Machine Learning Based on Kernel Function Controlled Gaussian Process Regression Method for In-depth Extrapolative Analysis of Covid-19 Daily Cases Drift Rates

Isabona, Joseph; Ojuh, Divine O.

doi:10.5815/ijmsc.2021.02.02

Cited by 6 publications

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The resultant precision output shows a high performance compared to other procedures also explored in the estimation process. A similar investigation involving the use of GPR for COVID-19 estimation has been reported using data from Nigeria [24].…”

Section: Related Workmentioning

confidence: 73%

Machine Learning-Based GPR with LBFGS Kernel Parameters Selection for Optimal Throughput Mining in 5G Wireless Networks

Isabona

Imoize

Ojo³

et al. 2023

Sustainability

View full text Add to dashboard Cite

Considering the ever-growing demand for an efficient method of deductive mining and extrapolative analysis of large-scale dimensional datasets, it is very critical to explore advanced machine learning models and algorithms that can reliably meet the demands of modern cellular networks, satisfying computational efficiency and high precision requirements. One non-parametric supervised machine learning model that finds useful applications in cellular networks is the Gaussian process regression (GPR). The GPR model holds a key controlling kernel function whose hyperparameters can be tuned to enhance its supervised predictive learning and adaptive modeling capabilities. In this paper, the limited-memory Broyden–Fletcher–Goldfarb–Shanno (LBFGS) with kernel parameters selection (KPS) algorithm is employed to tune the GPR model kernel hyperparameters rather than using the standard Bayesian optimization (BOP), which is computationally expensive and does not guarantee substantive precision accuracy in the extrapolative analysis of a large-scale dimensional dataset. In particular, the hybrid GPR–LBFGS is exploited for adaptive optimal extrapolative learning and estimation of throughput data obtained from an operational 5G new radio network. The extrapolative learning accuracy of the proposed GPR–LBFGS with the KPS algorithm was analyzed and compared using standard performance metrics such as the mean absolute error, mean percentage error, root mean square error and correlation coefficient. Generally, results revealed that the GPR model combined with the LBFGS kernel hyperparameter selection is superior to the Bayesian hyperparameter selection method. Specifically, at a 25 m distance, the proposed GPR–LBFGS with the KPS method attained 0.16 MAE accuracy in throughput data prediction. In contrast, the other methods attained 46.06 and 53.68 MAE accuracies. Similarly, at 50 m, 75 m, 100 m, and 160 m measurement distances, the proposed method attained 0.24, 0.18, 0.25, and 0.11 MAE accuracies, respectively, in throughput data prediction, while the two standard methods attained 47.46, 49.93, 29.80, 53.92 and 47.61, 52.54, 53.43, 54.97, respectively. Overall, the GPR–LBFGS with the KPS method would find valuable applications in 5G and beyond 5 G wireless communication systems.

show abstract

Section: Related Workmentioning

confidence: 73%

Machine Learning-Based GPR with LBFGS Kernel Parameters Selection for Optimal Throughput Mining in 5G Wireless Networks

Isabona

Imoize

Ojo³

et al. 2023

Sustainability

View full text Add to dashboard Cite

show abstract

“…Being nonparametric probabilistic models with a Kernel, Gaussian Process Regressions (GPRs) deal with a limited number of random variables having a multivariate distribution [28]. As all the linear combinations are assumed to be regularly distributed, Gaussian processes are governed by the concept of Normal distribution.…”

Section: Gaussian Process Regression Methodsmentioning

confidence: 99%

Development of Crop-Weather Models Using Gaussian Process Regression for the Prediction of Paddy Yield in Sri Lanka

Ekanayake¹,

Wickramasinghe²,

Jayasinghe³

2022

IJISA

View full text Add to dashboard Cite

This research introduces machine learning models using the Gaussian Process Regression (GPR) depicting the association between paddy yield and weather in Sri Lanka. All major regions in the island with most contribution to the total paddy production were considered in this research. The climatic factors of rainfall, relative humidity, minimum temperature, maximum temperature, average wind speed, evaporation, and sunshine hours were considered as input (independent) variables, while the paddy yield was the output (dependent) variable. The collinearity within each pair of independent and dependent variables was determined using Spearman’s and Pearson’s correlation coefficients. Data sets corresponding to the two main annual paddy cultivation seasons since 2009 were trained in MATLAB to develop crop-weather models. The most appropriate Kernel function was chosen from among four types of Kernels viz. Rational Quadratic, Exponential, Squared Exponential, and Matern 5/2 based on their degree of coherence in modeling. This approach exploits the full potential of GPR in developing highly accurate crop-weather models. The performance of the crop-weather models was measured by the Correlation Coefficient, Mean Absolute Percentage Error, Mean Squared Error, Root Mean Squared Error Ratio, Nash Number and the BIAS. All the GPR-based models proposed in this paper are highly accurate in terms of the aforementioned evaluation metrics. Accordingly, when the climatic data are known or projected, the paddy yield and thereby the harvest of Sri Lanka can be predicted precisely by using the proposed crop-weather models.

show abstract

“…В багатьох інформаційних експертних системах чи в інформаційних системах класифікації використовуються багатовимірні лінійні чи нелінійні регресії, коефіцієнти яких при змінних знаходяться методом найменших квадратів [1][2][3][4][5][6][7][8]. В роботах [9][10][11] була сформульована наступна проблема.…”

Section: вступunclassified

Метод Оцінки Коефіцієнтів При Лінійних Членах Багатовимірної Поліноміальної Регресії, Заданої Надлишковим Описом

Павлов¹,

Головченко²,

Ревич³

2022

збірник «Адаптивні Системи Автоматичного Управління Міжвідомчий

View full text Add to dashboard Cite

В роботі проведене розв’язання однієї задачі, що виникає при вирішенні проблеми оцінки коефіцієнтів членів багатовимірної поліноміальної регресії (БПР), заданої надлишковим описом за результатами умовного активного експерименту. В роботах проф. Павлова О. А. та його учнів ефективно вирішена проблема оцінки коефіцієнтів при нелінійних членах БПР з заданою малою дисперсією помилки (при умові точного розв’язання відповідних невироджених систем лінійних рівнянь). В цьому випадку гарантовано з надлишкового опису виключаються зайві нелінійні члени БПР. Задача, яка розв’язується у цій роботі, полягає у знаходженні обґрунтованого методавизначення зайвих лінійних членів БПР, заданої надлишковим описом, при використанні загального методу найменших квадратів при умові відомої (можливо, з точністю до невідомих числових значень її параметрів) функції щільності випадкової величини, що адитивно накладається на результати вимірювань вихідної змінної. Бібл. 11.

show abstract

Machine Learning Based on Kernel Function Controlled Gaussian Process Regression Method for In-depth Extrapolative Analysis of Covid-19 Daily Cases Drift Rates

Cited by 6 publications

References 13 publications

Machine Learning-Based GPR with LBFGS Kernel Parameters Selection for Optimal Throughput Mining in 5G Wireless Networks

Machine Learning-Based GPR with LBFGS Kernel Parameters Selection for Optimal Throughput Mining in 5G Wireless Networks

Development of Crop-Weather Models Using Gaussian Process Regression for the Prediction of Paddy Yield in Sri Lanka

Метод Оцінки Коефіцієнтів При Лінійних Членах Багатовимірної Поліноміальної Регресії, Заданої Надлишковим Описом

Contact Info

Product

Resources

About