Artigos Acadêmicos e Noticiosos
URI permanente desta comunidadehttps://repositorio.insper.edu.br/handle/11224/3226
Navegar
7 resultados
Resultados da Pesquisa
Artigo Científico Neonatal mortality prediction with routinely collected data: a machine learning approach(2021) ANDRE FILIPE DE MORAES BATISTA; Diniz, Carmen S. G.; Bonilha, Eliana A.; Kawachi, Ichiro; Chiavegatto Filho, Alexandre D. P.Background: Recent decreases in neonatal mortality have been slower than expected for most countries. This study aims to predict the risk of neonatal mortality using only data routinely available from birth records in the largest city of the Americas. Methods: A probabilistic linkage of every birth record occurring in the municipality of São Paulo, Brazil, between 2012 e 2017 was performed with the death records from 2012 to 2018 (1,202,843 births and 447,687 deaths), and a total of 7282 neonatal deaths were identified (a neonatal mortality rate of 6.46 per 1000 live births). Births from 2012 and 2016 (N = 941,308; or 83.44% of the total) were used to train five different machine learning algorithms, while births occurring in 2017 (N = 186,854; or 16.56% of the total) were used to test their predictive performance on new unseen data. Results: The best performance was obtained by the extreme gradient boosting trees (XGBoost) algorithm, with a very high AUC of 0.97 and F1-score of 0.55. The 5% births with the highest predicted risk of neonatal death included more than 90% of the actual neonatal deaths. On the other hand, there were no deaths among the 5% births with the lowest predicted risk. There were no significant differences in predictive performance for vulnerable subgroups. The use of a smaller number of variables (WHO’s five minimum perinatal indicators) decreased overall performance but the results still remained high (AUC of 0.91). With the addition of only three more variables, we achieved the same predictive performance (AUC of 0.97) as using all the 23 variables originally available from the Brazilian birth records. Conclusion: Machine learning algorithms were able to identify with very high predictive performance the neonatal mortality risk of newborns using only routinely collected data.Artigo Científico Wildfire-sourced fine particulate matter and preterm birth risks in Brazil: A nationwide population-based cohort study(2024) Zhang, Yiwen; Huang, Wenzhong; Xu, Rongbin; Ye, Tingting; Chen, Gongbo; Yue, Xu; Coêl , Micheline de Sousa Zanotti Stagliorio; Saldiva, Paulo Hilario Nascimento; Song, Jiangning; Guo, Yuming; Li, ShanshanWildfire-specific particulate matter with diameters ≤ 2.5 µm (PM2.5) is the key component of wildfire smoke, with potentially higher toxicity than PM2.5 from other sources. In this nationwide population-based cohort study, we included 22,163,195 births from Brazil during 2010–2019. Daily wildfire-specific PM2.5 was estimated through the chemical transport model. Time-varying Cox proportional hazards models were used to characterize the exposure-time-response (E-T-R) relationship between weekly wildfire-specific PM2.5 exposure and preterm birth (PTB) risks, followed by subgroup analyses. A 10 µg/m3 increment in wildfire-specific PM2.5 was associated with a hazard ratio of 1.047 (95 % confidence interval [CI]: 1.032–1.063) for PTB. Stronger associations between wildfire-specific PM2.5 and PTB were observed during earlier pregnancy, among female infants, and pregnant women < 18 years old, in ethnic minorities, with a length of education ≥ 11 years, from low-income or high temperature municipalities, and residing in North/Northeast regions. An estimated 1.47 % (95 % CI: 1.01 %–1.94 %) of PTBs were attributable to wildfire-specific PM2.5 in Brazil, increasing from 2010 to 2019. The PTBs attributable to wildfire-specific PM2.5 surpassed those attributed to non-wildfire PM2.5 (0.31 %, 95% CI: 0.09 %–0.57 %). Wildfire emerged as a critical source contributing to the PM2.5-linked PTBs. Prioritized fire management and emission control strategies are warranted for PTB prevention.Artigo Científico Digital Platform Ecosystems and Conglomerate Mergers: A Review of the Brazilian Experience(2022) Zingales, Nicolo; Renzetti, Bruno PolonioThis paper highlights some of the key challenges for the Brazilian merger control regime in dealing with mergers involving digital platform ecosystems (DPEs). After a quick introduction to DPEs, we illustrate how conglomerate effects that are raised by such mergers remain largely unaddressed in the current landscape for merger control in Brazil. The paper is divided in four sections. First, we introduce the reader to the framework for merger control in Brazil. Second, we identify the possible theories of harm related to conglomerate mergers, and elaborate on the way in which their application may be affected by the context of DPEs. Third, we conduct a review of previous mergers involving DPEs in Brazil, aiming to identify the theories of harm employed (and those that could have been explored) in each case. Fourth and finally, we summarize and results and suggest adaptations to the current regime, advancing proposals for a more consistent and predictable analysis.Artigo Científico The role of police contact and neighborhood experiences on legal socialization: Longitudinal evidence from adolescents in Brazil(2021) Piccirillo, Debora; García-Sánchez, Efraín; MARIANA CHIES SANTIAGO SANTOS; Gomes, Aline MizutaniThe legal socialization framework emphasizes the impor tance of adolescents’ encounters with police. We exam ine how different types of police contact and neighbor hood experiences are linked to the legal socialization pro cess among adolescents living in São Paulo, Brazil. Draw ing on 669 participants across three waves of panel data from the São Paulo Legal Socialization Study, results from the multilevel longitudinal model revealed that within person increases in vicarious police contact were linked to a decrease in police legitimacy over time. Exploratory analyses using a multilevel mediation model indicated that the observed negative effect of vicarious police contact was mediated by evaluations of police procedural justice. As for the between-person effects, voluntary police contact posi tively predicted police legitimacy over time. However, the effects of vicarious police contact and violent police contact were mediated by police procedural justice. Additionally, police legitimacy decreased for people who had more expo sure to violence and lower levels of fear of crime. The study revealed that the nature of police contact and levels of expo sure to violence can have important effects on adolescents’ perceptions of police legitimacy, even after accounting for procedural justice.Artigo Científico Networking and developing collaborative relationships: evidence of the auto-part industry of Brazil(2011) DANNY PIMENTEL CLARO; PRISCILA BORIN DE OLIVEIRA CLAROPurpose – This study aims to assess the moderating effect of the business network on the effects of between relational behavior and the effects of transaction-specific investments on joint actions. Design/methodology/approach – The study was a survey-based field study designed using theoretical support from marketing channels, transaction cost economics and network perspectives. Findings – The results show the importance of relational behavior and the network in coordinating joint actions, and this has relevant managerial implications for the coordination of a collaborative relationship. The characteristics of the relationship, its length as well as the size of each partner affect the collaborative efforts of the partners. Practical implications – Firms and managers should understand not only the dyadic relationships they are in but also the network structure. Dyadic characteristics affect collaboration, while the network also has effects on the collaboration of partners in vulnerable positions. Originality/value – The paper points out the role of the network as a countervailing safeguard for dyadic TSIs and network stability. Dyadic relationships are supported by the network.Artigo Científico Investigating the partial adjustment effect of Brazilian IPOs(2015) ANDREA MARIA ACCIOLY FONSECA MINARDI; Moita, Rodrigo Menon; Castanho, Rafael PlantierGlobal literature reports positive initial return in IPOs, or “money left on the table” by the issuing companies. One possible cause is that when the underwriter perceives high demand, she adjusts upward the offer price, but not the full fair price. This partial adjustment creates positive first day return, which is used to compensate informed investors for revealing truthful information during the book building process. We investigate Brazilian IPOs issued between 2004 and 2012 and find evidence similar to the findings in the US. The launching price is lower than the first day closing price, and underwriter increases the number of shares in the aftermarket with the overallotment option. The new shares issued in the aftermarket reduce the “money left on the table” and at the same time compensates informed investors. Surprisingly the underpricing is less than half of the one reported in the US.Artigo Científico Private Equity and Venture Capital Investments in Brazilian Companies in the Last 30 Years(2015) ANDREA MARIA ACCIOLY FONSECA MINARDI; Bassani, Rafael; Kanitz, Ricardo; Moreira Neto, José Candido; Pechlyie, KarimThis study presents an overview of Brazilian Private Equity and Venture Capital (PE and VC) investments in portfolio companies between 1982 and 2014. Main source of information is Private Placement Memoranda. According to our data, the majority of capital was invested in PE, and most of the deals were in late stage companies (PE type). We observe that deal size has a cyclical behavior for PE and VC, and that the cycle reflects the dynamic of the economic environment and regulatory changes. PE deals have a mean MoM of 3.4. Although 15% of the deals were total losses (or close to total losses), 18% of the deals had extraordinary performance (MoM greater than 5 times invested capital). VC deals have a mean MoM of 1.5, less than half of PE’s MoM. Total losses correspond to 39% of the deals, what is usual in the international industry, but only 5% of the deals were high performers. VC industry was basically sponsored by government resources before 2010, and our sample reflects this period. In an underdeveloped VC ecosystem, it is difficult to fund new rounds for early stage companies’ growth, and therefore to create a high number of outperformers. We expect to have higher returns for VC in future analysis, since there has been significant evolution in the Brazilian early stage ecosystem in the last three or four years.