Graph Convolutional Networks (GCNs) have gained prominence in collaborative filtering (CF) recommendation systems for capturing intricate signals using high-order structural data. However, GCN-based models focus solely on these signals, neglecting the sparse nature of data and overlooking important aspects like temporal signals in user preferences and baseline signals in users or items, leading to sub-optimal performance. To address these issues, this paper introduces a novel multi-component CF model that integrates GCNs with baseline and temporal components. The integrated model learns user and item representations from multiple perspectives, enhancing performance and robustness across various datasets. Experiments conducted on the MovieLens and Douban datasets demonstrate the superiority of this approach over state-of-the-art models, reducing RMSE by up to 4.7%, while improving NCDG by up to 5.1% compared to pure GCN-based CF ( https://github.com/tseesurenb/wise2024_v2.git ).