Meteorological Satellite Report Uncertainty and Negation Benchmark Dataset
- Records
- 3650
- Size
- 0.4GB
- Format
- JSON
- Provider
MOBIUSI INC- Updated
- v1.02026-11-04
This benchmark focuses on negation, modality, and uncertainty in web-based meteorological satellite reports, including statements such as no significant change observed, possible intensification, gradual weakening, and no expansion in extent yet. Each record provides structured annotations for whether a change is established, the confidence state, candidate change types, and supporting evidence spans from the source text. It supports meteorological remote-sensing text analysis, model training, and targeted evaluation of common errors in change understanding and structured generation.
Overview
Dataset Insights
Preview
Sample Examples
Schema
Technical Specifications
| Field | Type | Description |
|---|---|---|
| file_name | string | File name |
| report_text | string | The raw text of the web-based meteorological satellite report to be analyzed. |
| evidence_spans | array | Source-text spans supporting the judgment of whether a change is established, its confidence state, or its change type, with position and expression attributes. |
| confidence_state | string | Describes the certainty of the report's change judgment, distinguishing explicit conclusions, possible judgments, and cases where no judgment can be made. |
| change_established | boolean | Indicates whether the report explicitly states that the target change has occurred; negated statements such as no significant change observed are generally labeled false. |
| candidate_change_types | array | One or more change types identified in the report text; no significant change is also recorded as a candidate type. |
Trust
Compliance Statement
Authorization Type
Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use
Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization
No PII, no real company names, simulated scenarios follow industry standards
Compliance System
Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs
FAQ
Frequently Asked Questions
- What does the dataset contain?
- The dataset is based on web-based meteorological satellite reports. It provides structured labels for whether a change is established, the confidence state, candidate change types, and source-text evidence spans, with a focus on negation, possibility, and trend expressions.
- Which training and evaluation tasks does it support?
- It supports change establishment judgment, uncertainty detection, change type classification, evidence extraction, and structured information generation. It can also help diagnose errors involving statements such as no significant change observed or possible intensification.
- What are the dataset size and annotation specifications?
- The field design provides both judgment labels and source-text evidence, supporting result verification and error analysis. The supplied information does not specify the sample count or coverage period, so check the product details or dataset documentation for exact specifications before acquisition.
- How can I obtain the data and verify its license?
- Check the dataset product page for available access methods, file formats, and license terms. Before use, confirm that the permitted scope covers your research, training, or commercial deployment needs.
Can't find the data you need?
Post a request and let data providers reach out to you.
Get this Dataset
High-quality data, ready for your next AI breakthrough.
Cite this Work
@dataset{Mobiusi2026,
title={Meteorological Satellite Report Uncertainty and Negation Benchmark Dataset},
author={MOBIUSI INC},
year={2026},
url={https://www.mobiusi.com/datasets/74a8bcaf3423782600d3358999b1652d?dataset_scene_id=11},
urldate={2026-10-27},
keywords={meteorological satellite dataset,satellite report text,change detection benchmark,negation detection,uncertainty detection,remote sensing text,structured data generation,evidence span extraction},
version={1.0}
}Using this in research? Please cite us.