SYNTHESIS OF A BIT-BASED FEATURE USING THE K-NEAREST NEIGHBORS METHOD FOR DETECTING CONFLICT OBJECTS IN MEDICAL DATA
Keywords:
synthetic feature, K-nearest neighbors, binary sequence, conflict objects, local neighborhood, Juravlyev metric, Euclidean metric, heart disease dataset, pattern recognitionAbstract
The identification of hidden patterns and boundary objects remains one of the important problems in pattern recognition, data mining, and machine learning. Traditional classification methods focus mainly on assigning objects to predefined classes and often provide limited information about the local structural properties of the data. This paper proposes a method for synthesizing a new informative feature based on the local neighborhood structure of objects using the K-Nearest Neighbors (KNN) algorithm. The proposed approach transforms the class composition of the nearest neighbors into a binary sequence and subsequently into a synthetic numerical feature. This feature reflects the local environment of each object and can be used for the detection of conflict objects, analysis of class boundaries, and discovery of latent relationships between classes.
References
Guyon I., Elisseeff A. An Introduction to Variable and Feature Selection // Journal of Machine Learning Research. 2003. Vol. 3. P. 1157–1182.
Тухтабаев К.А., Эргашева Ш.Э. Аналитические выражения для вычисления значений латентных признаков // Proceedings of the Seminar dedicated to the memory of professor M.I. Isroilov, CMT2024. Tashkent, 2024. P. 159–162.
Tenenbaum J.B., de Silva V., Langford J.C. A Global Geometric Framework for Nonlinear Dimensionality Reduction // Science. 2000. Vol. 290. No. 5500. P. 2319–2323.
Ignatev N.A., Akbarov B.K., Tuhtabayev K.A. Deriving Analytical Expressions for Calculating the Values of Latent Features in Recognition Problems // Problems of Computational and Applied Mathematics. 2023. No. 6/1(54). P. 68–76.
Ergasheva Sh.E. Dimensionality Reduction of Feature Space Using Nonlinear Transformations of Heterogeneous Features // Education, Science and Innovative Ideas in the World. 2024.
Belkin M., Niyogi P. Laplacian Eigenmaps for Dimensionality Reduction and Data Representation // Neural Computation. 2003. Vol. 15. No. 6. P. 1373–1396.
Ignatev N.A. On Nonlinear Transformations of Features Based on the Functions of Objects Belonging to Classes // Pattern Recognition and Image Analysis. 2021. Vol. 31. No. 2. P. 197–204.
Downloads
Published
Issue
Section
How to Cite
