The paper production industry has witnessed significant advancements in automation, yet human expertise remains crucial for process control due to its complexity. This study proposes a knowledge and data-based assistance system to support machine operators in optimizing the production process. The system employs state/situation recognition through data analysis, utilizing machine learning algorithms and Nelson rules on sensor data to estimate or to detect the process state. By integrating error numbers from process control systems, the system generates suggestions for potential solutions. Remarkably, over 50 % of the cases reveal that the top-ranked suggestion proves to be the correct solution. Additionally, the proposed system facilitates the detection of long-term tendencies that often go unnoticed, enhancing the operator's ability to identify and address such trends. This research contributes to the ongoing efforts in automating paper production while leveraging the expertise of human operators in ensuring efficient and error-free processes.

Este trabalho tem como objetivo analisar o impacto da convergência às normas internacionais de contabilidade sobre a qualidade preditiva da informação contábil no Brasil. Particularmente, o estudo verifica o impacto da adoção do padrão International Financial Reporting Standards (IFRS) sobre: i) a acurácia das estimativas de lucro realizadas pelos analistas de mercado e ii) a dispersão dessas estimativas de lucro, além de verificar o comportamento do viés de previsão. Os dados da pesquisa foram extraídos da base Institutional Brokers Estimate System (I/B/E/S) e dos formulários de referência das empresas, no site da Comissão de Valores Mobiliários (CVM), no período de 2006 a 2012. A metodologia utilizada foi a de análise de dados em painel, com estimação de modelos de efeitos fixos e aleatórios. Para adequação dos modelos, foram utilizadas variáveis de controle comumente empregadas na literatura internacional, além de variáveis de ajuste para caso brasileiro. Os resultados do trabalho indicam que a adoção do padrão IFRS no Brasil ainda não contribuiu para melhora da qualidade preditiva da informação contábil, embora o viés de previsão tenha diminuído. A acurácia dos analistas de mercado diminuiu no período de adoção parcial do IFRS no Brasil e, no período de adoção obrigatória, as evidências encontradas não permitem concluir sobre a melhora da acurácia dos analistas. A dispersão das estimativas dos analistas de mercado aumentou no período de adoção parcial do IFRS e, no período de adoção obrigatória, não se verificou alteração no nível da mesma. Estes resultados contrariam as evidências dos estudos para a Europa e Austrália, as quais indicam elevação da qualidade preditiva das informações. Porém, assemelham-se aos resultados encontrados para o período inicial da adoção do padrão IFRS na Alemanha. A explicação para os resultados obtidos podem estar relacionadas (i) ao método de adoção do IFRS no Brasil - que incluiu um período de adoção parcial - diferentemente de outros países; e (ii) à necessidade de um tempo de aprendizado para as empresas e analistas de mercado, haja vista as alterações ainda serem recentes. Considera-se que o estudo contribui para a literatura de análise do impacto do padrão IFRS na qualidade preditiva da informação contábil no Brasil, podendo colaborar para as decisões de normatizadores sobre futuras alterações nos padrões contábeis brasileiros e auxiliar as decisões de investidores e analistas no mercado de capitais. / This study aims to analyze the impact of the adoption of International Financial Reporting Standards (IFRS) on the predictive quality of accounting information in Brazil. In particular, the study investigates the impact of the IFRS adoption on i) the accuracy of the profit forecasting by market analysts, and ii) the dispersion of these estimates, besides verifying the forecast bias. The data was extracted from the base of Institutional Brokers Estimate System (I\\/B\\E\\S) and from the forms of companies, on the website of Securities and Exchange Commission of Brazil (CVM) between 2006 and 2012. The empirical strategy employed involves the analysis of panel data and estimation of fixed-effects and random-effects models, considering control variables commonly found in international literature and specific variables for Brazilian reality. The results indicate that IFRS adoption in Brazil has not contributed to improve the predictive quality of accounting information, although the forecast bias has decreased. The forecasts accuracy decreases during the period of partial adoption of IFRS in Brazil and, for the period of mandatory adoption, this study has not found conclusive evidences about accuracy of analysts\' forecasts. Moreover, the dispersion of estimates has increased in the period of partial adoption of IFRS, however no evidence was found for the mandatory adoption period. These results are contrary to the evidences for the European and Australian cases, which suggest improvement of accounting information. Nevertheless, the results resemble the evidences encountered in Germany, particularly for the initial period of IFRS adoption. The results provided might be related (i) to the method of adoption in Brazil - which includes partial and mandatory adoption periods; and (ii) need for a time of learning period for companies and market analysts, considering the requirements are still recent. We believe that this study contributes to the literature that analyzes the impact of IFRS on the predictive quality of accounting information in Brazil. Also, might contribute to the standard-setting decisions on future changes in Brazilian accounting standards and assist the decisions of investors and research analysts.

Este trabalho tem como objetivo analisar o impacto da convergência às normas internacionais de contabilidade sobre a qualidade preditiva da informação contábil no Brasil. Particularmente, o estudo verifica o impacto da adoção do padrão International Financial Reporting Standards (IFRS) sobre: i) a acurácia das estimativas de lucro realizadas pelos analistas de mercado e ii) a dispersão dessas estimativas de lucro, além de verificar o comportamento do viés de previsão. Os dados da pesquisa foram extraídos da base Institutional Brokers Estimate System (I/B/E/S) e dos formulários de referência das empresas, no site da Comissão de Valores Mobiliários (CVM), no período de 2006 a 2012. A metodologia utilizada foi a de análise de dados em painel, com estimação de modelos de efeitos fixos e aleatórios. Para adequação dos modelos, foram utilizadas variáveis de controle comumente empregadas na literatura internacional, além de variáveis de ajuste para caso brasileiro. Os resultados do trabalho indicam que a adoção do padrão IFRS no Brasil ainda não contribuiu para melhora da qualidade preditiva da informação contábil, embora o viés de previsão tenha diminuído. A acurácia dos analistas de mercado diminuiu no período de adoção parcial do IFRS no Brasil e, no período de adoção obrigatória, as evidências encontradas não permitem concluir sobre a melhora da acurácia dos analistas. A dispersão das estimativas dos analistas de mercado aumentou no período de adoção parcial do IFRS e, no período de adoção obrigatória, não se verificou alteração no nível da mesma. Estes resultados contrariam as evidências dos estudos para a Europa e Austrália, as quais indicam elevação da qualidade preditiva das informações. Porém, assemelham-se aos resultados encontrados para o período inicial da adoção do padrão IFRS na Alemanha. A explicação para os resultados obtidos podem estar relacionadas (i) ao método de adoção do IFRS no Brasil - que incluiu um período de adoção parcial - diferentemente de outros países; e (ii) à necessidade de um tempo de aprendizado para as empresas e analistas de mercado, haja vista as alterações ainda serem recentes. Considera-se que o estudo contribui para a literatura de análise do impacto do padrão IFRS na qualidade preditiva da informação contábil no Brasil, podendo colaborar para as decisões de normatizadores sobre futuras alterações nos padrões contábeis brasileiros e auxiliar as decisões de investidores e analistas no mercado de capitais. / This study aims to analyze the impact of the adoption of International Financial Reporting Standards (IFRS) on the predictive quality of accounting information in Brazil. In particular, the study investigates the impact of the IFRS adoption on i) the accuracy of the profit forecasting by market analysts, and ii) the dispersion of these estimates, besides verifying the forecast bias. The data was extracted from the base of Institutional Brokers Estimate System (I\\/B\\E\\S) and from the forms of companies, on the website of Securities and Exchange Commission of Brazil (CVM) between 2006 and 2012. The empirical strategy employed involves the analysis of panel data and estimation of fixed-effects and random-effects models, considering control variables commonly found in international literature and specific variables for Brazilian reality. The results indicate that IFRS adoption in Brazil has not contributed to improve the predictive quality of accounting information, although the forecast bias has decreased. The forecasts accuracy decreases during the period of partial adoption of IFRS in Brazil and, for the period of mandatory adoption, this study has not found conclusive evidences about accuracy of analysts\' forecasts. Moreover, the dispersion of estimates has increased in the period of partial adoption of IFRS, however no evidence was found for the mandatory adoption period. These results are contrary to the evidences for the European and Australian cases, which suggest improvement of accounting information. Nevertheless, the results resemble the evidences encountered in Germany, particularly for the initial period of IFRS adoption. The results provided might be related (i) to the method of adoption in Brazil - which includes partial and mandatory adoption periods; and (ii) need for a time of learning period for companies and market analysts, considering the requirements are still recent. We believe that this study contributes to the literature that analyzes the impact of IFRS on the predictive quality of accounting information in Brazil. Also, might contribute to the standard-setting decisions on future changes in Brazilian accounting standards and assist the decisions of investors and research analysts.

This thesis is a case study conducted at Magna Electronics to explore the use of machinelearning techniques in improving the predictive quality of electronic control unit (ECU)within the automotive manufacturing. This thesis aims to apply interpretable machinelearning methods to predict potential future ECU failures early. With the interpretablemachine learning the goal is to identify predictive variables that lead to ECU failure andwhich can be used as support for decision making.Logistic Regression and Random Forest were chosen as the machine learning methods,which have been used in research of predictive quality and have different levels of interpretability.TreeSHAP was used on the Random Forest as the post-hoc method to furtherunderstand the results. The models’ performances were quantitatively evaluatedthrough metrics such as accuracy and area under precision-recall curve. Subsequently, thebest-performing models were further analyzed using confusion matrices, precision-recallcurves, and horizontal bar charts to assess the impact of predictive variables.The results of this thesis indicated that while Random Forest outperformed Logistic Regression,both models demonstrated limited capability in accurately predicting faulty ECUs,due to the low AUCPR scores. The precision-recall curves suggested performance near randomguess, highlighting the possible variability in parameter impact.This study has also identified significant challenges, such as data imbalance and mislabeling,which may have had a negative effect on the results. Given these issues, the thesisadvises caution in using these results for decision-making. Although, findings of this thesisunderscore the need for a cautious approach to interpreting model outputs, suggestingthat real-world application may require to use different models based on the specific goalsand context of the analysis.

Die seit Jahren anhaltende digitale Transformation erfährt durch neue, innovative Prozesse, Methoden und Technologien erneut ein atemberaubendes Wachstum in allen Bereichen. Entlang eines jeden Produktlebenszyklus werden unter den aktuellen Trends wie z.B. Systems Engineering, Industrie 4.0 und Internet of Things vielfältige Lösungen geschaffen, um vor allem die digitale Repräsentanz eines Produktes sowie der zu deren Herstellung notwendigen Produktionsmittel und der betreffenden Umgebung beim Betrieb des Produktes zu erschaffen bzw. auszubauen. Die digitale Repräsentanz, der sog. „Digitale Zwilling“ (oder auch 'Digitale Schatten') dient vor allem dazu, die Durchgängigkeit und Nachvollziehbarkeit aller produktrelevanten und -bezogenen Informationen sicherzustellen und für unterschiedlichste Szenarien und Stakeholder nutzbar zu machen. Die Informationen im Product Life-cycle Management (PLM) durchlaufen dabei unterschiedliche Reifegrade. In den Spezifikationsphasen werden die Informationen im Soll-Zustand auch gern als „Digitaler Master“ bezeichnet. [...]