OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Ucr Time Series Classification Deep Learning Baseline

A public Python repository containing MLP, fully convolutional network and ResNet baselines for end-to-end classification of univariate time series, with experiments on UCR benchmarks and methods for visualizing class-specific regions and network weights.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Ucr Time Series Classification Deep Learning Baseline — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it to study or compare deep-learning baselines on univariate time-series classification, particularly when UCR benchmark results or visual interpretation methods are relevant.

Who it is for

  • Researchers comparing neural-network approaches on time-series classification benchmarks.
  • Practitioners seeking a starting baseline for classifying univariate signals in areas such as finance, industry or healthcare.
  • Readers interested in visualizing which time-series regions contribute to a model’s class scores.

Documented uses

  • Compare MLP, FCN and ResNet classification error rates across UCR datasets.
  • Use an end-to-end deep-learning baseline as a starting point for time-series analytics.
  • Inspect class-specific time-series regions with class activation maps.
  • Visualize network filters or weights using Gramian Angular Summation Fields.
Before you choose

Requirements & limits

  • The repository includes three architectures: MLP, FCN and ResNet.
  • Its README reports testing error rates on 85 UCR datasets and compares them with several conventional methods.
  • CAM is described for FCN and ResNet models with global average pooling; the README also describes weight visualization with Gramian Angular Summation Fields.
  • The README warns that UCR data is unusually ideal relative to settings with skewed labels, non-stationarity and distribution or concept drift.
  • The README says no single approach is clearly best in its experiments; it suggests BOSS for readers preferring a white-box model and FCN or MLP with dropout as deep-learning baselines.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Choose an end-to-end baseline for time-series classification

Select an MLP, FCN, or ResNet as a deep-learning baseline for univariate time-series classification.

Before you start

  • Univariate time-series data.
  • A preference for an end-to-end deep-learning approach rather than a white-box model.

Steps

  1. Choose FCN or MLP with dropout as a first baseline; the README also lists ResNet among the repository's models.

  2. If using FCN, consider CAM or grad-CAM for a degree of model interpretability.

Expected outcome

A documented candidate baseline is selected for end-to-end time-series classification.

Guide limitations

  • The README does not provide installation, data preparation, training, or inference steps.
  • The README reports no significant difference between the mentioned deep-learning approaches and COTE or BOSS in its experiments; it does not establish one approach as universally best.
  • The dataset context is limited: the README notes that the UCR time series is kind, but the provided text ends before completing that discussion.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
7 November 2026
Original website title
cauchyturing/UCR_Time_Series_Classification_Deep_Learning_Baseline · Official homepage title · retained in its source language
How we review sourcesSuggest a correction