Abstract
<p>The Dansgaard-Oeschger (DO) events are one of the most striking examples of abrupt climate change in the Earth's history, representing temperature oscillations of about 8 to 16 degrees Celsius within a few decades. DO events have been studied extensively in paleoclimatic records, particularly in ice core proxies. Examples include the Greenland NGRIP record of oxygen isotopic composition.<br>This work addresses the anticipation of DO events using machine learning algorithms. We consider the NGRIP time series from 20 to 60 kyr b2k with the GICC05 timescale and 20-year temporal resolution. Forecasting horizons range from 0 (nowcasting) to 400 years. We adopt three different machine learning algorithms (random forests, support vector machines, and logistic regression) in training windows of 5 kyr. We perform validation on subsequent test windows of 5 kyr, based on timestamps of previous DO events' classification in Greenland by Rasmussen et al. (2014). We perform experiments with both sliding and growing windows.<br>Results show that predictions on sliding windows are better overall, indicating that modelling is affected by non-stationary characteristics of the time series. The three algorithms' predictive performance is similar, with a slightly better performance of random forest models for shorter forecast horizons. The prediction models' predictive capability decreases as the forecasting horizon grows more extensive but remains reasonable up to 120 years. Model performance deprecation is mostly related to imprecision in accurately determining the start and end time of events and identifying some periods as DO events when such is not valid.</p>

Abstract

Abstract

Abstract
Starting from a gap identified in the literature regarding the use of social networks by newspapers to disseminate urgent news, this article aims to study strategies of journalistic content in social media, particularly in the context of a public crisis and to compare the effectiveness of different types of news disseminated in this medium, namely in terms of reach and generated interaction. The following research question was defined: how popular was public health news in Brazil during the covid-19 pandemic? Based on contributions in the literature, a quantitative study was carried out, using the content analysis technique. The study enable to better understand the sharing behavior of news in Twitter, the consumption behavior of newspaper readers on social networks and the generation of news during the pandemic.

Abstract
Predawn leaf water potential (Psi(pd)) is widely used to assess plant water status. Also, pigments concentration work as proxy of canopy's water status. Spectral data methods have been applied to monitor and assess crop's biophysical variables. This work developed two models to estimate Psi(pd) using a hand-held spectroradiometer (400-1010 nm) to obtain canopy and foliar reflectance in four dates of 2018 and a pressure chamber to measure Psi(pd). Two modelling approaches, combining spectral data and several machine learning algorithms (MLA), were used to estimate Psi(pd) in a commercial vineyard in the Douro Wine Region. The first approach estimated Psi(pd) through vine's canopy reflectance; several vegetation indices (VIs) were computed and selected, namely the SPVi(opt)(1_)(950;596;521;) SPVIopt2_896;880;901; PRI_CI2(opt_539;560,573;716 )and NPCIopt_983;972, as well as a time-dynamic variable based on Psi(pd) (Psi(pd)_(0)). The second modelling approach is based on pigments' concentrations; several VIs were optimized for non-correlated pigments of vine's leaves, assessed by its hyperspectral reflectance. The following variables for Psi(pd) estimation were selected through stepwise forward method: Psi(pd)_(0); NRIgreen_LUT520;532; NRIgreen_LWC540;551. The B-MARS algorithm performed the best results for both modelling approaches, presenting a RRMSE in both validation modelling approaches between 13-14%.

Abstract
In the past decade, we have witnessed the widespread adoption of Deep Neural Networks (DNNs) in several Machine Learning tasks. However, in many critical domains, such as healthcare, finance, or law enforcement, transparency is crucial. In particular, the lack of ability to conform with prior knowledge greatly affects the trustworthiness of predictive models. This paper contributes to the trustworthiness of DNNs by promoting monotonicity. We develop a multi-layer learning architecture that handles a subset of features in a dataset that, according to prior knowledge, have a monotonic relation with the response variable. We use two alternative approaches: (i) imposing constraints on the model's parameters, and (ii) applying an additional component to the loss function that penalises non-monotonic gradients. Our method is evaluated on classification and regression tasks using two datasets. Our model is able to conform to known monotonic relations, improving trustworthiness in decision making, while simultaneously maintaining small and controllable degradation in predictive ability.