Thus, the proposed high-level technique works on the components instead of whole classes.
Thus, the high-level technique proposed here considers components instead of classes.
HL-KAOG takes advantages provided by the K-associated optimal graph and the high-level technique for data classification.
Again, considering Figure 1, the high-level technique is able to detect the pattern formed by gray cycle class and put the test instance into it.
For this reason, we propose a non-parametric network formation method based on the KAOG to work together with the high-level technique.
Differently of the previous works, the high-level technique proposed in this work is non-parametric not only at the network construction phase but also at the classification phase.
where α J denotes a component α in which its instances belong to class J, P y ( α J ) establishes the association produced by a low-level classifier between the instance y and the component α, and H y ( α J ) points to an association produced by the high-level technique between y and the component α.
The general idea behind H y ( α J ) is very simple: (i) KAOG network finds a set of components based on the purity measure, (ii) high-level technique examines these components in relation to the insertion of a test instance y, and (iii) the probabilities for each component α are obtained.
Finally, the general framework given by (16) combines low- and high-level techniques.
In the second step, HL-KAOG employs the hybrid low- and high-level techniques to classify the test instances.
Cui, Jiang, Huang, Xu, Gui, and Zhang in "POS: A High-Level System to Simplify RealTime Stream Application Development on Storm" focus on efficient and high-level techniques for processing streaming data.
