Akademik Veri Yönetim Sistemi
Recent Developments and the New Directions of Research, Foundations, and Applications, Shahnaz N. Shahbazova,Ali M. Abbasov,Vladik Kreinovich,Janusz Kacprzyk,Ildar Z. Batyrshin, Editör, Springer, London/Berlin , Heidelberg, ss.317-328, 2023
This paper focuses on developing an efficient framework using Improved Principal Component Analysis (IPCA) and hybrid neural networks based machine learning techniques for estimating the quality of red wine and white wine datasets. IPCA is a dimensionality reduction technique based on cumulative sum improved PCA. The proposed machine learning classifier is introduced by integrating the Fuzzy Neural Networks (FNN) and Improved Crow Search Algorithm (ICSA). In this model, the parameters of the FNN are automatically tuned using the ICSA. The ICSA is developed by improving the dynamic awareness probability, local search and global search abilities of the standard crow search algorithm to overcome the slow convergence and local optimum problem. This hybrid classifier model of ICSA-FNN improves the classification accuracy for the wine data and provides highly accurate results within less computation time. Experiments are conducted using the Red Wine and White Wine datasets from UCI Machine Learning Repository. The results showed that the proposed quality estimation framework using IPCA and ICSA-FNN has higher performance than the existing models in terms of accuracy, precision and computation time.