Abstract
<title>Abstract</title> <p>The reproducibility of electrospinning is strongly influenced by the ambient environment, as temperature and relative humidity affect solvent evaporation, jet stability, and fiber formation. In practice, these variables are often documented only retrospectively, which limits the ability of researchers to anticipate weather-driven shifts in laboratory conditions and in turn can hinder reproducibility. Here, we present a machine-learning-based decision-support framework for forecasting indoor laboratory relative humidity from local meteorological measurements and forecasts and internal climate measurements. To preserve room-specific influences such as orientation, ventilation context, and local thermal buffering, we trained and evaluated separate regression models for four laboratories within the same building rather than assuming a single shared indoor-outdoor transfer function. We then used site-specific weather forecasts to assess predictive performance and uncertainty across forecast horizons from 1 to 7 days. To examine user-related indoor variability without relying on direct occupancy sensing, we additionally analyzed working- and non-working-day patterns using a calendar-based occupancy proxy and Wasserstein-distance heatmaps. The resulting framework provides a seven-day indoor humidity forecast with lead-day-specific prediction intervals intended to support experiment planning under weather-responsive indoor conditions. Because the study was conducted in four laboratories within a single building, the trained models should be understood as site-specific rather than directly transferable to other laboratory environments. The broader contribution of this work therefore lies in the forecasting and decision-support workflow, which can be adapted to other settings using local data. At its current stage, the framework is intended to indicate likely indoor relative humidity shifts and their uncertainty; the experimental significance of those shifts remains to be interpreted by the researcher in the context of the specific polymer system and target fiber outcome.</p>