Meta-Learning Enhanced Trade Forecasting: A Neural Framework Leveraging Efficient Multicommodity STL Decomposition DOI Open Access
Bohan Ma, Yushan Xue, J. Chen

et al.

International Journal of Intelligent Systems, Journal Year: 2024, Volume and Issue: 2024, P. 1 - 21

Published: April 3, 2024

In the dynamic global trade environment, accurately predicting values of diverse commodities is challenged by unpredictable economic and political changes. This study introduces Meta-TFSTL framework, an innovative neural model that integrates Meta-Learning Enhanced Trade Forecasting with efficient multicommodity STL decomposition to adeptly navigate complexities forecasting. Our approach begins partition value sequences into seasonal, trend, residual elements, identifying a potential 10-month cycle through Ljung–Box test. The employs dual-channel spatiotemporal encoder for processing these components, ensuring comprehensive grasp temporal correlations. By constructing spatial graphs leveraging correlation matrices graph embeddings introducing fused attention multitasking strategies at decoding phase, surpasses benchmark models in performance. Additionally, integrating meta-learning fine-tuning techniques enhances shared knowledge across import export predictions. Ultimately, our research significantly advances precision efficiency forecasting volatile scenario.

Language: Английский

Companies’ green competitiveness: Justifying the role of marketing determinants DOI Open Access

Olena Chygryn,

Oleksii Lyulyov, Tetyana Pimonenko

et al.

AIP conference proceedings, Journal Year: 2024, Volume and Issue: 3033, P. 020018 - 020018

Published: Jan. 1, 2024

Today, companies' green competitiveness is developing under the influence of a wide set economic, social, environmental, corporate and marketing determinants. The purpose paper to justify role determinants in competitiveness. provided analysis semantic core information-commercial scientific analytics on search queries related using Google Analytics Trends allowed identifying following groups its provision: sustainable strategic development (green strategies, supply chains, logistics, pricing); media information networks, platforms, digital tools, web tools); targeted brand, advertising, promotion). systematic combination correlation Godrick-Prescott smoothing filtering method time series issues enterprises three different vector trends: increasing level commercial interest field (at request Internet users) accompanied by community both trend cyclical components (positive statistically significant correlation); an uneven growth part users (negative object dominant researchers are such as brand promotion (results components).

Language: Английский

Citations

0

An approach in the process of constructing a research methodology in management science DOI Open Access
Henryk Dźwigoł

AIP conference proceedings, Journal Year: 2024, Volume and Issue: 3033, P. 020001 - 020001

Published: Jan. 1, 2024

The main idea of the article is to draw attention necessity combining multiple methods in research process with a view obtaining comprehensive answer questions posed. an attempt demonstrate rationale for using triangulation processes. Triangulation presented as methodology procedure and condition enhancing reliability research. advantages triangulating are also discussed. highlights importance raising quality management sciences; it respond challenges civilization, which determined by science economy.

Language: Английский

Citations

0

Early diagnosis of the causes of Holstein cows’ extinction by polymorphisms BGH-ALUI and BIGF-1-SNABI DOI Open Access

B Makhatov,

Gulzhan Mussayeva,

Gulshat Shaikamal

et al.

AIP conference proceedings, Journal Year: 2024, Volume and Issue: 3033, P. 020025 - 020025

Published: Jan. 1, 2024

In order to ensure sustainable development and system management of livestock farming, the expansion scientific research boundaries interdisciplinary focus science is becoming increasingly important. this regard, at present stage in field animal breeding, especially intensive possibility early forecasting long-term productive use animals with high levels productivity important, which turn puts forward need combine efforts different branches science. For example, selection more attention paid not only appearance, but also genotype animals, particular isolation individual species genotype, a certain allelic state genes. last years, scientists agrarian university have been actively studying somatotropin cascade genes, it was decided deepen prevalence alleles gene breed, Holstein breed cattle, widely spread Kazakhstan. The laboratory DNA-typing on polymorphisms bGH-AluI, bIGF-1-SnaBI carried out analysis genetic structure researched populations out. genotypes each were established. aim article evaluate association pair combinations life expectancy cows Kazakh breeding studied. there observed strengthening phenotypic effects towards an increase milk indicators combination¹ 1 bGH-AluILL-bIGF-1-SnaBIAA. record diplotype bGH-AluILV-bIGF-1-SnaBIAA, combined effect relation separately taken by trait total yield.

Language: Английский

Citations

0

Economic and mathematical modelling of estimating the use of basic production resources of agricultural formations DOI Open Access
Zharas Ainakulov,

Kulmuhanbet Akhmetov,

Серик Оспанов

et al.

AIP conference proceedings, Journal Year: 2024, Volume and Issue: 3033, P. 020022 - 020022

Published: Jan. 1, 2024

Based on the materials of specific agricultural formations in Almaty region, this paper presents a methodological approach to assessing effectiveness using resource potential. Approved economic-mathematical model optimizing production and industrial structure, justifying optimal parameters elements potential through which there are revealed reserves for increasing efficiency The research results show that use economic mathematical optimization models will significantly increase formations. Practical recommendations given substantiate all types enterprises region based methodology.

Language: Английский

Citations

0

Meta-Learning Enhanced Trade Forecasting: A Neural Framework Leveraging Efficient Multicommodity STL Decomposition DOI Open Access
Bohan Ma, Yushan Xue, J. Chen

et al.

International Journal of Intelligent Systems, Journal Year: 2024, Volume and Issue: 2024, P. 1 - 21

Published: April 3, 2024

In the dynamic global trade environment, accurately predicting values of diverse commodities is challenged by unpredictable economic and political changes. This study introduces Meta-TFSTL framework, an innovative neural model that integrates Meta-Learning Enhanced Trade Forecasting with efficient multicommodity STL decomposition to adeptly navigate complexities forecasting. Our approach begins partition value sequences into seasonal, trend, residual elements, identifying a potential 10-month cycle through Ljung–Box test. The employs dual-channel spatiotemporal encoder for processing these components, ensuring comprehensive grasp temporal correlations. By constructing spatial graphs leveraging correlation matrices graph embeddings introducing fused attention multitasking strategies at decoding phase, surpasses benchmark models in performance. Additionally, integrating meta-learning fine-tuning techniques enhances shared knowledge across import export predictions. Ultimately, our research significantly advances precision efficiency forecasting volatile scenario.

Language: Английский

Citations

0