For more than a year, COVID-19 has changed societies all over the world and put massive strains on its healthcare systems. In an attempt to aid in prioritizing medical resources, this thesis uses dynamic regression with ARIMA errors to forecast the number of hospitalizations related to COVID-19 two weeks ahead in Uppsala County. For this purpose, 100 models are created and their ability to forecast hospitalizations two weeks ahead for weeks 15-17 of 2021 for the different municipalities in Uppsala County is evaluated using root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). The best performing models are then utilized to forecast hospitalizations for weeks 19-22. The results show that the models perform well during periods of increasing numbers of hospitalizations during early 2021, while they perform less well during the last weeks of May 2021 where hospitalizations numbers have been falling dramatically. This recent decrease in forecasting performance is believed to be caused by an increase in vaccination coverage, which is not accounted for in the models.

Background Forecasts are used as a basis for decision making and they mainly affect decisions at strategic and tactical levels in a company or organization. There are two different methods to perform forecasts. The first one is a qualitative method where a n expert or group of experts tell about the future. The second one is a quantitative method where forecast are produced by mathematical and statistical models. This study used a quantitative method to build a forecast model and took into account external f actors in forecasting the sales volume of Bosch Rexroth’s hydraulic motors. There is a very wide range of external factors and only a limited selection had been analyzed in this study. The selection of the variables was based on the markets where Bosch Rexroth products are used, such as mining. Purpose This study aimed to develop five predictive models: one model for the global sales volume, one model each for sales volume in USA and China and one model each for sales volume of CA engine and Viking engine. By identifying external factors that showed significant relationship in various time lags with Bosch Rexroth’s sales volume, the forecasts 2016 and 2017 were produced. Methods The study used a combination of multiple linear regression and a Box - Jenkins AR MA errors to analyze the association of external factors and to produce forecasts. Externa l factors such as commodity prices, inflation and exchange rates between different currencies were taken into account. By using a cross - correlation function between external factors and the sales volume, significant external factors in different time lags were identified and then put into the model. The forecasting method used is a Causal forecasting model. Conclusions The global sales volume of Bosch Rexroth turned out to be affected by the historical price of copper in three different time lags , one, six and seven months . From 2010 to 2015, the copper price have been continuously dropping which explain s the downward trend of the sales volume. The sales volume in The U SA showed a significant association by the price of coal with three and four time lags. This means that the change of coal price takes three and four months before it affects the sales volume in the USA. The market in China showed to be affected by the development of the price of silver. The volume of sales is affected by the price of silver by four and six time lags. CA engine also displayed association with the price of copper at the same time lags as in the global sales volume. On the other hand, Viking engine showed no significant association at all with any of the external factors that were analyzed in this study. The forecast for global mean sales volume will be between 253 to 309 units a month for May 2016 – December 2017. Mean sales volume in USA projected to be in between 24 to 32 units per month. China's mean sales volume is expected to be in between 42 to 81 units a month. Mean sales volume of CA engine has a forecast of 175 to 212 units a month. While the mean s ales of Viking engine projected to stay in a constant volume of 25 units per month.

[pt] O transporte marítimo de petróleo e derivados é componente fundamental da cadeia de suprimento da indústria do petróleo, integrando fornecedores e clientes localizados em regiões geográficas distintas. Neste contexto, os valores de fretes praticados possuem grande impacto no comércio internacional destes bens. O objetivo deste trabalho é verificar o desempenho de modelos de Regressão Dinâmica em previsões de frete marítimo de curto prazo do mercado spot de uma rota de exportação de petróleo do oeste da África para a China, comparar a capacidade preditiva do modelo com métodos tradicionais, vastamente discutidos na literatura, como Amortecimento Exponencial e modelos ARIMA e projetar cenários para avaliar como as variáveis explicativas presentes no modelo de Regressão Dinâmica proposto neste estudo afetam o frete marítimo. O produto desenvolvido nesta dissertação mostrou a viabilidade de os modelos univariados e causais serem utilizados como ferramenta de previsão da taxa frete de navios petroleiros. Como forma de validação, os resultados foram comparados aos obtidos com a metodologia vigente em uma grande empresa de petróleo do Brasil. O protótipo de sistema de previsão proposto, via Regressão Dinâmica, apresentou resultados satisfatórios e desempenho superior ao obtido através da metodologia da empresa de petróleo. / [en] Crude oil and oil products seaborne transportation is a key component of the petroleum industry supply chain, integrating suppliers and customers located in different geographic regions. In this context, the freight rates practiced have a great impact on the international trade of these goods. This work aims to verify the performance of Dynamic Regression models in short-term maritime freight forecasts of the spot market of an oil export route from West Africa to China, to compare the predictive capacity of the model with traditional methods, widely discussed in the literature, such as Exponential Smoothing and ARIMA models and to design scenarios to evaluate how the explanatory variables present in the Dynamic Regression model proposed in this study affect freight rate. The product developed in this dissertation showed the viability of the univariate and causal models being used as a forecasting tool for the oil tankers freight rate. As a form of validation, the results were compared to those obtained with the methodology of a large Brazilian oil company. The proposed prediction system prototype, through Dynamic Regression model, presented satisfactory results and performance superior to that obtained through the methodology of the oil company.

[pt] O mercado de smartphones é muito sensível a preços devido à alta competitividade comercial e à constante evolução tecnológica. A elasticidade de preços e a elasticidade cruzada são fundamentais para ajustes de previsões de vendas para evitar rupturas e/ou altos volumes de estoques. Este trabalho apresenta uma proposta de modelagem para cálculo de elasticidade de preços e elasticidade cruzada utilizando variáveis causais abordando aspectos internos e externos de uma empresa operadora de serviços de telecomunicações com várias lojas próprias. A escolha das variáveis é resultante da parceria entre profissionais de Logística, Marketing e Vendas fornecendo apoio técnico aos Planejadores de Demanda. Para se calcular a elasticidade de preços, a modelagem baseada em Regressão Dinâmica indicou utilização das variáveis: preços de concorrentes internos (representando a canibalização), disponibilidade (para lançamento e phase-out de produtos), loja aberta (diferenciando dos dias de lojas fechadas com vendas nula) e fator diário (cadenciando as vendas diárias), proporcionando resultados satisfatórios e demonstrando aplicabilidade comercial do modelo proposto. / [en] The smartphone market is very price sensitive due to high commercial competitiveness and constant technological evolution. Price elasticity and cross-elasticity are critical for adjusting sales forecasts to avoid disruptions and / or high inventory volumes. This work presents a modeling proposal for calculation of price elasticity and cross elasticity using causal variables, addressing the internal and external aspects of a telecommunications service operator with several own stores. The variable s choice is the result of a partnership between professionals in Logistics, Marketing and Sales providing technical support to Demand Planners. In order to calculate price elasticity, modeling based on dynamic regression indicated the use of variables: internal competitors prices (typifying cannibalization), availability (for launching and phase-out of products), open store (differentiating from the days of closed stores with zero sales) and daily factor (daily sales rhythm), providing satisfactory results and demonstrating commercial applicability of the proposed model.

This thesis aims to fit and compare different time series models namely the ARIMA-model, conditional heteroscedastic models and lastly a dynamic regression model with ARIMA error to Bitcoin closing price data that spans over 5 consecutive years. The purpose is to evaluate if the sanction on Russia had any effect on the cryptocurrency Bitcoin. After giving a very brief introduction to time series models and the nature of the error term, we describe the models that we want to compare. Quite early in on, autocorrelation was detected and that the time series were nonstationary. Additionally, as we are dealing with financial data, we found that the best alternative was to transform the data into logarithmic return and we then took the first difference. As we then detected a very large outlier, we decided to replace the extreme value with the mean of the two adjacent observations as we suspected it would affect the forecast interval. The dataset with first differenced log-returns was used in the ARIMA model but it turned out that there was no autocorrelation which indicated that returns in financial assets are uncorrelated across time and therefore unpredictable. The conditional heteroscedastic models, the ARCH and the GARCH models turned out to be best suitable for our data, as there was an ARCH-effect present. We could conclude that the GARCH(1,1) model using student t-distribution had the best fit, which had the lowest AIC and the highest log likelihood. In order to study the effect of the sanctions on Bitcoin volatility a dynamic regression model was used by allowing the error term to contain autocorrelation and include an independent dummy variable. The model showed that the Russian invasion of Ukraine did not, surprisingly, have any effect on the Bitcoin closing price.

[pt] O objetivo do presente trabalho é a geração de uma série mensal de carga elétrica livre das variações de calendário e de temperatura. Para tal, foram comparadas duas abordagens, uma totalmente empírica e outra híbrida com métodos empíricos e modelagens de regressão dinâmica, para identificar a mais adequada para a retirada desses ofensores. Os dados utilizados são provenientes de observações diárias de cada um dos quatro subsistemas que integram o Sistema Interligado Nacional (SIN), porém a ideia é produzir séries mensais do SIN e não apenas de cada um dos subsistemas. A série trimestral do PIB foi utilizada para decidir qual abordagem melhor ajustou os dados de Carga. A série mensal de carga ajustada do SIN será utilizada para subsidiar decisões, de compra e venda de energia nos leilões, das empresas distribuidoras de energia elétrica. / [en] This thesis proposes a method to generate monthly load series free of variations coming from two sources: calendar and temperature. Two approaches were considered, one totally empirical and another one called hybrid, as it use empirical procedure to remove the calendar effect and a dynamic regression type of model to remove the temperature effects. The data set used comes found to daily observations from each one of the four subsystems that form the SIN (Brazilian Integrated Grid). However the final task is to obtain a unique monthly series for the SIN and not only the four subsystems monthly series. The quarterly PIB series was used to check the performance of the two proposed methods. Such adjusted series are quite important tools to hold on the decision of acquisitions and dailes of energy in the energy audits.

[pt] Esta dissertação tem como objetivo principal a proposição de modelos para previsão, em um curto prazo, do número de processos que são ajuizados em desfavor de uma empresa do setor elétrico. A metodologia utilizada consiste em, a partir de uma análise exploratória dos dados, construir modelos usando uma estratégia bottom-up, ou seja, parte-se de um modelo simples e processa-se seu refinamento até encontrar um modelo apropriado que mais se adeque à realidade. Partiu-se então de um modelo auto projetivo indo até uma formulação de um modelo de regressão dinâmica. Os modelos são então comparados segundo alguns critérios, basicamente no que tange à sua eficiência preditiva. Conclui-se ao final sobre a eficiência de se utilizar modelos de regressão dinâmica para este tipo de previsão tendo em vista a presença de correlação serial dos resíduos, comumente presentes nas séries econômicas. Propõe-se, ao final, uma ferramenta para, a partir dos valores estimados, analisar a viabilidade econômica de estimular ou desestimular as medidas responsáveis pela geração de processos contra a empresa. / [en] The aim of this dissertation is to develop short term models to forecast the number of judicial process in electric sector companies. From the methodology point of view, data is analyzed and models using bottom-up strategy is developed. In other words, a simple model is improved step by step until a proper model that fits well the reality is found. From a univariate model it ends up in a dynamic regression model. The models obtained in this study are compared according to some criterion, mainly forecast accuracy. In the end the conclusion is about the efficiency of dynamic regression models for this kind of forecast, which one presents data with serial correlation of residues, commonly present in economic series. In the end, from the estimated values, it´s proposed a mechanism to analyze the economic viability, to encourage or not, actions which are responsible for instigating judicial processes against the company.

Bakgrund: Denna uppsats riktar in sig på hotell i Stockholm och all data som anges gäller för staden som helhet. Inom Hotellbranschen finns det tre vedertagna nyckeltal som kan sägas beskriva hur det går ekonomiskt för ett hotell. Då hotellen till stor del styrs efter dessa tre nyckeltal så är det av stort intresse för varje enskilt hotell att jämföra sina egna värden med marknadens värden på dessa nyckeltal. Om prognoser utförs på dessa nyckeltal borde det vara av stort intresse för varje hotell att ta del av dessa prognoser för att på så vis kunna reglera prissättningen utefter hur marknaden kommer att se ut den närmaste tiden. Syfte: Ta fram modeller som utifrån framtida evenemang och framtida bokningsläge prognostiserar hotellmarknadens Beläggning och Rumsintäkter. Utifrån dessa prognoser beräknas nyckeltalen Beläggning, Snittpris och intäkt per disponibelt rum på dagsnivå ett år fram i tiden, det vill säga för år 2016. Metod: Då datamaterialet består av tidsserier med tillhörande förklarande variabler används en typ av dynamisk regressionsmodell. Dessa modeller är utformade för att hantera tidsseriedata med tillhörande förklarande variabler. Modellen som används kallas för regression med ARMA-fel och syftar till att en multipel regression anpassas och en lämplig ARMA-modell tas fram för att förklara feltermerna. På så vis modelleras även autokorrelationen som annars finns kvar i feltermerna. Resultat: Modellen för Beläggningen består av fyra förklarande variabler och feltermerna antas följa en AR-struktur. Rumsintäkterna prognostiseras med en modell med sju förklarande variabler, även för denna modell antas feltermerna följa en AR-struktur. Det tycks också finnas en säsong i data vilken också modelleras i form av en AR-struktur för de båda modellerna. Prognosen för nyckeltalen ser till största del ut att följa föregående års mönster, och evenemangs-typen Event ger oftast en hög skattning i förhållande till månaden. Evenemangstypen Högtid tycks ge en negativ effekt och Bokningsläget har en positiv effekt för båda modellerna. Slutsats: Modellerna anses välanpassade men det krävs mer bearbetning på de förklarande variablerna där till exempel event bör grupperas in beroende på vad för slags event det är. För att prognostisera rumsintäkter bör en variabel som förklara hotellens prisjusteringar modelleras. / Background: This thesis targets hotels in Stockholm with aggregated data for the city. In the hotel market there’s three key indicators of particular interest and can be said describes how the market goes. Because of how much influence these key indicator have on the hotels it’s in great interest for the hotels to compare themselves with the market values. If these key indicators where forecasted it would perhaps be of great interest for the hotels to buy these forecasts to be able to control the room pricing in advance. Purpose: Develop forecasting models due to future event and bookings with occupancy and room revenue as response variables. The key indicators revenue per available rooms and average price is then calculated through these forecasts for the year 2016. Method: Since data consist of response variables (called output series) where the future values this series depends on past values of this series and a multiple set of related time series and external events (called input series) a dynamic regression called “regression with ARMA errors” where used. The method implies that you suit a multiple regression where the error terms are modelled with an appropriate ARMA model. Results: The model for occupancy consist of four dependent variables and the model for the room revenue contain seven dependent variables. The error terms for these models include an autoregressive model with both seasonal and non-seasonal orders. The forecast for the key indicators seems to follow the same pattern as previous years, where the event type Event more often than not gives a high estimate in relation to the current month. The event type Holiday seems to have a negative impact and bookings has a small positive effect for both models. Conclusions: The models seems to fit data well but the input series needs more processing where the variable event seems to need some subgrouping. To forecast the room revenue is seems like a variable explaining price changes need to be constructed.

The goal of the thesis is to accurately predict future values of a company’s website visits and to estimate the uncertainty of those predictions. To achieve this, a dynamic regression model with an ARIMA error term is considered, using advertisement spending with lags and dummy variables for Black Friday and weekdays as predictors. After dividing the data into a training set and a test set, the order of the ARIMA error term is specified using the Box-Jenkins methodology. The initial model is then run through a backward elimination algorithm, which selects two models based on the Akaike Information Criterion and Bayes Information Criterion. As expected, the model selected using Bayes Information Criterion is more conservative in its choice of variables than the model specified using the Akaike Information Criterion. The forecasts made on the test set are complemented with normal and bootstrap-based prediction intervals in order to estimate the uncertainty of the predictions. These are then compared to the forecasts made using a simpler model, consisting of only the ARIMA error term. The thesis concludes that the dynamic regression models are twice as accurate as the simpler model and that they were on average off by 14% from the actual values. The prediction intervals for the dynamic regression models are slightly too pessimistic as they overstate the uncertainty of the model by about 10 percentage points in the 80% prediction interval and by 5 percentage points in the 95% prediction interval. There is no practical discrepancy in prediction power between the model selected using the Akaike Information Criterion and the one using Bayes Information Criterion. The accuracy of the prediction intervals is higher than in the simpler model even though both dynamic regression models have more residual autocorrelation.

The scarce resources of public health makes the administrator manage the destination of resources, aiming to rationalize and optimize its collection, in order to improve the assistance to patients because the hospital is a public institution and does not get profits but promotes the community well-being. Thus, the hospital infection is acquired after the patient comes to the hospital of after he goes home and might be associated with his staying in hospital or with hospital procedures. This cost must be avoided. Once the complete eradication is not impossible, it is necessary to analyze and to control the monthly cost of the main antibiotics used for its treatment so that there is enough knowledge to foresee the resource collection to buy them. In this context, the main reason of this research is to carry out a forecast of the monthly cost and of the resource collection needed to purchase those medicine used in the treatment of hospital infections at the University Hospital of Santa Maria. To do so, a methodology for forecast by dynamic and multiple linear regressions was used. They were combined with to a multivariate technique by principal components. The technique of principal components was used to eliminate the multiple linearity existing among the original variants so, the resulting principal components were used as variables in the construction of the model of multiple linear regression and of dynamic regression. Therefore, these methodologies are applied to a case study of public health, in order to foresee and to conclude about which model is more suitable to forecast the monthly cost of antibiotics in hospital infections. The results obtained from the two models were considered satisfactory but the model of dynamic regression was chosen to be more suitable because it presented a mean absolute percentage error (MAPE). Finally, the findings might be a managerial tool for hospital administration when they offer subsides for the budget of planning and of the resource finances, especially in a time when resources are globally scarce, making health even more expensive. / Os escassos recursos da saúde pública impõem ao administrador gerenciar a destinação dos recursos buscando racionalizar e otimizar sua alocação, permitindo, desta forma, melhorar o atendimento aos pacientes, pois o hospital, sendo uma entidade pública, não tem por objetivo o lucro, mas sim promover o bem estar da comunidade. Com isso, a infecção hospitalar que é adquirida após a internação do paciente e se manifesta durante a internação ou mesmo após a alta, podendo ser relacionada com a internação ou procedimentos hospitalares, deve ser evitada. Uma vez que sua total erradicação não é possível, se faz necessário analisar e controlar o custo mensal dos principais antibióticos utilizados no seu tratamento a fim de se ter embasamento suficiente para prever a alocação de recursos para sua aquisição. Nesse contexto, o principal objetivo desta pesquisa é realizar a previsão do custo mensal e de alocação de recursos necessários para aquisição de medicamentos utilizados no tratamento de infecções hospitalares no Hospital Universitário de Santa Maria. Para isso, utilizou-se a metodologia de previsão por regressão linear múltipla e de regressão dinâmica combinada com a técnica multivariada de componentes principais que foi utilizada para eliminar a multicolinearidade existente entre as variáveis originais. Com isso, as componentes principais resultantes foram utilizadas como variáveis independentes na construção do modelo de regressão linear múltipla e de regressão dinâmica. Portanto, essas metodologias são aplicadas a um estudo de caso na saúde pública, a fim de fazer previsões e tirar conclusões a respeito de qual modelo é mais adequado para realizar a previsão do custo mensal dos antibióticos em infecções hospitalares. Os resultados obtidos nos dois modelos foram considerados satisfatórios, mas foi escolhido, como modelo mais adequado para realizar as previsões, o modelo de regressão dinâmica, porque apresentou o menor erro percentual absoluto médio (MAPE). Por fim, as previsões encontradas, podem se constituir em uma ferramenta gerencial para a administração hospitalar ao fornecer subsídios para o planejamento orçamentário e financeiro dos recursos, especialmente em uma época em que há escassez de recursos em escala global, com reflexos muito intensos nos custos da saúde.