Explore the vast range of titles available.

Most machine learning methods assume that the input data distribution is the same in the training and testing phases. However, in practice, this stationarity is usually not met and the distribution of inputs differs, leading to unexpected performance of the learned model in deployment. The issue in which the training and test data inputs follow different probability distributions while the input–output relationship remains unchanged is referred to as covariate shift. In this paper, the performance of conventional machine learning models was experimentally evaluated in the presence of covariate shift. Furthermore, a region-based evaluation was performed by decomposing the domain of probability density function of the input data to assess the classifier’s performance per domain region. Distributional changes were simulated in a two-dimensional classification problem. Subsequently, a higher four-dimensional experiments were conducted. Based on the experimental analysis, the Random Forests algorithm is the most robust classifier in the two-dimensional case, showing the lowest degradation rate for accuracy and F1-score metrics, with a range between 0.1% and 2.08%. Moreover, the results reveal that in higher-dimensional experiments, the performance of the models is predominantly influenced by the complexity of the classification function, leading to degradation rates exceeding 25% in most cases. It is also concluded that the models exhibit high bias toward the region with high density in the input space domain of the training samples.

Building energy simulation (BES) models rely on a variety of different input data, and the more accurate the input data are, the more accurate the model will be in predicting energy use. The objective of this paper is to show a method for obtaining higher accuracy in building energy simulations of existing buildings by combining time diaries with data from logged measurements, and also to show that more variety is needed in template values of user input data in different kinds of buildings. The case studied in this article is a retirement home in Linköping, Sweden. Results from time diaries and interviews were combined with logged measurements of electricity, temperature, and CO2 levels to create detailed occupant behavior schedules for use in BES models. Two BES models were compared, one with highly detailed schedules of occupancy, electricity use, and airing, and one using standardized input data of occupant behavior. The largest differences between the models could be seen in energy losses due to airing and in household electricity use, where the one with standardized user input data had a higher amount of electricity use and less losses due to airing of 39% and 99%, respectively. Time diaries and interviews, together with logged measurements, can be great tools to detect behavior that affects energy use in buildings. They can also be used to create detailed schedules and behavioral models, and to help develop standardized user input data for more types of buildings. This will help improve the accuracy of BES models so the energy efficiency gap can be reduced.

In this study, a village input–output table (VIOT) is built from household survey data from a rural village in a developing country to capture the interdependency between households through their transactions in 2016. This VIOT is a simple, but useful tool for understanding the economic transactions among villagers. The main findings of this work are that lower- and middle-income households mainly depend on commodities supplied by non-poor counterparts, especially by four higher-income households, who are not only producers, but also traders of commodities in the village and play key roles in the village economy, and that the IO interdependency among non-poor households is stronger than that among poor households. Additionally, this paper describes a first trial application of the VIOT method to develop economic policies with goals such as poverty reduction and trade expansion in the village.

While quality of life (QOL) is the result of satisfying human needs, our current provision strategies result in global environmental degradation. To ensure sustainable QOL, we need to understand the environmental impact of human needs satisfaction. In this paper we deconstruct QOL, and apply the fundamental human needs framework developed by Max-Neef et al to calculate the carbon and energy footprints of subsistence, protection, creation, freedom, leisure, identity, understanding and participation. We find that half of global carbon emissions are driven by subsistence and protection. A similar amount are due to freedom, identity, creation and leisure together, whereas understanding and participation jointly account for less than 4% of global emissions. We use 35 objective and subjective indicators to evaluate human needs satisfaction and their associated carbon footprints across nations. We find that the relationship between QOL and environmental impact is more complex than previously identified through aggregated or single indicators. Satisfying needs such as protection, identity and leisure is generally not correlated with their corresponding footprints. In contrast, the likelihood of satisfying needs for understanding, creation, participation and freedom, increases steeply when moving from low to moderate emissions, and then stagnates. Most objective indicators show a threshold trend with respect to footprints, but most subjective indicators show no relationship, except for freedom and creation. Our study signals the importance of considering both subjective and objective satisfaction to assess QOL-impact relationships at the needs level. In this way, resources could be strategically invested where they strongly relate to social outcomes, and spared where non-consumption satisfiers could be more effective. Through this approach, decoupling human needs satisfaction from environmental damage becomes more attainable.

Due to relatively high costs and labor required for experimental profiling of the full target space of chemical compounds, various machine learning models have been proposed as cost-effective means to advance this process in terms of predicting the most potent compound-target interactions for subsequent verification. However, most of the model predictions lack direct experimental validation in the laboratory, making their practical benefits for drug discovery or repurposing applications largely unknown. Here, we therefore introduce and carefully test a systematic computational-experimental framework for the prediction and pre-clinical verification of drug-target interactions using a well-established kernel-based regression algorithm as the prediction model. To evaluate its performance, we first predicted unmeasured binding affinities in a large-scale kinase inhibitor profiling study, and then experimentally tested 100 compound-kinase pairs. The relatively high correlation of 0.77 (p < 0.0001) between the predicted and measured bioactivities supports the potential of the model for filling the experimental gaps in existing compound-target interaction maps. Further, we subjected the model to a more challenging task of predicting target interactions for such a new candidate drug compound that lacks prior binding profile information. As a specific case study, we used tivozanib, an investigational VEGF receptor inhibitor with currently unknown off-target profile. Among 7 kinases with high predicted affinity, we experimentally validated 4 new off-targets of tivozanib, namely the Src-family kinases FRK and FYN A, the non-receptor tyrosine kinase ABL1, and the serine/threonine kinase SLK. Our sub-sequent experimental validation protocol effectively avoids any possible information leakage between the training and validation data, and therefore enables rigorous model validation for practical applications. These results demonstrate that the kernel-based modeling approach offers practical benefits for probing novel insights into the mode of action of investigational compounds, and for the identification of new target selectivities for drug repurposing applications.

Disasters are often followed by a large-scale stimulus supporting the economy through the built environment, which can last years. During this time, official economic indicators tend to suggest the economy is doing well, but as activity winds down, the sentiment can quickly change. In response to the damaging 2011 earthquakes in Canterbury, New Zealand, the regional economy outpaced national economic growth rates for several years during the rebuild. The repair work on the built environment created years of elevated building activity. However, after the peak of the rebuilding activity, as economic and employment growth retracts below national growth, we are left with the question of how the underlying economy performs during large scale stimulus activity in the built environment. This paper assesses the performance of the underlying economy by quantifying the usual, demand-driven level of building activity at this time. Applying an Input-Output approach and excluding the economic benefit gained from the investment stimulus reveals the performance of the underlying economy. The results reveal a strong growing underlying economy, and while convergence was expected as the stimulus slowed down, the results found that growth had already crossed over for some time. The results reveal that the investment stimulus provides an initial 1.5% to 2% growth buffer from the underlying economy before the growth rates cross over. This supports short-term economic recovery and enables the underlying economy to transition away from a significant rebuild stimulus. Once the growth crosses over, five years after the disaster, economic growth in the underlying economy remains buoyant even if official regional economic data suggest otherwise.

Reducing food loss and waste (FLW) is a global priority under UN SDG 12.3, yet Taiwan has lacked stage-specific FLW data and systematic valuation of its environmental and economic implications. This study addresses these gaps by integrating localized FLW estimates from the APEC-FLOWS database with an enhanced analytical framework—the Environmentally Extended Input–Output Valuation (EEIO-V) model. The EEIO-V extends conventional input–output analysis by monetizing multiple environmental burdens, including greenhouse gases, air pollutants, wastewater, and solid waste, thereby linking FLW reduction to tangible economic benefits and policy design. The simulations reveal substantial differences in environmental cost reductions across supply chain stages, with downstream interventions delivering the largest benefits, particularly in reducing air pollution and greenhouse gases. By contrast, upstream measures contribute relatively smaller improvements. These findings highlight the novelty of EEIO-V in bridging environmental valuation with system-level FLW analysis, and they provide actionable insights for designing cost-effective, stage-specific strategies that prioritize downstream interventions to advance Taiwan’s sustainability and policy goals.