Skip to main content
CARLA HQ
Agriculture, Forestry & Fishing Commerce & Industry use casesNACE A03.21RegressionTarget: number_of_ringsOpenML #42726 ↗

Abalone

Benchmark aquaculture dataset containing 4,177 records designed to predict the age and shell ring count of abalones using non-destructive physical biometric measurements.

✨ Try in CarlaInstall Chrome Extension ↗100% In-Browser WebGPU • Zero Cloud Upload
4,177Records (Rows)
8Predictive Features
8 / 1Numeric / Categorical
0.0%Missing Value Ratio
2.03RMSE
11.68sWebGPU Mean Split
BUSINESS CONTEXT & OBJECTIVEAquaculture & Fisheries Management (A03.21)

Business Objective: Abalone

01Business Context

In commercial shellfish aquaculture and marine fisheries management, determining abalone maturity traditionally requires destructive cone slicing, chemical staining, and microscopic ring counting under a lab microscope.

02Analytical Objective

Predict the exact number of shell rings (representing age in years + 1.5) non-destructively based on rapidly measurable physical dimensions and mass attributes.

03Economic & Decision Impact

Eliminates destructive testing and tedious lab inspections, accelerating commercial grading throughput while preventing regulatory fines from harvesting underage stock. Underestimating age results in premature harvesting and revenue loss, while overestimating leads to unnecessary feed and holding expenditures.

ML Benchmark Narrative

Standard classifiers like Random Forest, LightGBM, and XGBoost reach 24–28% top-1 exact accuracy on the full 28-class multiclass setup due to substantial morphological overlap between adjacent age groups. In contrast, tabular foundation models like TabICL and Carla leverage continuous in-context spatial reasoning directly within spreadsheet workflows to deliver accurate age estimations without manual feature scaling or hyperparameter tuning.

Source Origin:Department of Primary Industry and Fisheries, Tasmania & University of Tasmania
Creator:Warwick J. Nash, Tracy L. Sellers, Simon R. Talbot, Andrew J. Cawthorn, Wes B. Ford, and Sam Waugh (1994)
License:Creative Commons Attribution 4.0 International (CC BY 4.0)
FEATURE SPECIFICATION & DATA DICTIONARY

Target & Feature Column Definitions

Exhaustive business definitions, measurement units, roles, and target variables across all 9 columns.

ColVariable / Feature NameRoleTypeUnit / ScaleDescription & Business Meaning
Anumber_of_ringsNumber of Shell RingsTARGETNUMrings
Count of growth rings stained and identified under microscope; adding 1.5 yields the chronological age in years.
BsexBiological SexFeatureCAT
Biological sex classification of the abalone: 'M' (Male), 'F' (Female), or 'I' (Infant/Sexually Immature).
Clength_mmShell LengthFeatureNUMmm
Longest linear dimension of the abalone shell (scaled in mm).
Ddiameter_mmShell DiameterFeatureNUMmm
Measurement of the shell diameter perpendicular to the length axis (scaled in mm).
Eheight_mmShell HeightFeatureNUMmm
Vertical height of the shell including the meat core (scaled in mm).
Fwhole_weight_gWhole WeightFeatureNUMgrams
Total live weight of the intact abalone organism (scaled in grams).
Gshucked_weight_gShucked Meat WeightFeatureNUMgrams
Weight of the edible abalone meat extracted after shucking (scaled in grams).
Hviscera_weight_gViscera WeightFeatureNUMgrams
Weight of the internal gut organs measured after post-harvest bleeding (scaled in grams).
Ishell_weight_gDry Shell WeightFeatureNUMgrams
Weight of the calcified shell structure measured after thermal drying (scaled in grams).
DATASET PREVIEW

Interactive Data Table

Explore rows, feature values, and target labels for Abalone.

Loading dataset...
LOCAL EXECUTION TELEMETRY

TabICLv2 WebGPU Benchmark Results

Complete performance metrics on persisted IID or non-IID evaluation splits, executed 100% locally in the browser sandbox.

Mean Split Latency?Mean wall-clock execution time per persisted evaluation split running 100% locally via WebGPU inside Chrome.Explore Latency Guide →
11.68s9 splits • 8 ensembles/split
Total Evaluation Runtime?Total wall-clock duration across all selected splits, including in-context encoding, WebGPU shader execution, and result aggregation.Explore Split Protocol →
105.14sIncludes warmup & sync
Evaluation Protocol?Precomputed row indices ensure the browser and Python runners evaluate identical out-of-sample observations without leakage.Explore Split Protocol →
3×3 SplitsRepeated IID • 9 total
TabICLv2 Score?Primary task performance score (RMSE) achieved by TabICLv2 across out-of-sample evaluation splits.Explore Metric Formula →
2.03RMSE
Evaluation SplitTrain RowsTest Rows
Split Duration?Wall-clock inference time taken for this evaluation split running 100% locally via WebGPU.Learn more →
R² Score?Coefficient of determination measuring variance explained by the model.Learn more →
MAE?Mean Absolute Error in original target units, robust to outliers.Learn more →
RMSE?Root Mean Squared Error, penalizing large residual deviations.Learn more →
MAPE?Mean Absolute Percentage Error relative to true target values.Learn more →
RMSLE?Root Mean Squared Logarithmic Error, measuring proportional error scales.Learn more →
R1 / F12,7841,39313.81s0.63541.381.91
R1 / F22,7851,39213.77s0.57351.452.09
R1 / F32,7851,39213.70s0.59111.462.11
R2 / F12,7841,39313.57s0.60791.442.06
R2 / F22,7851,3924.60s0.60931.482.12
R2 / F32,7851,39213.72s0.59051.371.91
R3 / F12,7841,39313.49s0.59911.421.99
R3 / F22,7851,39213.86s0.57691.512.21
R3 / F32,7851,3924.60s0.63371.361.89
Mean ± Std11.68s0.6019 ± 0.02091.42
UNIFIED BENCHMARK LEADERBOARD

Foundation Models vs Traditional ML Baselines

Side-by-side evaluation of our in-browser Carla engine (WebGPU), open-source tabular foundation models, and OpenML baselines.

RankAlgorithm / ModelModel FamilyRuntime
R² Score?Coefficient of determination (R²) measuring proportion of variance explained.Google ML Guide & Details →
MAE?Mean Absolute Error measuring average prediction error magnitude.Learn more →
RMSE?Root Mean Squared Error penalizing large prediction errors.Learn more →
Reference / Repo
#1Google TabFM v1.0Tabular Foundation ModelPyTorch/Python0.60531.42google-research/tabfm ↗
#2TabPFN v3Tabular Foundation ModelPyTorch/Python0.60441.42PriorLabs/tabpfn ↗
#3EXAONE TabularTabular Foundation ModelPyTorch/Python0.60291.42LGAI-Research/EXAONE-Tabular ↗
#4Carla Engine (TabICLv2 WebGPU)Tabular Foundation ModelBrowser0.60191.42100% In-Browser
#5Full TabICLv2 (PyTorch Reference)Tabular Foundation ModelPyTorch/Python0.60141.42soda-inria/tabicl ↗
#6nanotabicl VanillaTabular Foundation ModelPyTorch/Python0.60121.42soda-inria/nanotabicl ↗
#7Streaming nanotabiclTabular Foundation ModelPyTorch/Python0.59941.42soda-inria/nanotabicl ↗
STATISTICAL DISTRIBUTIONS

Variable Schema & Summary Distributions

Observed numerical ranges, category cardinalities, missing rates, and sample values across 4,177 rows.

ColVariable NameTypeMissingDistinctSummary Stats / DistributionSample Values
Anumber_of_ringsTARGETNUM028[1, 29] μ=9.9 σ=3.21579
BsexCAT033 cats: M, I, FMMF
Clength_mmNUM0134[0.075, 0.815] μ=0.5 σ=0.10.4550.350.53
Ddiameter_mmNUM0111[0.055, 0.65] μ=0.4 σ=0.10.3650.2650.42
Eheight_mmNUM051[0, 1.13] μ=0.1 σ=0.00.0950.090.135
Fwhole_weight_gNUM02,429[0.002, 2.8255] μ=0.8 σ=0.50.5140.22550.677
Gshucked_weight_gNUM01,515[0.001, 1.488] μ=0.4 σ=0.20.22450.09950.2565
Hviscera_weight_gNUM0880[0.0005, 0.76] μ=0.2 σ=0.10.1010.04850.1415
Ishell_weight_gNUM0926[0.0015, 1.005] μ=0.2 σ=0.10.150.070.21
FREQUENTLY ASKED QUESTIONS

Frequently Asked Questions: Abalone

Common questions regarding the Abalone dataset, machine learning task formulations, and in-browser tabular inference.

What is the Abalone dataset used for?

The Abalone dataset is a canonical tabular benchmark used to predict the biological age (number of shell rings) of abalone mollusks from non-destructive physical measurements such as length, diameter, height, and weight.

Can I run zero-server predictions on this dataset in Google Sheets?

Yes, Carla HQ enables zero-server in-context tabular prediction directly in spreadsheets using TabICL foundation models.

What machine learning models perform best on the Abalone dataset?

Gradient boosted trees (LightGBM, XGBoost) and in-context tabular foundation models (TabICL / Carla) achieve the highest predictive accuracy by capturing non-linear relationships between physical dimensions and ring counts.

LIVE EVALUATION IN GOOGLE SHEETS

Test TabICLv2 on Abalone Yourself

Open the pre-loaded Google Sheet and let Carla configure the target and task for local, zero-cloud tabular machine learning.

Step 1

Launch Carla

Click once to open the spreadsheet and Carla side panel together.

Step 2

Review the Setup

Carla selects the dataset target and task from this page automatically.

Step 3

Evaluate & Predict

Run predictions and compute metrics with zero server uploads.