Human activity recognition leaderboard
OPPORTUNITY: macro-F1 (ZARA protocol, training-free LLM agents)
3 reported results for human activity recognition on OPPORTUNITY (ZARA protocol, training-free LLM agents), ranked by Macro-F1 on HAR activities. Every row cites its source publication.
Last updated
Recent changes
From the hub's dated sync history: new results as they are added, and SOTA stars as they are awarded or move.
Best reported result · macro-F1 · higher is better
92.5 (%)by ZARA (2025)
Open weightsNo result on this board has released weights on record.
3 results1 with code0 open weightsYears 2025–2026
Progress over time · macro-F1 · higher is better
Best so farNew bestOther resultProtocols can differ between points; check each row below.
Chart data (3 results)
| Year | Model | Open weights | macro-F1 |
|---|---|---|---|
| 2025 | ZARA (new best) | No or unknown | 92.5 (%) |
| 2026 | CHEAR | No or unknown | 85.13 (%) |
| 2026 | OntoAR-HAR | No or unknown | 91.5 ± 1.6 (%) |
| Rank | Model | macro-F1 ↑ | Year | Code | Open weights | Size | Protocol |
|---|---|---|---|---|---|---|---|
| 1 | ZARA zara-training-free-motion-reasoning-llm-agents-acl-2026 (ZARA, Gemini-2.0-Flash, training-free LLM agents) | 92.5 (%) | 2025 | Code | Closed | — | Subject-disjointLab Evaluation detailsZARA protocol: knowledge base from seen subjects, class-balanced held-out-subject test, 4 classes (Table 1). Label set: 4 classes Sensors in dataset: IMU Sampling rate: 30 Hz |
| 2 | OntoAR-HAR ontoar-har-ontology-guided-agentic-rag-har-2026 (OntoAR-HAR, Gemini 3.0 Flash) | 91.5 ± 1.6 (%) | 2026 | No code link recorded | Unknown | — | Subject-disjointLab Evaluation detailsMean of ten samples Label set: as ZARA table (4 classes) Sensors in dataset: IMU Sampling rate: 30 Hz |
| 3 | CHEAR chear-confusion-hierarchy-evidence-aware-llm-har-2026 (CHEAR, Gemini 3 Flash) | 85.13 (%) | 2026 | No code link recorded | Unknown | — | Subject-disjointLab Evaluation detailsZARA 92.5 in the same table Label set: as ZARA table (4 classes) Sensors in dataset: IMU Sampling rate: 30 Hz |
Rows are ranked only within this task × dataset × metric and one protocol group. Protocols can still differ in detail; read each method's evaluation details before comparing. Year, code, open weights and size come from each model's or paper's record; a dash or “Unknown” means the record does not say.
This page ranks every result ChatPPG has recorded for human activity recognition on OPPORTUNITY, measured by Macro-F1 on HAR activities (higher is better). Each number is copied from the publication cited in its row, and ranks apply only within this dataset and metric and this protocol group.
Evaluation protocol: ZARA-protocol family: training-free LLM agents; knowledge base from seen subjects, class-balanced held-out-subject test.
About the dataset
Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
Sources on this board
ZARA: Training-Free Motion Time-Series Reasoning via Evidence-Grounded LLM AgentsarXiv preprint · 2025
OntoAR-HAR: Ontology-Guided Agentic RAG with Personalized Memory for Human Activity RecognitionarXiv 2610.15292 · 2026
CHEAR: Confusion-Hierarchy Evidence-Aware Reasoning for Cost-Efficient Human Activity RecognitionarXiv 2610.15298 · 2026
Related leaderboards
Other metrics on OPPORTUNITY
Human activity recognition on other datasets

Human activity recognition
MHEALTH: F1
4 results

Human activity recognition
PAMAP2: F1
4 results

Human activity recognition
PAMAP2: macro-F1 (LOSO, wrist IMU)
4 results

Human activity recognition
WISDM AR v1.1: accuracy
4 results

Human activity recognition
CAPTURE-24: Cohen's kappa (Subject-wise 5-fold CV)
3 results

Human activity recognition
CAPTURE-24: macro-F1 (Subject-wise 5-fold CV)
3 results

Human activity recognition
MHEALTH: macro-F1 (ZARA protocol, training-free LLM agents)
3 results

Human activity recognition
PAMAP2: macro-F1 (ZARA protocol, training-free LLM agents)
3 results

Human activity recognition
RealWorld HAR: macro-F1 (Subject-wise 5-fold CV)
3 results

Human activity recognition
UCI Human Activity Recognition Using Smartphones (UCI-HAR): F1
3 results

Human activity recognition
USC-HAD: F1
3 results

Human activity recognition
CAPTURE-24: Cohen's kappa (4-class activity intensity)
1 result
Show 14 more leaderboards

Human activity recognition
CAPTURE-24: macro-F1 (4-class activity intensity)
1 result

Human activity recognition
CAPTURE-24: macro-F1 (3-class activity intensity)
1 result

Human activity recognition
DEO: balanced accuracy
1 result

Human activity recognition
hapt: macro-F1
1 result

Human activity recognition
har-6bench-avg: macro-F1
1 result

Human activity recognition
HHAR: macro-F1
1 result

Human activity recognition
MHEALTH: accuracy
1 result

Human activity recognition
MHEALTH: macro-F1 (IMU-Bench protocol)
1 result

Human activity recognition
opentslm-har-cot: macro-F1
1 result

Human activity recognition
PAMAP2: accuracy
1 result

Human activity recognition
prosi: macro-F1
1 result

Human activity recognition
RealWorld HAR: macro-F1 (4-class activity intensity)
1 result

Human activity recognition
μDPad: accuracy
1 result

Human activity recognition
WISDM AR v1.1: macro-F1
1 result
Data and submissions
Downloads of every leaderboard (JSON and CSV) are free for members. Sign in to download.Missing a published result? Email hello@chatppg.com with the paper link and the exact table it comes from.
Cite or reuse this
CC BY 4.0ChatPPG compiled this data. You can copy, adapt and republish it, including in articles, slides, papers and AI tools, as long as you credit ChatPPG and link to this page. Please also cite the original studies for any number you use.
Attribution
Source: ChatPPG, OPPORTUNITY macro-F1 leaderboard (ZARA protocol, training-free LLM agents): human activity recognition, https://chatppg.com/research/leaderboards/har/opportunity-macro-f1-har-zara-protocol (CC BY 4.0)
HTML link
Source: <a href="https://chatppg.com/research/leaderboards/har/opportunity-macro-f1-har-zara-protocol">ChatPPG, OPPORTUNITY macro-F1 leaderboard (ZARA protocol, training-free LLM agents): human activity recognition</a> (<a href="https://creativecommons.org/licenses/by/4.0/" rel="license">CC BY 4.0</a>)Cite the page URL and the date you accessed it. Please also cite the original publications listed above.
APA
ChatPPG. (2026). OPPORTUNITY macro-F1 leaderboard (ZARA protocol, training-free LLM agents): human activity recognition [Data set]. ChatPPG Research Hub. https://chatppg.com/research/leaderboards/har/opportunity-macro-f1-har-zara-protocol
BibTeX
@misc{chatppg_har_opportunity_macro_f1_har_zara_protocol,
author = {{ChatPPG}},
title = {{OPPORTUNITY macro-F1 leaderboard (ZARA protocol, training-free LLM agents): human activity recognition}},
year = {2026},
month = oct,
howpublished = {\url{https://chatppg.com/research/leaderboards/har/opportunity-macro-f1-har-zara-protocol}},
note = {ChatPPG Research Hub}
}CITATION.cff
cff-version: 1.2.0 message: "If you use this resource, please cite it as below." type: dataset title: "OPPORTUNITY macro-F1 leaderboard (ZARA protocol, training-free LLM agents): human activity recognition" authors: - name: "ChatPPG" url: "https://chatppg.com/research/leaderboards/har/opportunity-macro-f1-har-zara-protocol" date-released: "2026-10-19"