# Data vs classifiers, who wins?

arXiv: https://arxiv.org/abs/2107.07451 · HTML: https://mlsanity.com/paper/2107.07451/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 43 references that are in the map.
- First posted on arXiv 15 Jul 2021.

## Builds on (most cited first, 5 of 5)

- [Adversarial NLI: A New Benchmark for Natural Language Understanding](https://mlsanity.com/paper/1910.14599/) (2019)
- [OpenML: networked science in machine learning](https://mlsanity.com/paper/1407.7722/) (2014)
- [Decoding machine learning benchmarks](https://mlsanity.com/paper/2007.14870/) (2020)
- [Analysing Results from AI Benchmarks: Key Indicators and How to Obtain Them](https://mlsanity.com/paper/1811.08186/) (2018)
- [Reducing the Effects of Detrimental Instances](https://mlsanity.com/paper/1406.2237/) (2014)

## Built on by (most cited first, 2 of 2)

- [HPOBench: A Collection of Reproducible Multi-Fidelity Benchmark Problems for HPO](https://mlsanity.com/paper/2109.06716/) (2021)
- [carps: A Framework for Comparing N Hyperparameter Optimizers on M Benchmarks](https://mlsanity.com/paper/2506.06143/) (2025)

## Abstract (arXiv)

The experiments covered by Machine Learning (ML) must consider two important aspects to assess the performance of a model: datasets and algorithms. Robust benchmarks are needed to evaluate the best classifiers. For this, one can adopt gold standard benchmarks available in public repositories. However, it is common not to consider the complexity of the dataset when evaluating. This work proposes a new assessment methodology based on the combination of Item Response Theory (IRT) and Glicko-2, a rating system mechanism generally adopted to assess the strength of players (e.g., chess). For each dataset in a benchmark, the IRT is used to estimate the ability of classifiers, where good classifiers have good predictions for the most difficult test instances. Tournaments are then run for each pair of classifiers so that Glicko-2 updates performance information such as rating value, rating deviation and volatility for each classifier. A case study was conducted hereby which adopted the OpenML-CC18 benchmark as the collection of datasets and pool of various classification algorithms for evaluation. Not all datasets were observed to be really useful for evaluating algorithms, where only 10% were considered really difficult. Furthermore, the existence of a subset containing only 50% of the original amount of OpenML-CC18 was verified, which is equally useful for algorithm evaluation. Regarding the algorithms, the methodology proposed herein identified the Random Forest as the algorithm with the best innate ability.