computing, skewed distribution, cost-sensitive, streams,
ensemble, nonparametric methods, graph mining, feature
construction and selection, sample bias, transfer
learning, commercial data mining systems.
Klappentext
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 the training set. This
assumption is more often violated than satisfied.
There are usually new and unknown
patterns that traditional hypotheses will either
ignore or misclassify.
