Prediction of Landslide Displacement Based on the Combined VMD-Stacked LSTM-TAR Model

Remote Sensing, Journal Year: 2022, Volume and Issue: 14(5), P. 1164 - 1164

Published: Feb. 26, 2022

The volatility of the cumulative displacement landslides is related to influence external factors. To improve prediction nonlinear changes in landslide caused by influences, a new combined forecasting model has been proposed. Variational modal decomposition (VMD) was used obtain trend and fluctuation sequences original sequence displacement. First, we established stacked long short time memory (LSTM) network introduced rainfall reservoir water levels as influencing factors predict sequence; next, threshold autoregressive (TAR) sequence, following which were superimposed predicted landslide. Finally, VMD-stacked LSTM-TAR combination based on variational decomposition, network, built. Taking Baishuihe Three Gorges Reservoir area an example, through comparison with results VMD-recurrent neural network-TAR, VMD-back propagation VMD-LSTM-TAR, proposed noted have high accuracy, it provided novel approach for volatile

Language: Английский

---

Air Pollution Prediction Using Long Short-Term Memory Variants

Lecture notes on data engineering and communications technologies, Journal Year: 2024, Volume and Issue: unknown, P. 122 - 132

Published: Jan. 1, 2024

Language: Английский

---

A novel long-term water absorption and thickness swelling deep learning forecast method for corn husk fiber-polypropylene composite

Case Studies in Construction Materials, Journal Year: 2022, Volume and Issue: 17, P. e01268 - e01268

Published: June 29, 2022

Investigating long-term water absorption (WA) and thickness swelling (TS) behaviors of wood plastic composites demand long working hours high laboratory costs. However, using artificial intelligence methods, these can be predicted in far less time with a low degree error. This paper aims to predict the WA TS cornhusk fiber (CHF) propylene (PP) composite deep learning field's short-term memory (LSTM) method. We assessed network LSTM performance based on mean square error (MSE), root (RMSE), absolute (MAE), percentage (MAPE). The experimental tests were performed CHF/PP three different filler percentages over period 0–1500 h. predictions carried out for 200, 400, 600, 800, 1000 h construct database identify how many training data are required meet MAPE criterion 2% between actual data. results show that 200 is adequate method achieve this metric. Furthermore, metrics validate applicability proposed All manufacturing codes attached.

Language: Английский

---

An Enhancement Method Based on Long Short-Term Memory Neural Network for Short-Term Natural Gas Consumption Forecasting

Energies, Journal Year: 2023, Volume and Issue: 16(3), P. 1295 - 1295

Published: Jan. 26, 2023

Artificial intelligence models have been widely applied for natural gas consumption forecasting over the past decades, especially short-term forecasting. This paper proposes a three-layer neural network model that can extract key information from input factors and improve weight optimization mechanism of long memory (LSTM) to effectively forecast consumption. In proposed model, convolutional (CNN) layer is adopted features among various affecting computing efficiency. The LSTM able learn save long-distance state through gating overcomes defects gradient disappearance explosion in recurrent network. To solve problem encoding sequences as fixed-length vectors, attention (ATT) used optimize assignment weights highlight sequences. Apart comparisons with other popular models, performance robustness are validated on datasets different fluctuations complexities. Compared traditional two-layer (CNN-LSTM LSTM-ATT), mean absolute range normalized errors (MARNE) Athens Spata improved by more than 16% 11%, respectively. comparison single LSTM, back propagation network, support vector regression, multiple linear regression methods, improvement MARNE exceeds 42% Athens. coefficient determination 25%, even high-complexity dataset, Spata.

Language: Английский

---

A Multi-Attention Approach Using BERT and Stacked Bidirectional LSTM for Improved Dialogue State Tracking

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(3), P. 1775 - 1775

Published: Jan. 30, 2023

The modern digital world and associated innovative state-of-the-art applications that characterize its presence, render the current age a captivating era for many worldwide. These innovations include dialogue systems, such as Apple’s Siri, Google Now, Microsoft’s Cortana, stay on personal devices of users assist them in their daily activities. systems track intentions by analyzing speech, context looking at previous turns, several other external details, respond or act form speech output. For these to work efficiently, state tracking (DST) module is required infer conversation processing states up state. However, developing DST tracks exploit effectively accurately challenging. notable challenges warrant immediate attention scalability, handling unseen slot-value pairs during training, retraining model with changes domain ontology. In this article, we present new end-to-end framework combining BERT, Stacked Bidirectional LSTM (BiLSTM), multiple mechanism formalize classification problem address aforementioned issues. BERT-based encodes user’s system’s utterances. BiLSTM extracts contextual features mechanisms calculate between hidden utterance embeddings. We experimentally evaluated our method against approaches over variety datasets. results indicate significant overall improvement. proposed scalable terms sharing parameters it considers instances training.

Language: Английский

---

Aedes aegypti mosquito movements analysis and sex classification using computer vision and deep learning

Published: March 6, 2024

Mosquitoes are vectors of diseases, carrying viruses, parasites, and bacteria that infect millions people around the world. Understanding their flight patterns behaviours is crucial for disease modelling, ecological research, developing effective control methods. Traditional manual methods analysing mosquito records time-consuming limited, whilst automated could provide a viable alternative. In this study, recognition, monitoring, classification movements was done using artificial intelligence (AI), particularly, computer vision deep learning. Two experiments were carried out: first experiment assessed system's capacity to accurately detect classify direction various classifiers, with models such as Gated Recurrent Unit (GRU) Convolutional Neural Network model Long Short-Term Memory (CNN-LSTM). Results show high accuracy rate 96.67%. The second showed ability identify between male female Aedes aegypti mosquitoes CNN based on movement heatmaps. levels 89.84% 99.73%.

Language: Английский

---

Risk Analysis in Customer Relationship Management via QRCNN-LSTM and Cross-Attention Mechanism

Journal of Organizational and End User Computing, Journal Year: 2024, Volume and Issue: 36(1), P. 1 - 22

Published: Sept. 28, 2024

Risk analysis is an important business decision support task in Customer Relationship Management (CRM), involving the identification of potential risks or challenges that may affect customer satisfaction, retention rates, and overall performance. To enhance risk CRM, this paper combines advantages QRCNN-LSTM cross-attention mechanisms for modeling. The model sequence modeling with deep learning architectures commonly used natural language processing tasks, enabling capture both local global dependencies data. mechanism enhances interactions between different input data parts, allowing to focus on specific areas features relevant CRM analysis. By applying analysis, empirical evidence demonstrates approach can effectively identify provide data-driven decisions.

Language: Английский

---

The analysis of agricultural Internet of things product marketing by deep learning

The Journal of Supercomputing, Journal Year: 2022, Volume and Issue: 79(4), P. 4602 - 4621

Published: Sept. 28, 2022

Language: Английский

---

Intensified customer churn Prediction: Connectivity with weighted Multi-Layer Perceptron and Enhanced Multipath Back Propagation

Expert Systems with Applications, Journal Year: 2024, Volume and Issue: unknown, P. 125993 - 125993

Published: Dec. 1, 2024

Language: Английский

---

Revealing User Behavior Trends through Time-series Analysis of Google Analytics Data

Published: Aug. 21, 2024

Language: Английский

---