What is it about?
Given is a set of objects and a set of features. All objects should be classified into two classes by a classificator. Classificator is a binary decision tree, where features are sitting in nodes. Leaves of this tree contain the misclassification error. When we build a decision tree to classify objects into two classes we need to control the misclassification error and the complexity of the decision tree. Large tree is inpractical and uncomfortable to work with. We research how a priory expert knowledge about each class can decrease the misclassification error.
Featured Image
Photo by Adarsh Kummur on Unsplash
Why is it important?
Very often some prior information about classes and objects from previous life is available. It is important to incorporate this information into classification process. This will help to decrease the tree complexity and the error.
Perspectives
Decision trees could be used for classification patients in a clinic, identification defects in production and many others. Decision trees are intuitively clear to non-mathematicians. No specific knowledge is required.
Dr. Alexander Litvinenko
Rheinisch Westfalische Technische Hochschule Aachen
Read the Original
This page is a summary of: The influence of prior knowledge on the expected performance of a classifier, Pattern Recognition Letters, November 2003, Elsevier,
DOI: 10.1016/s0167-8655(03)00099-0.
You can read the full text:
Contributors
The following have contributed to this page
