ts-classification

Time SeriesTime-Series-Library

Description

Time Series Classification: Custom Model Design

Objective

Design and implement a custom deep learning model for multivariate time series classification. Your code goes in the Model class in models/Custom.py. Three reference implementations (DLinear, TimesNet, PatchTST) are provided as read-only.

Evaluation

Trained and evaluated on three UEA datasets:

EthanolConcentration — spectral data classification
FaceDetection — MEG brain imaging classification
Handwriting — accelerometer-based character recognition

Training uses RAdam optimizer, CrossEntropyLoss, patience=10. Metric: accuracy (higher is better).

Code

Custom.py

EditableRead-only

1import torch
2import torch.nn as nn
3import torch.nn.functional as F
4
5
6class Model(nn.Module):
7    """
8    Custom model for time series classification.
9
10    Forward signature: forward(x_enc, x_mark_enc, x_dec, x_mark_dec, mask=None)
11    - x_enc: [batch, seq_len, enc_in] — input time series
12    - x_mark_enc: [batch, seq_len] — padding mask (1=valid, 0=padding)
13    - x_dec: not used (None)
14    - x_mark_dec: not used (None)
15

Additional context files (read-only):

Time-Series-Library/models/DLinear.py
Time-Series-Library/models/TimesNet.py
Time-Series-Library/models/PatchTST.py
Time-Series-Library/layers/AutoCorrelation.py
Time-Series-Library/layers/Autoformer_EncDec.py
Time-Series-Library/layers/Conv_Blocks.py
Time-Series-Library/layers/Crossformer_EncDec.py
Time-Series-Library/layers/Embed.py
Time-Series-Library/layers/FourierCorrelation.py
Time-Series-Library/layers/SelfAttention_Family.py
Time-Series-Library/layers/StandardNorm.py
Time-Series-Library/layers/Transformer_EncDec.py

Results

Show per-seed results

Model	Type	accuracy EthanolConcentration ↑	accuracy FaceDetection ↑	accuracy Handwriting ↑
dlinear	baseline	0.289	0.682	0.231
patchtst	baseline	0.285	0.685	0.254
timesnet	baseline	0.304	0.667	0.333
timesnet	baseline	0.319	0.675	0.335
timesnet	baseline	0.304	0.669	0.320
timesnet	baseline	0.304	0.669	0.346
anthropic/claude-opus-4.6	vanilla	0.313	0.631	0.519
deepseek-reasoner	vanilla	0.288	0.562	0.258
google/gemini-3.1-pro-preview	vanilla	0.267	0.649	0.267
openai/gpt-5.4-pro	vanilla	0.284	0.638	0.511
anthropic/claude-opus-4.6	agent	0.288	0.589	0.555
deepseek-reasoner	agent	0.283	0.597	0.469
google/gemini-3.1-pro-preview	agent	0.271	0.658	0.262
openai/gpt-5.4-pro	agent	0.286	0.651	0.462

Agent Conversations

anthropic/claude-opus-4.6

6 steps

deepseek-reasoner

6 steps

google/gemini-3.1-pro-preview

7 steps

openai/gpt-5.4-pro

6 steps