This article proposes a genetic algorithm-based approach to optimize the filling of missing NaN values in a dataset. The focus is on selecting NaN values in the dataset directly corresponding to the results of the classification task. In the proposed method, each individual is represented as a chromosome in the form of a vector of all missing values. The search space is bounded by the given intervals for numerical attributes, and by the set of appropriate categories for categorical attributes. The accuracy indicator of the Random Forest ensemble model was used as the fitness function in the genetic algorithm.
| Mualliflar | Sariyev, Shohruh |
|---|---|
| Jurnal | Al-Farg'oniy avlodlari |
| Nashr sanasi | 2025-12-20 |
| Son | 4 |
| Betlar | 138-144 |
| Til | Rus |
Генетический алгоритм, МО, ИИ, СЛ, KNN, Genetic algorithm, ML, AI, Random Forest, KNN, Genetik algoritm, mashinaviy o‘qitish, sun’iy intellekt, Random Forest, KNN
The article describes the development of a regression mathematical model to evaluate the thermal performance of a parabolic trough solar collector (PTC). The model accounts for key factors such as solar flux density and…
This article discusses the increasing threat of malware on social networks and highlights some effective ways to safeguard user data. Social media platforms have become vital channels for communication, marketing, and…
In the field of digital imaging, accurate estimation of noise level allows solving many image processing tasks in terms of image quality improvement and image restoration. In this research work, Gaussian noise, which…
The article proposes an approach to designing territorially branched road networks and their associated engineering structures as an integrated system. The approach is presented as a step-by-step algorithm that includes…
The article considers the software of an intelligent information system based on the presentation of production knowledge, as well as the results of experimental verification of the developed algorithms. The…
Discrete mathematics is one of the most important scientific areas that form the theoretical basis of modern cryptography. Discrete areas such as prime number theory, modular arithmetic, finite fields, combinatorics…
This study analyzes modern intelligent systems that use machine learning models to determine an individual’s solvency and increase accuracy. The study highlights the principles of operation of traditional credit rating…
Matrix reflections are one of the main mechanisms of modern cryptographic algorithms. They play an important role in increasing the diffusion strength, creating protection against linearity, forming complex structures…
The widespread adoption of Software-Defined Networking (SDN) in medium-scale information and communication systems increases network management flexibility but introduces security risks due to the centralized…
This article presents a theoretical and experimental investigation of the applicability of deep learning techniques in modern cryptanalysis. Traditional cryptanalytic methods are computationally intensive, as they…