A Convolutional Transformer Model for Multivariate Time Series Prediction

Kim, Dong-Keon; Kim, Kwangsu

doi:10.1109/access.2022.3203416

Cited by 12 publications

(3 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following this preprocessing step, the data is inputted into a CNN(Convolutional Neural Network) for classification, capitalizing on its effectiveness in analyze a series of abstract symbols to determine their meaning and context. Another contribution involves in suggesting a model that combines CNN(Convolutional Neural Network)s with a transformer architecture to predict the multivariate time series data [5]. This approach captures both temporal patterns and intervariable relationships concurrently, leading to improved prediction accuracy in contrast to current methods.…”

Section: Literature Surveymentioning

confidence: 99%

Food Delivery Time Prediction Using CNN- LSTM

2024

IJSREM

View full text Add to dashboard Cite

To tackle the challenge of predicting food delivery times, the study introduces a novel approach crucial for streamlining food delivery service operations. The proposed model combines CNN (Convolution Neural Network)s and (LSTM) Long Short-Term Memory networks, leveraging the unique strengths of each design. The CNN component analyzes delivery data trends on a daily and weekly basis, identifying geographical patterns such as rush hours and seasonal variations. Meanwhile, the LSTM component focuses on recognizing spatiotemporal relationships and retaining data over time to effectively model sequential data. Compared to standard procedure, the model's simultaneous incorporation of spatial and temporal information leads to more precise predictions. This enhanced accuracy holds proven strategy for food delivery services, enabling them to more accurately estimate delivery times, reduce delays, and ultimately improve overall customer satisfaction. Keywords: (LSTM) Long short term memory, Time prediction, CNN (Convolution Neural Networks).

show abstract

Section: Literature Surveymentioning

confidence: 99%

Food Delivery Time Prediction Using CNN- LSTM

2024

IJSREM

View full text Add to dashboard Cite

show abstract

“…where temporal data processing is mainly reflected in time series prediction tasks [7]. The Transformer model consists of a position encoder, encoder, and decoder, and its structure is shown in Figure 1 below.…”

Section: International Journal Of Scientific Advances Issn: 2708-7972mentioning

confidence: 99%

Application of Autoformer to Short-Term Traffic Flow Prediction

Luo,

2024

International Journal of Scientific Advances

View full text Add to dashboard Cite

This paper investigates the application of Autoformer model in the field of short-term traffic flow prediction. Traffic flow prediction is crucial for urban planning and traffic management, and is important for relieving traffic congestion and improving traffic efficiency. Autoformer, as a Transformer model based on the self-attention mechanism, provides a powerful modeling capability for the task of traffic flow prediction through its ability to automatically learn feature relationships. The paper first introduces the background and importance of traffic flow prediction and outlines current prediction methods and their limitations. Subsequently, the structure and working principle of Autoformer are described in detail, elucidating the advantages over traditional models. Autoformer’s self-attention mechanism can effectively capture the long-range dependencies in the input sequences and better adapt to the dynamic changes of traffic flow. To verify the performance of Autoformer in short-time traffic flow prediction, experiments are conducted using real traffic flow datasets. The results show that Autoformer significantly improves the prediction accuracy and generalization ability compared to traditional time series models. In addition, model interpretive analysis was conducted, revealing the advantages of Autoformer for automatic extraction of traffic flow features and correlation modeling.

show abstract

“…Chen Y et al [13] developed the TPM model, a stock price trend prediction model based on the encoderdecoder framework, which can adaptively predict the volatility and duration of stock prices. Kim D K et al [14] extracted temporal features of input data using CNN and explained the correlations between variables using attention mechanisms. Mishev K et al [15] explored the effectiveness and performance of various sentiment analysis methods combining text representation techniques and machine learning classifiers.…”

Section: Introductionmentioning

confidence: 99%

HPMG-Transformer: HP Filter Multi-Scale Gaussian Transformer for Liquor Stock Movement Prediction

Huang

2024

IEEE Access

View full text Add to dashboard Cite

Predicting complex, volatile, and nonlinear financial stock prices is a challenging task due to the multitude of factors and inherent uncertainty that influence the financial market. This paper proposes a novel approach based on a neural network model that combines the Hodrick-Prescott (HP) filter and a Multi-Scale Gaussian transformer to address these challenges. The proposed method enhances the local features of the model by employing a multi-scale Gaussian transformer. Firstly, the time series of stocks is decomposed into long-term and short-term fluctuations using the HP filter. Next, the encoded long-term and short-term series are input into a multi-scale Gaussian transformer. Finally, a Multi-Scale Gaussian prior is introduced to further enhance the local features of the transformer and improve the relative positional information features of the time series. Compared to popular recurrent neural networks such as RNN, LSTM, GRU, and state-of-the-art baseline models, ours model (HPMG-Transformer) has the unique advantage of capturing both extremely long-term and short-term dependencies in stock time series. Experimental results demonstrate the significant advantages of our proposed model in predicting stock trends in the China A-shares market, New York Stock Exchange (NYSE) and NASDAQ market .INDEX TERMS stock price, artificial neural network, time series, HP filter, transformer.

show abstract

A Convolutional Transformer Model for Multivariate Time Series Prediction

Cited by 12 publications

References 35 publications

Food Delivery Time Prediction Using CNN- LSTM

Food Delivery Time Prediction Using CNN- LSTM

Application of Autoformer to Short-Term Traffic Flow Prediction

HPMG-Transformer: HP Filter Multi-Scale Gaussian Transformer for Liquor Stock Movement Prediction

Contact Info

Product

Resources

About