Time series forecasting is crucial in various fields such as financial markets and weather prediction. Although mainstream deep learning models like RNNs and CNNs have made some progress in capturing short-term patterns, they still fall short in handling long-range dependencies and complex interactions. The transformer model enhances modeling capabilities through self-attention mechanisms, but its complexity and sensitivity to noise limit its applications. To address these challenges, we propose the KANMTS model, which integrates the advantages of Kolmogorov–Arnold networks (KAN) and multi-layer perceptrons (MLP). KANMTS leverages the non-linear mapping capabilities of KAN to improve the capture of complex patterns and dependencies in multivariate time series while maintaining the computational simplicity of MLP. Experimental results show that KANMTS outperforms existing methods in terms of forecasting performance and resource utilization efficiency, with a more pronounced accuracy improvement on large-scale datasets. We also explore the interpretability of KANMTS using symbolic regression and visualization methods. Furthermore, KANMTS demonstrates good generalization and integration capabilities, making it suitable for various application scenarios. This study provides a simple, efficient, and more interpretable solution for multivariate time series forecasting, advancing the field of time series analysis.

Recently, studies have reported that protein glycosylation plays an important role in the occurrence and development of cancer. Gastric cancer is a common cancer with high morbidity and mortality owing to most gastric cancers are discovered only at an advanced stage. Here, we aim to discover novel specific serum glycanbased biomarkers for gastric cancer. A lectin microarray with 50 kinds of tumor-associated lectin was used to detect the glycan profiles of serum samples between early gastric cancer and healthy controls. Then lectin blot was performed to validate the differences. The result of the lectin microarray showed that the signal intensities of 13 lectins showed significant differences between the healthy controls and early gastric cancer. Compared to the healthy, the normalized fluorescent intensities of the lectins PWA, LEL, and STL were significantly increased, and it implied that their specifically recognized GlcNAc showed an especially elevated expression in early gastric cancer. Moreover, the binding affinity of the lectins EEL, RCA-II, RCA-I, VAL, DSA, PHA-L, UEA, and CAL were higher in the early gastric cancer than in healthy controls. These glycan structures containing GalNAc, terminal Galβ 1-4 GlcNAc, Tri/tetraantennary N-glycan, β-1, 6GlcNAc branching structure, α-linked fucose residues, and Tn antigen were elevated in gastric cancer. While the two lectins CFL GNL reduced their binding ability. In addition, their specifically recognized N-acetyl-D-galactosamine structure and (α-1,3) mannose residues were decreased in early gastric cancer. Furthermore, lectin blot results of LEL, STL, PHA-L, RCA-I were consistent with the results of the lectin microarray. The findings of our study clarify the specific alterations for glycosylation during the pathogenesis of gastric cancer. The specific high expression of GlcNAc structure may act as a potential early diagnostic marker for gastric cancer.