Search papers, labs, and topics across Lattice.
This paper introduces MARCUS, a novel model for rent prediction that leverages missing data as a valuable contextual signal rather than treating it as mere noise. By employing a three-stage approach鈥擨ntra Learning, Inter Learning, and Fusion鈥擬ARCUS effectively encodes observed features alongside missing patterns and enhances cross-modal interactions based on modality reliability. The model outperforms existing methods, achieving a 51.35% reduction in mean absolute error (MAE) on Sydney data and 12.62% on New York data, highlighting the significance of incorporating missingness in urban data representation.
Treating missing data as a contextual signal, MARCUS dramatically improves rent prediction accuracy, slashing MAE by over 51% in some cases.
Multimodal urban data has expanded the applications of urban region representation learning, such as functional zone identification and real estate appraisal, but also introduces challenges caused by data incompleteness. Existing studies usually handle missing data through imputation, treating missingness as noise while ignoring its potential semantic value. To address this issue, we propose MARCUS, a missing-aware region representation model that treats missingness as a contextual urban signal. MARCUS models missingness in three stages: Intra Learning jointly encodes observed features and missing patterns, Inter Learning estimates modality reliability to guide cross-modal interaction, and Fusion uses missing-aware and time-aware gating to generate the final region embedding. We apply MARCUS to rent prediction, a task with long-term trends and seasonal fluctuations, using real-world datasets from Sydney and New York. Experimental results show that MARCUS achieves state-of-the-art performance, reducing MAE by 51.35% on Sydney and 12.62% on New York compared with the best baselines. Additional experiments, including an imputation-based ablation study and randomized additional-missingness analysis, further demonstrate the effectiveness of the proposed method.
