phoneme_sound_classOpenML #1489 ↗Phoneme
Benchmark dataset comprising 5,404 acoustic observations across 5 spectral harmonic features designed to classify speech vowel phonemes into nasal versus oral sound categories.
Business Objective: Phoneme
01Business Context
Analytical speech recognition pipelines and automated voice assistants require robust, speaker-independent phonetic feature extraction to accurately transcribe spoken language in real time under noisy conditions.
02Analytical Objective
Classify acoustic vowel segments as either nasal or oral based on the energy-normalized amplitudes of their first five spectral harmonics.
03Economic & Decision Impact
Accurate phoneme disambiguation prevents cascading phonetic transcription errors in automated speech-to-text and IVR systems, directly reducing downstream semantic misunderstanding and customer call escalation costs.
Standard tree-based ensembles (XGBoost, LightGBM, and Random Forest) routinely achieve between 87% and 90% classification accuracy on this benchmark by capturing complex nonlinear spectral harmonic boundaries. Tabular foundation models (TabICL / Carla) deliver comparable zero-shot classification performance directly in-context without manual hyperparameter optimization or local model training.
Target & Feature Column Definitions
Exhaustive business definitions, measurement units, roles, and target variables across all 6 columns.
| Col | Variable / Feature Name | Role | Type | Unit / Scale | Description & Business Meaning |
|---|---|---|---|---|---|
| A | phoneme_sound_classPhoneme Sound Class | TARGET | NUM | — | Target phonetic articulation category representing nasal vowel sounds (Class 1) versus oral vowel sounds (Class 2). |
| B | harmonic_1_normalized_amplitudeHarmonic 1 Normalized Amplitude | Feature | NUM | ratio | Normalized amplitude of the 1st spectral harmonic (AH1/Ene), signed positive for local spectral maxima and negative otherwise. |
| C | harmonic_2_normalized_amplitudeHarmonic 2 Normalized Amplitude | Feature | NUM | ratio | Normalized amplitude of the 2nd spectral harmonic (AH2/Ene), signed positive for local spectral maxima and negative otherwise. |
| D | harmonic_3_normalized_amplitudeHarmonic 3 Normalized Amplitude | Feature | NUM | ratio | Normalized amplitude of the 3rd spectral harmonic (AH3/Ene), signed positive for local spectral maxima and negative otherwise. |
| E | harmonic_4_normalized_amplitudeHarmonic 4 Normalized Amplitude | Feature | NUM | ratio | Normalized amplitude of the 4th spectral harmonic (AH4/Ene), signed positive for local spectral maxima and negative otherwise. |
| F | harmonic_5_normalized_amplitudeHarmonic 5 Normalized Amplitude | Feature | NUM | ratio | Normalized amplitude of the 5th spectral harmonic (AH5/Ene), signed positive for local spectral maxima and negative otherwise. |
Interactive Data Table
Explore rows, feature values, and target labels for Phoneme.
TabICLv2 WebGPU Benchmark Results
Complete performance metrics on persisted IID or non-IID evaluation splits, executed 100% locally in the browser sandbox.
| Evaluation Split | Train Rows | Test Rows | Split Duration?Wall-clock inference time taken for this evaluation split running 100% locally via WebGPU.Learn more → | Accuracy?Proportion of correct test predictions across all classes.Google ML Guide & Details → | ROC-AUC?Area under the ROC curve evaluating ranking capability across all thresholds.Google ML Guide & Details → | F1 Score?Harmonic mean of Precision and Recall, robust against class imbalance.Learn more → | Precision?Proportion of predicted positives that were actually positive.Google ML Guide & Details → | Recall?Proportion of actual positive cases successfully captured.Google ML Guide & Details → |
|---|---|---|---|---|---|---|---|---|
| R1 / F1 | 3,602 | 1,802 | 5.26s | 91.68% | 0.9705 | 0.8585 | 0.8569 | 0.8601 |
| R1 / F2 | 3,603 | 1,801 | 5.10s | 92.06% | 0.9709 | 0.8639 | 0.8681 | 0.8598 |
| R1 / F3 | 3,603 | 1,801 | 5.10s | 92.39% | 0.9764 | 0.8694 | 0.8769 | 0.8620 |
| R2 / F1 | 3,602 | 1,802 | 5.10s | 90.95% | 0.9729 | 0.8452 | 0.8492 | 0.8412 |
| R2 / F2 | 3,603 | 1,801 | 5.15s | 92.62% | 0.9766 | 0.8715 | 0.8895 | 0.8542 |
| R2 / F3 | 3,603 | 1,801 | 5.14s | 92.28% | 0.9732 | 0.8675 | 0.8750 | 0.8601 |
| R3 / F1 | 3,602 | 1,802 | 5.10s | 92.56% | 0.9726 | 0.8699 | 0.8942 | 0.8469 |
| R3 / F2 | 3,603 | 1,801 | 5.14s | 90.84% | 0.9665 | 0.8433 | 0.8457 | 0.8409 |
| R3 / F3 | 3,603 | 1,801 | 5.13s | 90.67% | 0.9696 | 0.8382 | 0.8546 | 0.8223 |
| Mean ± Std | — | — | 5.13s | 91.78% ± 0.73% | 0.972 ± 0.003 | 0.859 ± 0.012 | 0.868 | 0.850 |
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.
| Rank | Algorithm / Model | Model Family | Runtime | Predictive Accuracy?Proportion of correct test predictions across cross-validation splits.Google ML Guide & Details → | ROC-AUC?Area under the ROC curve evaluating ranking capability.Learn more → | F-Measure?Harmonic mean of precision and recall.Learn more → | Reference / Repo |
|---|---|---|---|---|---|---|---|
| #1 | Google TabFM v1.0 | Tabular Foundation Model | PyTorch/Python | 92.98% | 0.9798 | 0.8793 | google-research/tabfm ↗ |
| #2 | EXAONE Tabular | Tabular Foundation Model | PyTorch/Python | 91.96% | 0.9724 | 0.8617 | LGAI-Research/EXAONE-Tabular ↗ |
| #3 | TabPFN v3 | Tabular Foundation Model | PyTorch/Python | 91.95% | 0.9726 | 0.8609 | PriorLabs/tabpfn ↗ |
| #4 | Carla Engine (TabICLv2 WebGPU) | Tabular Foundation Model | Browser | 91.78% | 0.9721 | 0.8586 | 100% In-Browser |
| #5 | Streaming nanotabicl | Tabular Foundation Model | PyTorch/Python | 91.77% | 0.9722 | 0.8578 | soda-inria/nanotabicl ↗ |
| #6 | nanotabicl Vanilla | Tabular Foundation Model | PyTorch/Python | 91.77% | 0.9722 | 0.8583 | soda-inria/nanotabicl ↗ |
| #7 | KStar | Machine Learning Model | Python | 91.73% | 0.9674 | 0.9162 | OpenML #1672792 ↗ |
| #8 | Full TabICLv2 (PyTorch Reference) | Tabular Foundation Model | PyTorch/Python | 91.72% | 0.9722 | 0.8573 | soda-inria/tabicl ↗ |
| #9 | RandomForest | Random Forest | Python | 91.52% | 0.9643 | 0.9149 | OpenML #568789 ↗ |
| #10 | AttributeSelectedClassifier RandomForest | Random Forest | Python | 91.21% | 0.9622 | 0.9118 | OpenML #569670 ↗ |
| #11 | Bagging RandomForest | Random Forest | Python | 90.88% | 0.9623 | 0.9085 | OpenML #1672965 ↗ |
| #12 | RandomCommittee RandomTree | Decision Tree | Python | 90.71% | 0.9512 | 0.9062 | OpenML #1761214 ↗ |
| #13 | IB1 | Machine Learning Model | Python | 90.43% | 0.8761 | 0.9035 | OpenML #1665152 ↗ |
| #14 | AdaBoostM1 IBk | AdaBoost | Python | 90.43% | 0.8763 | 0.9035 | OpenML #1673270 ↗ |
| #15 | Bagging IBk | k-Nearest Neighbors (k-NN) | Python | 90.10% | 0.9281 | 0.9000 | OpenML #1674041 ↗ |
| #16 | AttributeSelectedClassifier IBk | k-Nearest Neighbors (k-NN) | Python | 90.06% | 0.8715 | 0.8997 | OpenML #569498 ↗ |
| #17 | Bagging RandomTree | Decision Tree | Python | 90.04% | 0.9480 | 0.8989 | OpenML #1672373 ↗ |
Variable Schema & Summary Distributions
Observed numerical ranges, category cardinalities, missing rates, and sample values across 5,404 rows.
| Col | Variable Name | Type | Missing | Distinct | Summary Stats / Distribution | Sample Values |
|---|---|---|---|---|---|---|
| A | phoneme_sound_classTARGET | NUM | 0 | 2 | [1, 2] μ=1.3 σ=0.5 | 111 |
| B | harmonic_1_normalized_amplitude | NUM | 0 | 5,336 | [-2.9337, 3.8286] μ=0 σ=1.0 | 0.489927-0.6412650.870593 |
| C | harmonic_2_normalized_amplitude | NUM | 0 | 5,312 | [-3.0387, 3.6648] μ=0 σ=1.0 | -0.4515280.109245-0.459862 |
| D | harmonic_3_normalized_amplitude | NUM | 0 | 5,308 | [-2.7959, 2.6309] μ=0 σ=1.0 | -1.047990.292130.578159 |
| E | harmonic_4_normalized_amplitude | NUM | 0 | 5,336 | [-2.4853, 3.0469] μ=0 σ=1.0 | -0.598693-0.9168040.806634 |
| F | harmonic_5_normalized_amplitude | NUM | 0 | 4,499 | [-2.3675, 4.5867] μ=0 σ=1.0 | -0.0204180.2402230.835248 |
Frequently Asked Questions: Phoneme
Common questions regarding the Phoneme dataset, machine learning task formulations, and in-browser tabular inference.
What is the Phoneme dataset used for?
The Phoneme dataset is an acoustic speech processing benchmark used to train and evaluate machine learning models in distinguishing nasal vowel sounds from oral vowel sounds based on five normalized harmonic spectral amplitudes.
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 Phoneme dataset?
Non-linear gradient boosting algorithms (such as LightGBM and XGBoost), Random Forests, and modern tabular foundation models like TabICL achieve the strongest benchmark scores, typically reaching 87% to 90% classification accuracy.
Test TabICLv2 on Phoneme Yourself
Open the pre-loaded Google Sheet and let Carla configure the target and task for local, zero-cloud tabular machine learning.
Launch Carla
Click once to open the spreadsheet and Carla side panel together.
Review the Setup
Carla selects the dataset target and task from this page automatically.
Evaluate & Predict
Run predictions and compute metrics with zero server uploads.
Provenance & Attribution
OpenML Dataset 1489