List your company
ResearchFormerly WizWand

Time-series Classification on Epilepsy (test)

97.23AccuracyTS-TCC
65.790873.952982.11590.2771Jun 26, 2021Apr 15, 2022Feb 2, 2023Nov 23, 2023Sep 11, 2024Jul 1, 2025Apr 21, 2026
Updated 6mo ago

Evaluation results

MethodLinks
TS-TCCEvaluation Protocol=linear evaluation, Encoder State=frozen pretrained2021.0697.2395.54—
Baseline stagedm (number of modalities)=1, dm (number of variates)=3, Tm (sequence length)=206, C (number of classes)=4, Train size=137, Test size=1382026.0397.1——
SupervisedEvaluation Protocol=end-to-end supervised training, Encoder State=jointly trained2021.0696.6694.52—
CPCEvaluation Protocol=linear evaluation, Encoder State=frozen pretrained2021.0696.6194.44—
Baseline end-to-endm (number of modalities)=1, dm (number of variates)=3, Tm (sequence length)=206, C (number of classes)=4, Train size=137, Test size=1382026.0396.4——
MF-ESN2026.0496.1——
SimCLREvaluation Protocol=linear evaluation, Encoder State=frozen pretrained2021.0696.0593.53—
PULSEPre-training Dataset=EEG, Fine-tuning Head=2-layer MLP, Fine-tuning Epochs=40, Optimizer=Adam, Learning Rate=0.0003, Aggregation=Average of five random seeds2025.1195.82—98.51
MARS2026.0495.6——
REBARPre-training Dataset=EEG, Fine-tuning Head=2-layer MLP, Fine-tuning Epochs=40, Optimizer=Adam, Learning Rate=0.0003, Aggregation=Average of five random seeds2025.1195.27—98.33
PatchTSTPre-training Dataset=EEG, Fine-tuning Head=2-layer MLP, Fine-tuning Epochs=40, Optimizer=Adam, Learning Rate=0.0003, Aggregation=Average of five random seeds2025.1195.03—98.04
DSVAEPre-training Dataset=EEG, Fine-tuning Head=2-layer MLP, Fine-tuning Epochs=40, Optimizer=Adam, Learning Rate=0.0003, Aggregation=Average of five random seeds2025.1194.97—98.17
LFADSPre-training Dataset=EEG, Fine-tuning Head=2-layer MLP, Fine-tuning Epochs=40, Optimizer=Adam, Learning Rate=0.0003, Aggregation=Average of five random seeds2025.1194.71—98.01
TS2VecPre-training Dataset=EEG, Fine-tuning Head=2-layer MLP, Fine-tuning Epochs=40, Optimizer=Adam, Learning Rate=0.0003, Aggregation=Average of five random seeds2025.1193.95—95.87
SSL-ECGEvaluation Protocol=linear evaluation, Encoder State=frozen pretrained2021.0693.7289.15—
SimCLRPre-training Dataset=EEG, Fine-tuning Head=2-layer MLP, Fine-tuning Epochs=40, Optimizer=Adam, Learning Rate=0.0003, Aggregation=Average of five random seeds2025.1193.52—97.52
Random InitializationEvaluation Protocol=linear evaluation, Encoder State=randomly initialized2021.0690.2681.12—
ESN2026.0486.8——
Previous workm (number of modalities)=1, dm (number of variates)=3, Tm (sequence length)=206, C (number of classes)=4, Train size=137, Test size=1382026.0367——