.png){kind=icon}
.png){kind=icon}
Return
Abductive hypothesis testing in cognitive IoT sensor network
.png){kind=icon}
PRE
AI
DOI：10.1007/s10586-025-05128-z.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Recently, IoT research has focused on objects that can learn, reason, and perceive their environment. This led to cognitive IoT (CIoT), which adds brain-like intelligence to IoT devices. Many CIoT inferential operations involve testing hypotheses that become more complicated with diverse and huge datasets. Therefore, this work presents abductive hypothesis testing (AHT) that employs a total variation regularizer for noisy data and probabilistic clustering to tolerate missing entries. Further, the plausibility value is determined for each cluster to determine its informative value for binary dataset conversion for generating the concept lattice and significant data. In this way, the large datasets are compressed into smaller ones. We then use the minimum akaike information criterion to build a copula for data heterogeneity and adjust p-values before abductive hypothesis testing. The study then assesses the algorithm's effectiveness using 21.25 years of weather data and concludes that the suggested technique excels over competing approaches.
Keywords:
Abductive
CIoT
Massive and heterogeneous data
Concept lattice
Hypothesis testing
Journal
C
IF:
0
Papers:
691
Citations:
1