Closed-Loop Data Reduction for Edge Time-Series: annotated corpus, search records and reproduction scripts

Published: 4 November 2026| Version 1 | DOI: 10.17632/tbh4253ysg.1
Contributors:
Iago Martínez,
,
,
,
,

Description

Supplementary data and methods for the article “Closed-Loop Data Reduction for Edge Time-Series: A Design-Oriented Survey” (Internet of Things, Elsevier). The package contains the 128-record annotated corpus (107 primary, 21 contextual records), lexical reporting-audit patterns, evidence logs and source manifest, Scopus and IEEE Xplore search records (bibliographic metadata only), reconciliation and inclusion records, Python reproduction scripts, and a README. Source articles are not included; text-audit reruns require independently obtained source PDFs.

Files

Steps to reproduce

Unzip the package and read README.md. The database reconciliation is reproduced with python reproduce_crosscheck.py (Python 3, standard library only), which recomputes the Scopus/IEEE Xplore counts and union from the archived exports and checks them against crosscheck/reconciliation.json. The lexical reporting audit requires independently obtained source PDFs: run python prepare_audit_sources.py SOURCE_DIRECTORY --output CANONICAL_DIRECTORY to match them by SHA-256 against source_manifest.csv, then python reproduce_text_audit.py CANONICAL_DIRECTORY --output AUDIT_DIRECTORY (Python 3.12, PyMuPDF 1.28.2). Database exports contain bibliographic metadata only; abstracts and keywords were removed to comply with provider redistribution terms.

Institutions

Categories

Computer Science, Wireless Sensor Network, Data Compression, Time Series Analysis, Internet of Things, Edge Computing

Licence