2026-10-13T11:47:00Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34029993 2026-09-30T05:23:34Z category_26941 item_type_12 month_year_09_2026

Inductive Discovery of Industrial Pollution Patterns Machine-Learning Sensor Data Fusion in Heterogeneous Environmental Systems Davina Liberata Dallo (25141658) Environmental management pattern discovery; unsupervised learning; machine olfaction; sensor drift; data fusion; industrial pollution monitoring. <h2 dir="ltr"><i>: Industrial pollution patterns are typically sought in supervised settings, where a model is trained to reproduce labels supplied by a reference method. This article argues that the most informative patterns in industrial environmental monitoring are precisely the ones for which no labels exist an uncharacterized emission source, an emerging nuisance regime, a novel interaction between process state and meteorological condition and that the field therefore requires an inductive programme alongside its supervised one. We formulate pattern discovery on heterogeneous IoT streams as a five-stage inductive procedure: representation learning from unlabelled sensor responses, cross-stream fusion on a common substrate, candidate pattern emergence, attribution to a physical or process cause, and stabilization under sensor drift. Reviewing the evidence, we show that each stage is individually demonstrated but that the stages have not been composed, and that two obstacles dominate the composition. The first is drift, which is directional rather than random and therefore destructively entangled with the very pattern structure the procedure seeks to recover; the second is the instability of feature selection under heterogeneous instrumentation, where the same physical variable yields an informative calibration model on one instrument and none on another. We identify four testable propositions and conclude that inductive discovery in this domain is limited less by algorithmic capability than by the absence of labelled and drift-annotated streams, which makes deliberate instrumentation of the discovery process the binding requirement.</i></h2><p></p> 2026-08-31T00:00:00Z Text Preprint 10.6084/m9.figshare.34029993.v1 https://figshare.com/articles/preprint/Inductive_Discovery_of_Industrial_Pollution_Patterns_Machine-Learning_Sensor_Data_Fusion_in_Heterogeneous_Environmental_Systems/34029993 CC BY 4.0