A novel attLSTM framework combining the attention mechanism and bidirectional LSTM for demand forecasting

Expert Systems with Applications, Год журнала: 2024, Номер 254, С. 124409 - 124409

Опубликована: Июнь 7, 2024

Язык: Английский

---

The Predictive Grey Forecasting Approach for Measuring Tax Collection

Journal of risk and financial management, Год журнала: 2024, Номер 17(12), С. 558 - 558

Опубликована: Дек. 13, 2024

Taxation serves as a vital lifeline for government revenue, directly contributing to national development and the welfare of its citizens. Ensuring efficiency effectiveness tax collection process is essential maintaining sustainable economic framework. This study investigates (a) trends patterns direct collection, (b) cost (c) proportion in total (d) tax-to-GDP ratio India. By utilizing novel grey forecasting model (GM (1,1)), this attempted predict future India’s collections, through which it aims provide concurrent accurate outlook on ensuring resources are optimally allocated country’s growth. Results revealed that has consistently increased past two decades, also improved significantly. On contrary, decreased regularly, indicating collection. Forecasting shows from expected reach INR 30.67 trillion 2029–30, constituting around 54.41% tax, leaving behind collections indirect at 25.70 trillion. Such findings offer insights could enhance revenue management strategies with policy decisions relevant economists, government, other stakeholders understand

Язык: Английский

---

Study on Demand Forecast and Influencing Factors of Fruit and Vegetable Cold Chain Logistics in Guangzhou City—Based on Grey Theory

Operations Research and Fuzziology, Год журнала: 2024, Номер 14(04), С. 125 - 133

Опубликована: Янв. 1, 2024

Язык: Английский

---

Prediction of provincial Digital Economy Development Index based on grey combination forecasting model

Grey Systems Theory and Application, Год журнала: 2024, Номер unknown

Опубликована: Ноя. 6, 2024

Purpose In many instances, the data exhibits periodic and trend characteristics. However, indices like Digital Economy Development Index (DEDI), which pertains to science, technology, policy economy, may occasionally display erratic behaviors due external influences. Thus, address unique attributes of digital this study integrates principle information prioritization with nonlinear processing techniques accurately forecast rapid anomalous data. Design/methodology/approach The proposed method utilizes new priority GM(1,1) model alongside an optimized BP neural network achieved through gradient descent technique (GD-BP). Initially, provincial Economic (DEDI) is derived using entropy weight approach. Subsequently, original time response equation undergoes alteration initial value, parameter fine-tuned Particle Swarm Optimization (PSO). Next, GD-BP addresses residual error. Ultimately, prediction outcome grey combination forecasting (GCFM) by merging findings from both NIPGM(1,1) Findings Using DEDI Jiangsu Province as a case study, researchers demonstrate effectiveness model. This achieves mean absolute percentage error 0.33%, outperforming other methods. Research limitations/implications First all, limited access, it impossible obtain more comprehensive dataset related Province. Secondly, according test results GCFM 2011 2020 2021 2023, can be seen that are consistent actual development situation, but cannot guarantee correctness long-term forecasting, so only suitable for short-term forecasting. Originality/value article proposes based on principles processing.

Язык: Английский

---

Parameter optimization of a novel grey model EGM(1, 1, $ \sum{} $t^c) and its application in China's GDP per capita prediction

Journal of Industrial and Management Optimization, Год журнала: 2024, Номер 21(1), С. 454 - 473

Опубликована: Июль 2, 2024

The GM(1, 1) model, i.e. the first-order univariate grey is most important prediction but it considerable inaccurate in of fast-growing sequences. To improve model prediction, this paper makes improvements following two aspects based on traditional model: (1) improves accumulated generating sequence original sequence, properly making a quantitative transformation sequence; (2) model's structure, extending action into superposition power function expression. We call new extended EGM(1, 1, $ \sum $t^c) with action, which function. gives parameter estimation method and time response equation for simulation prediction. builds proposed compares to seven other models predicting China's GDP per capita. Results show that built has high precision, its precision significantly superior comparison models.

Язык: Английский

---

A Diversified Integrated Model for Seasonal Product Demand Prediction

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Сен. 18, 2024

Product demand forecasting is the core link of an intelligent supply chain. The article discusses characteristics seasonal fast-moving consumer goods and presents a diversified stacked regression model (RXOEL-X) that combines linear multi-machine learning models. This utilizes stacking strategy adopts ElasticNet model, combined with L1 L2 regularization to handle complex relationships in data prevent overfitting. Empirical evaluation using real from leading beverage companies demonstrates model's superiority over other time series techniques for smart chains.

Язык: Английский

---

Research on demand prediction model and application of sustainable logistics of fresh aquatic products based on machine learning

Опубликована: Июнь 21, 2024

Язык: Английский

---

Prediction Method of Tea Disease Based on HP Filter and Ant Colony Grey Model

Опубликована: Дек. 16, 2023

Язык: Английский

---