# Robust Taxi Fare Prediction Under Noisy Conditions: A Comparative Study of GAT, TimesNet, and XGBoost

arXiv: https://arxiv.org/abs/2507.20008 · HTML: https://mlsanity.com/paper/2507.20008/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 15 references that are in the map.
- First posted on arXiv 26 Jul 2025.

## Builds on (most cited first, 9 of 9)

- [Semi-Supervised Classification with Graph Convolutional Networks](https://mlsanity.com/paper/1609.02907/) (2016)
- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- [XGBoost: A Scalable Tree Boosting System](https://mlsanity.com/paper/1603.02754/) (2016)
- [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://mlsanity.com/paper/2012.07436/) (2020)
- [Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting](https://mlsanity.com/paper/2106.13008/) (2021)
- [FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecasting](https://mlsanity.com/paper/2201.12740/) (2022)
- [Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shift](https://mlsanity.com/paper/1810.11953/) (2018)
- [Energy-based Out-of-Distribution Detection for Graph Neural Networks](https://mlsanity.com/paper/2302.02914/) (2023)
- [Can bin-wise scaling improve consistency and adaptivity of prediction uncertainty for machine learning regression ?](https://mlsanity.com/paper/2310.11978/) (2023)

## Abstract (arXiv)

Precise fare prediction is crucial in ride-hailing platforms and urban mobility systems. This study examines three machine learning models-Graph Attention Networks (GAT), XGBoost, and TimesNet to evaluate their predictive capabilities for taxi fares using a real-world dataset comprising over 55 million records. Both raw (noisy) and denoised versions of the dataset are analyzed to assess the impact of data quality on model performance. The study evaluated the models along multiple axes, including predictive accuracy, calibration, uncertainty estimation, out-of-distribution (OOD) robustness, and feature sensitivity. We also explore pre-processing strategies, including KNN imputation, Gaussian noise injection, and autoencoder-based denoising. The study reveals critical differences between classical and deep learning models under realistic conditions, offering practical guidelines for building robust and scalable models in urban fare prediction systems.