Publication
Building a Forecasting Model for Interval Time Series Based on Improvements in Data Interpolation: Accepted October 2026
| Summary: | This study proposes a novel forecasting model for interval time series and makes several significant contributions. First, the variation series is partitioned into an appropriate number of clusters. Second, a data interpolation method based on a fuzzy clustering technique is introduced. Finally, a forecasting principle is constructed by combining deep learning and statistical methods. The proposed model is described in detail along with its implementation steps and can be applied to real-world series through a program developed in the Python programming language. Experimental results on various real-world series from different domains demonstrate that the proposed model is stable, effective, and outperforms many existing models. |
|---|---|
| Main Authors: | Lethi-Thu, Thuy |
| Other Authors: | Nguyen-Ngoc, Nhi; Thai-Minh, Trong; Vo-Van, Tai; Vo-Van, Tai |
| Subject: | deep learning forecasting model interval time series interpolation |
| Year: | 2026 |
| Country: | Portugal |
| Document type: | article |
| Access type: | unknown |
| Associated institution: | Instituto Nacional de Estatística |
| Language: | English |
| Origin: | REVSTAT-Statistical Journal |
| Summary: | This study proposes a novel forecasting model for interval time series and makes several significant contributions. First, the variation series is partitioned into an appropriate number of clusters. Second, a data interpolation method based on a fuzzy clustering technique is introduced. Finally, a forecasting principle is constructed by combining deep learning and statistical methods. The proposed model is described in detail along with its implementation steps and can be applied to real-world series through a program developed in the Python programming language. Experimental results on various real-world series from different domains demonstrate that the proposed model is stable, effective, and outperforms many existing models. |
|---|
Funded activities
No funded projects found.