- Sirdaryo viloyati Guliston shahar
- Dushanba-Shanba: 9.00 - 18.00
Determining the minimum cover based on the grouping of class objects
Annotatsiya
The method of dividing the training sample into non-intersecting groups of objects
based on the property of their connection according to the defined subset of boundary
objects of classes is considered. Grouping is used to find the coverage of the sample with
reference objects. The formation of a new feature space for representing objects is
described by nonlinear mapping of non-intersecting set of features onto the number axis.
Kalit so‘zlar
image recognition, logical patterns, data cluster analysi
Foydalanilgan adabiyotlar
According to the sequential exclusion principle used in the coverage search process,
the E0 sample is divided into two subsets: Eed reference set and control set Ek, E0 = Eed
∪Ek. At the beginning of the process Eed = E0, Ek = ∅. Arrangement by recess values
{𝑅𝑆}𝑆∈𝐺𝑢
,𝑢 = 1,...,𝛿, is used to determine the candidate for exclusion from the list of
reference objects according to the Gu group. The selection idea consists of finding the
minimum number of standards, where the recognition algorithm according to (4) remains
correct (recognizing objects without errors) at E0. Let’s assume that the numbering of
groups from interconnected objects reflects the order | G1 |≥ ··· ≥| Gδ | and by group
Gp, p = 1,...,δ, No reference objects were selected. Candidates for Eed removal are selected
sequentially starting from S ∈ Gp with minimal RS value. If the inclusion of S ∈ Ek violates
the correctness of the decisive rule (4), then S returns to the set E