Vegetable oil production is characterized by high variability in output indicators due to nonlinear interactions between raw material parameters, equipment modes, and heat and mass transfer conditions. Existing approaches to applying machine learning in this field, as a rule, do not account for the impact of hyperparameter adjustments on forecasting quality across specific technological stages. The article presents a systematic methodology for adjusting model parameters (Ridge regression, SVR, GBM, LSTM) applied to three key tasks: predicting residual oil content in oil cake, color index during bleaching, and free fatty acid content during deodorization. In a set of 1000 observations, including technological, energy, and laboratory indicators, it is shown that adjusting hyperparameters improves forecast quality compared to basic model configurations; the best results were obtained for Gradient Boosting and LSTM.
| Mualliflar | Umidjon Ruziev, M.K. Shodiev, A.T. Rajabov |
|---|---|
| Jurnal | Кимёвий технология. Назорат ва бошқарув |
| Nashr sanasi | 2026-04-30 |
| Jild | 2026 |
| Son | 2 |
| Betlar | 52-58 |
| Til | en |
| DOI | 10.59048/2181-1105.1780 |
DOI: 10.59048/2181-1105.1780 · Maqolaning asl sahifasi
With the rapid advancement of Industry 4.0 technologies, intelligent manufacturing and big data platforms are profoundly transforming the production models of the traditional edible oil industry. The edible oil…
This research provides a detailed guideline for implementing and evaluating Proportional Integral Derivative (PID) control frameworks in lower limb rehabilitation exoskeleton robotics. It examines the role of control…
This article aims to enhance the reliability, efficiency, and diagnostic capabilities of gas-liquid separators operating in hazardous technological processes. In the study, the separator is considered as a critical…
This article presents an algorithm for simultaneously estimating the parameters and state coordinates of a multidimensional control object when some of its state variables are not directly measured. The inability to…
This paper presents a comprehensive optimization study of hydrogen production systems through alkaline water electrolysis powered by photovoltaic energy sources, with a specific focus on Central Asian applications. The…
A new design of an induction transducer has been developed for measuring linear and torsional vibrations in different directions with high sensitivity by constructing an inertial element consisting of four mutually…
This paper examines the development of regularized algorithms for synthesizing control devices in control systems for polynomial objects, described by multidimensional Volterra functional series. The synthesis problem…
This study reports the results of a comprehensive physicochemical investigation of the urea–L-valine binary system, conducted to validate its potential as a modified base for liquid nitrogen fertilizers. By utilizing…
The computation of eigenvalues and eigenvectors under uncertainty is a fundamental problem in fuzzy linear algebra and decision analysis. When matrix elements are represented by fuzzy numbers, classical spectral methods…
The work provides an analytical review of the current state of modeling and controlling the carbonization process of ammoniated brine. It analyzes modern methods of mathematical, simulation, and intelligent modeling of…