Skip to content
Contact UsStart Building

Automotive & Resale

Used-car asking prices

Estimate a used car’s asking price to support listing-price decisions.
AI-generated representative illustration of a used compact hatchback with realistic paint wear and a neutral clean silhouette.
AI-generated task illustration. Representative scene; not a customer site or benchmark observation.
From known examples to a prediction

Available inputs

  • Model, engine power and vehicle age
  • Mileage, previous owners and location

Known examples

1,538 recorded examples

Prediction target

The listed vehicle asking price
2.8%lower prediction error

vs. LightGBM (tuned ensemble), the strongest conventional result in this benchmark.

Used car pricesRMSE · Lower is better

1,538 examples · 7 inputs · 30 evaluation splits

01000
  1. Causilo
    708.7920
  2. LightGBM
    729.4337
  3. CatBoost
    732.6860
  4. Random forest
    735.3989
  5. Extra trees
    735.6991
  6. XGBoost
    737.1119
  7. Linear regression
    764.2413
  8. K-nearest neighbors
    782.1517

The task in practice

The benchmark contains 1,538 historical Fiat 500 listings with seven inputs covering vehicle specifications, usage history and location. The target is the seller’s asking price, rather than a completed sale price.

  1. Use listings with asking prices. Pair historical Fiat 500 listing attributes with their recorded prices.
  2. Describe another listing. Supply the same specification, usage and location fields.
  3. Review the proposed asking price. Use the estimate as one reference when comparing similar listings and setting a proposed asking price.

Scope. The target is an asking price, not a completed sale, and the task covers the Fiat 500 family. Validate on recent listings before using estimates in a new market.

Dataset, sources and evaluation

Used car prices dataset. Original source and curated OpenML task linked below. Original dataset source.

The TabArena-curated OpenML version (46907) supplies the dataset and recorded target, price. Comparison scores come from the retained results_per_split_official.csv evaluation snapshot. These pages describe the task and compare the recorded means; source records are not republished here.

RMSE measures prediction error on the dataset’s recorded target scale. A lower value is better. The percentage compares the reported mean with the strongest conventional result; it does not measure a change in the target itself.

Causilo uses its default configuration. Each conventional family uses its strongest reported mean across default, tuned and tuned-ensemble settings. The chart compares reported means, without a claim of statistical significance, customer deployment or operational savings.

Relative error change = (strongest conventional mean − Causilo mean) ÷ strongest conventional mean × 100. A negative change is shown as higher error.

How to read these comparisons · All configurations and original values.