Continuous Trust Scoring for Three-Level Pre-Chain Admission of IoT Sensor Streams
Blockchain can make Internet of Things (IoT) records tamper-evident after submission, but sensor measurements still require quality assessment before ledger admission. We formulate this task as continuous estimation of a rule-derived data-quality trust proxy and propose PreTrust-TCA for multivariate sensor streams. Each sample contains 48 consecutive observations encoded by a two-block residual causal temporal convolutional network with additive temporal attention. The shared representation feeds a continuous score head and an auxiliary risk head; the score supports FULL 0.80
Authors
- Maotong Qin
- Luwei Wang (ORCID: https://orcid.org/0009-0006-0518-8972)
- Yunxiang Zhang
- Yue Chen (ORCID: https://orcid.org/0000-0002-0190-5253)
- Chen Li
Institutions
- Jiamusi University (CN)
- Shenzhen University (CN)
Publication Details
- Journal
- Applied Sciences
- Published
- 2026-09-11
- DOI
- https://doi.org/10.3390/app16189011
- Primary Topic
- Data Stream Mining Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Basic and Applied Basic Research Foundation of Guangdong Province