Time-series Data Modeling Guided by Visual Information
Ke Zhang, Junzheng Wang, Xiang Zhang, Yu Shi, Ting Wang · 2023
Time-series data typically contain issues such as missing data and noise, which can impact the model's precision and stability. This paper proposes a Transformer structure-based visual information-guided temporal data modeling algorithm to address the issues as mentioned above. The algorithm effectively captures the time-series structure of the time-series data, thereby enhancing the model's precision and stability. To evaluate the performance of the proposed algorithm, a dataset containing visual information aligned with time-series data is compiled, and a comprehensive quantitative and qualitative analysis is performed. Conduct a comprehensive quantitative and qualitative analysis. The results indicate that visual information can assist time-series data in capturing the intricate dynamics of the time-series data, thereby enhancing the performance of the proposed algorithm and facilitating its comprehension. The results indicate that visual information can assist time-series data in capturing the complex dynamics of time-series data, and thus in comprehending and predicting their behavior and trends. The application of this algorithm will advance research in the field of modeling and predicting time series data. Applying this algorithm will advance research and practice in modeling and forecasting time series data.