A comparative study of machine learning methods for identifying the 15 CIE standard skies DOI
Emmanuel Imuetinyan Aghimien, Danny H.W. Li, Ernest K.W. Tsang

et al.

Journal of Building Physics, Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 5, 2024

For energy-efficient building designs, the solar irradiance and daylight illuminance derived from CIE standard skies are useful. Over time, sky luminance distributions have been used to identify these skies, but sparingly measured. Thus, use of available climatic variables has become a viable alternative. Nevertheless, it is necessary determine if could correctly skies. This study addresses lack distribution measurement by classifying using measured data in Hong Kong. The classification approach was improved machine learning (ML) method. comparative analysis, five popular ML algorithms i.e., decision tree (DT), k-nearest neigbhour (KNN), light gradient boosting (LGBM), random forest (RF) support vector machines (SVM) were used. findings show that accuracies 68.1, 73.1, 74.3, 74.5, 75.4% obtained for DT, KNN, SVM, LGBM, RF models, respectively. Similarly, F1 scores 66.6, 70.2, 71.8, 72.1 72.9%, models. result shows model gave best performance while DT performed least. Also, all models would classify with reasonable accuracy. Furthermore, feature importance done, found K d , T v t α, sun, cld most important input parameters classification. Lastly, vertical ( G VT ) VL estimated predicted proposed Upon predictions, observed ranged 14.7 24.6% 13.8 19.9%. Generally, predictions less than 20%, which good

Language: Английский

A dynamic method for preparing microarray gene expression data in disease classification system DOI
Hemant B. Mahajan,

K. T. V. Reddy

Journal of Ambient Intelligence and Humanized Computing, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 24, 2025

Language: Английский

Citations

0
Dynamic Approach for Pre-processing of Microarray Gene Expression Data DOI
Hemant B. Mahajan,

K. T. V. Reddy

Lecture notes on data engineering and communications technologies, Journal Year: 2025, Volume and Issue: unknown, P. 99 - 110

Published: Jan. 1, 2025

Language: Английский

Citations

0
Research on Imbalanced Data Regression Based on Confrontation DOI Open Access
Xiaowen Liu, Huixin Tian

Processes, Journal Year: 2024, Volume and Issue: 12(2), P. 375 - 375

Published: Feb. 13, 2024

The regression model has higher requirements for the quality and balance of data to ensure accuracy predictions. However, there is a common problem imbalanced distribution in real datasets, which directly affects prediction models. In order solve imbalance regression, considering continuity target value correlation using idea optimization confrontation, we propose an IRGAN (imbalanced generative adversarial network) algorithm. Considering context information disappearance deep network gradient, constructed generation module designed composite loss function. early stages training, gap between generated samples large, easily causes non-convergence. A correction train internal relationship state action as well subsequent reward samples, guide generate alleviate non-convergence training process. corrected are input into discriminant module. On this basis, confrontation used high-quality original samples. proposed method tested fields aerospace, biology, physics, chemistry. similarity comprehensively measured from multiple perspectives evaluate proves superiority Regression performed on balanced processed by algorithm, it proven that algorithm can improve terms problem.

Language: Английский

Citations

0
Integromics: Tracking the Multi-omic Expanse in Theragnostics DOI

Shambhavee Srivastav,

Lavanya Lavanya,

Anupama Sharma Avasthi

et al.

Published: Jan. 1, 2024

Language: Английский

Citations

0
A comparative study of machine learning methods for identifying the 15 CIE standard skies DOI
Emmanuel Imuetinyan Aghimien, Danny H.W. Li, Ernest K.W. Tsang

et al.

Journal of Building Physics, Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 5, 2024

For energy-efficient building designs, the solar irradiance and daylight illuminance derived from CIE standard skies are useful. Over time, sky luminance distributions have been used to identify these skies, but sparingly measured. Thus, use of available climatic variables has become a viable alternative. Nevertheless, it is necessary determine if could correctly skies. This study addresses lack distribution measurement by classifying using measured data in Hong Kong. The classification approach was improved machine learning (ML) method. comparative analysis, five popular ML algorithms i.e., decision tree (DT), k-nearest neigbhour (KNN), light gradient boosting (LGBM), random forest (RF) support vector machines (SVM) were used. findings show that accuracies 68.1, 73.1, 74.3, 74.5, 75.4% obtained for DT, KNN, SVM, LGBM, RF models, respectively. Similarly, F1 scores 66.6, 70.2, 71.8, 72.1 72.9%, models. result shows model gave best performance while DT performed least. Also, all models would classify with reasonable accuracy. Furthermore, feature importance done, found K d , T v t α, sun, cld most important input parameters classification. Lastly, vertical ( G VT ) VL estimated predicted proposed Upon predictions, observed ranged 14.7 24.6% 13.8 19.9%. Generally, predictions less than 20%, which good

Language: Английский

Citations

0