Two real time series, one discussing the area of energy, other discussing the area of economy. By the energetic area we will be dealing with the electric power consumption in the USA, by the economic area we will be dealing with the progress of index PX50. We will try to approve the validity of hypothesis that with some test functions we will be able to set down the accidental unit distribution in these two time series.

Predicting sales is an important step for many business processes. Several forecasting methods have been applied to uncountable different problems, however with no present research found in the area of renewable energy certificates. Thus, this study aims to examine the possibility of developing a model based on traded volumes of certificates, where a comparison between simpler and more complex models explores the general increased interest in machine learning models. Therefore, five different models are tested with monthly sales data: the statistical model Seasonal Autoregressive Integrated Moving Average, the machine learning models Support Vector Regression and Extreme Gradient Boosting and further the neural networks Long Short-Term Memory and Bidirectional Long Short-Term Memory. Extensive data preparation is operated by taking into account seasonality and trends where data transformations are applied in addition to feature engineering. To evaluate the models, non-aggregated monthly forecasts as well as aggregated predictions of two and three months are examined. The results show that it is feasible to model the sales volumes of renewable energy certificates. As expected, the models generally perform better when evaluated based on aggregated monthly predictions. Also, when considering both evaluation strategies, the Seasonal Autoregressive Integrated Moving Average, Support Vector Regression and Extreme Gradient Boosting are the only models showing better performance compared to a baseline model. The proposed solution to enable smarter and more efficient trading decisions today is a combination of the aggregated two months and quarterly predictions of the Seasonal Autoregressive Integrated Moving Average and Support Vector Regression models. Considering an expected expansion of relevant available data for the company, the recommendation for the future is to specifically further develop the machine learning models with an anticipation of improved performance and valuable feature importance insights.

No description available.

The objective of this thesis is to forecast the number of people registered at the Swedish Public Employment Service (Arbetsförmedlingen) that will manage to get employment each month and examine how accurate the forecasts are. The Swedish Public Employment Service is a government-funded agency in Sweden working to keep the unemployment rate low. When someone is unemployed or looking for a new job, he or she can register at the Swedish Public Employment Service. Being able to forecast well how many are expected to get employment could be useful when planning and making decisions. It could also be used as an indicator of how well the Swedish Public Employment Service manages to perform and thus how well the tax money is used. The models employed for forecasting were the seasonal autoregressive integrated moving average (SARIMA) and the long short-term memory (LSTM). A persistence model is also used as a baseline. The persistence model is a very simple model and the other models are therefore expected to outperform it. For the LSTM model, the use of both univariate and multivariate approaches will be explored in order to examine if the model can be improved with more data. Results from the experiments performed showed that a multivariate LSTM performed the lowest root mean squared error (RMSE) and is therefore considered the best model. However, the robustness of the model over time needs further research. / Syftet med detta arbete är att göra prognoser på hur många av de registrerade på Arbetsförmedlingen som kommer att få arbete en viss månad och undersöka hur noggranna dessa prognoser blir. Arbetsförmedlingen är en skattefinansierad myndighet i Sverige som arbetar med att hålla arbetslösheten låg. När någon är arbetslös eller letar efter ett arbete kan man registrera sig hos Arbetsförmedlingen. Att kunna göra bra prognoser på hur många som kommer att få arbete skulle kunna vara användbart vid planering och beslutfattande. Det skulle också kunna användas som en indikator på hur väl Arbetsförmedlingen använder skattepengarna. De modeller som har använts är seasonal autoregressive integrated moving average (SARIMA) och long short-term memory (LSTM). En persistensmodell används också som baslinje. Persistensmodellen är en enkel modell och därför förväntas de andra modellerna prestera bättre. För LSTM-modellen kommer användningen av både envariabla och flervariabla tillvägagångssätt att undersökas för att testa om mer data kan förbättra modellen. Resultat från experimenten visar att det var en LSTM-modell med flera variabler som presterade lägst root mean squared error (RMSE) och anses därför vara den bästa modellen. Det behövs dock ytterligare studier för att undersöka modellens stabilitet över tid.

The purpose of this thesis was to investigate whether some successful ideas in NLP, such as word2vec, are applicable to time series prob- lems or not. More specifically, we are interested to assess a combina- tion of previously proven methods such as SAX and Word2vec. Based on a rolling window approach, we applied SAX to create words for each window. These words formed a corpus on which we performed Word2vec, which served as inputs in a time series forecasting setting. We found that for forecasting horizons of longer length, our proposed method showed an improvement over statistical models under certain conditions. The findings suggest that bringing tools from the natural language processing domain into the time series domain may be an ef- fective idea. Further research is necessary to broaden the knowledge of these types of methods by testing alternative options for the cre- ation of words. Hopefully, this work will motivate other researchers to investigate this type of solution further.

This study examines three applications of control charts used for monitoring syndromic data with different characteristics. The first part develops a seasonal autoregressive integrated moving average (SARIMA) based surveillance chart, and compares it with the CDC Early Aberration Reporting System (EARS) W2c method using both authentic and simulated data. After successfully removing the long-term trend and the seasonality involved in syndromic data, the performance of the SARIMA approach is shown to be better than the performance of the EARS method in terms of two key surveillance characteristics, the false alarm rate and the average time to detect the outbreaks. In the second part, we propose a generalized likelihood ratio (GLR) control chart to detect a wide range of shifts in the mean of Poisson distributed biosurveillance data. The application of a sign function on the original GLR chart statistics leads to downward-sided, upward-sided, and two-sided GLR chart statistics in an unified framework. To facilitate the use of such charts in practice, we provide detailed guidance on developing and implementing the GLR chart. Under the steady-state framework, this study indicates that the overall GLR chart performance in detecting a range of shifts of interest is superior to the performance of traditional control charts including the EARS method, Shewhart charts, EWMA charts, and CUSUM charts. There is often an excessive number of zeros involved in health care related data. Zero-inflated Poisson (ZIP) models are more appropriate than Poisson models to describe such data. The last part of the dissertation considers the GLR chart for ZIP data under a research framework similar to the second part. Because small sample sizes may influence the estimation of ZIP parameters, the efficiency of MLEs is investigated in depth, followed by suggestions for improvement. Numerical approaches to solving for the MLEs are discussed as well. Statistics for a set of GLR charts are derived, followed by modifications changing them from two-sided statistics to one-sided statistics. Although not a complete study of GLR charts for ZIP processes, due to limited time and resources, suggestions for future work are proposed at the end of this dissertation. / Ph. D.

La Tension dans un Service d’Urgences (SU) est un déséquilibre entre le flux de charge des soins et la capacité de prise en charge sur une durée suffisante pouvant entrainer des conséquences néfastes au bon fonctionnement. Elle se reflète par la surcharge des locaux, l’allongement des délais de traitement et d’attente. Ce qui provoque à la fois l’insatisfaction des patients et l’anxiété du personnel. Cette thèse s’inscrit dans le cadre du projet HOST financé par le programme ANR-TECSAN-2011 afin d’élaborer un Système d'Aide à la Gestion de la Tension (SAGeT) assurant trois objectifs:1. L’évaluation multicritère grâce à une panoplie d’indicateurs agrégés par la logique floue afin de résoudre la subjectivité du ressentie humain de la tension. Chaque scénario d’évaluation déclenche des règles de décision spécifiques ciblant ainsi des points de défaillance à surveiller.2. L’anticipation de la demande sur différents horizons temporels : l’application des méthodes SARIMA et SARIMAX est justifiée par la saisonnalité des chroniques de visites et l’influence de certains paramètres externes (épidémies, vacances, météo). De plus, la qualité de l’information venant de l’historique a été améliorée par une recomposition d’historique basée sur la vraisemblance journalière.3. L’amélioration de la gestion des flux et le pilotage de l’activité puisque l’utilisation de SAGeT comme un tableau de bord offre une vue macro sur l’ensemble de l’activité (lits occupés, patients en attente, durées de passages prévisionnelles et allongements excessifs). Les simulations traitent des vrais scénarios de tension observés entre 2011 et 2013 dans le SU Pédiatriques Jeanne de Flandre du CHRU-Lille. / He strain in an Emergency Department (ED) is an imbalance between the total demand load of healthcare treatment and resources ability to support it during a convenient horizon, which may results negative consequences on the smooth running of the activity. It is reflected by overcrowding, longer treatment and waiting times which causes both patients dissatisfaction and anxiety of personnel. This thesis is part of the HOST project funded by the ANR-TECSAN-2011 program to develop a Management Support System of Strain (MSSS) ensuring three objectives:1. Multi-criteria evaluation through a variety of indicators aggregated by fuzzy logic to solve the subjectivity of the human feeling of strain. Each evaluation scenario involves specific decision rules targeting to supervise failure points.2. Demand forecasting through several time horizons: applying SARIMA and SARIMAX methods is justified by the time series seasonality of visits and the influence of some external parameters (epidemics, holidays, weather). In addition, the quality of the historical information has been improved by a history rebuilding based on the daily likelihood.3. Improving flow management and activity monitoring since the use of MSSS as a dashboard provides a macro view of the whole activity (beds occupied, waiting, estimated length of stay, excessive elongation).The simulations address real strain scenarios observed between 2011 and 2013 in the Pediatric ED Jeanne de Flandre of the Regional University Hospital of Lille (France).

Submitted by Adriana Bezerra Bessa (adrianabbessa@gmail.com) on 2018-05-09T00:07:09Z No. of bitstreams: 1 Tese_AdrianaBessa_versaofinal.pdf: 4846338 bytes, checksum: 5d2e8d52cd770e8fd17a4a9adee180d2 (MD5) / Approved for entry into archive by Thais Oliveira (thais.oliveira@fgv.br) on 2018-05-10T17:26:20Z (GMT) No. of bitstreams: 1 Tese_AdrianaBessa_versaofinal.pdf: 4846338 bytes, checksum: 5d2e8d52cd770e8fd17a4a9adee180d2 (MD5) / Approved for entry into archive by Suzane Guimarães (suzane.guimaraes@fgv.br) on 2018-05-11T12:30:07Z (GMT) No. of bitstreams: 1 Tese_AdrianaBessa_versaofinal.pdf: 4846338 bytes, checksum: 5d2e8d52cd770e8fd17a4a9adee180d2 (MD5) / Made available in DSpace on 2018-05-11T12:30:08Z (GMT). No. of bitstreams: 1 Tese_AdrianaBessa_versaofinal.pdf: 4846338 bytes, checksum: 5d2e8d52cd770e8fd17a4a9adee180d2 (MD5) Previous issue date: 2018-04-24 / A previsão de vendas é um aspecto crítico para maior parte das organizações, já que permite tornar o processo de planejamento mais eficiente, impactando assim nos resultados a serem obtidos pelas empresas. Entre as diversas técnicas de previsão, temos o grupo de métodos estatísticos clássicos e os métodos avançados, que trazem uma contribuição no tratamento das não linearidades. É neste contexto, que surge o problema desta dissertação: Quais são as técnicas que apresentam maior acurácia quando aplicadas para previsão de vendas no varejo de moda? Para responder a esta questão, esse trabalho avaliou dez métodos de previsão: Naive, SARIMA, SARIMA com exógenas, SARIMA GARCH, SARIMA GARCH com exógenas, método atual utilizado pela empresa estudada, rede neural MLP, rede neural MLP com exógenas, rede neural recorrente LSTM e rede neural recorrente LSTM com exógenas para quatro séries de quantidades vendidas de categorias de produtos distintas de uma empresa varejista do setor de moda. É fundamental destacar, que de forma casual, a pesquisa identificou que as quatro séries semanais de vendas dos produtos analisados são estacionárias, considerando um período longo de dez anos, o que por si só já é um resultado relevante. A análise dos diversos métodos de previsão para cada série de produto mostrou que os métodos avançados superaram os métodos estatísticos clássicos e, mais especificamente, a rede neural recorrente LSTM foi a que apresentou a maior precisão. Sendo assim, não há dúvidas que adoção dos métodos avançados para as empresas, que atuam no varejo de moda, pode trazer melhorias significativas em termos de gestão de estoque, de gestão da cadeia de abastecimento e de gestão de caixa, garantindo um aumento de eficiência e dos resultados das mesmas. De forma prática, para a empresa estudada foi obtido um incremento de acuracidade de 54,32%. / The sales forecasting is a critical aspect for most organizations, since it allows to make the planning process more efficient, thus impacting the results to be obtained by the companies. Among the various forecasting techniques, we have the group of classical statistical methods and the advanced methods, which make a contribution in the treatment of nonlinearities. It is in this context, that the problem of this dissertation arises: What are the techniques that present the greatest accuracy when applied to forecast sales in fashion retail? In order to answer this question, this study evaluated ten predictive methods: Naive, SARIMA, SARIMA with exogenous, SARIMA GARCH, SARIMA GARCH with exogenous, current method used by the studied company, MLP neural network, MLP neural network with exogenous, recurrent neural network LSTM and LSTM recurrent neural network with exogenous for four series of quantities sold from product categories distinct from a retailer in the fashion industry. It is important to highlight that, on a casual basis, the research identified that the four weekly series of sales of the analyzed products are stationary, considering a long period of ten years, which in itself is already a relevant result. The analysis of the various prediction methods for each product series showed that the advanced methods overcame the classic statistical methods and, more specifically, the recurrent neural network LSTM was the one that presented the highest precision. Therefore, there is no doubt that adoption of the advanced methods for companies that operate in fashion retail can bring significant improvements in terms of inventory management, supply chain management and cash management, ensuring an increase in efficiency and in its results. In practice, for the company studied, an accuracy increase of 54.32% was obtained.

[pt] Nesta dissertação é desenvolvida uma metodologia para previsão de demanda mensal de energia elétrica considerando cenários de racionamento. A metodologia usada consiste em, a partir das taxas de crescimento da série temporal, identificar e eliminar os efeitos do racionamento de energia elétrica através da aplicação de Modelos Lineares Dinâmicos. São analisadas também estruturas de intervenção nos modelos estatísticos de Box & Jenkins e Holt & Winters. Os modelos são então comparados segundo alguns critérios, basicamente no que tange à sua eficiência preditiva. Conclui-se ao final sobre a eficiência da metodologia proposta, dado a grande dificuldade para solucionar o problema a partir dos modelos estatísticos de Box & Jenkins e Holt & Winters. Esta solução é então proposta como a mais viável para criar cenários de racionamento e pósracionamento de energia para ser utilizado por agentes do sistema elétrico nacional. / [en] In this dissertation, a methodology is developed to forecast monthly demand of electric energy, considering the rationing scenery. The methodology is based on, taking the growth rate from the time series, identify and eliminate the effects of electric energy rationing, using Dynamic Linear Models. It is also analyzed intervention structures in the statistics models of Box & Jenkins and Holt & Winters. The models are compared according to some criterions, mainly forecast accuracy. At the end, we concluded that the methodology proposed is more efficient, due to the difficult to solve the problem using the statistics models with intervention. This solution is proposed as the best among them to create scenery during the energy rationing and after energy rationing, to be used by the national electric system agents.

Since the 2008 financial crisis, the interest for the subject area of modelling non-maturity deposits has been growing quickly. The area has been widely analysed from the perspective of a traditional bank where customers foremost have transactional and salary deposits. However, in recent year the Swedish banking sector has become more digitized. This has opened up opportunities for more niche banking actors to establish themselves on the market. Therefore, this study aims to examine how the theories developed and previously used in modelling liquidity volumes at traditional banks can be used at a niche bank focused on savings and investments. In this study the topics covered are short-rate modelling using Vasicek's model, liquidity volume modelling using SARIMA and SARIMAX modelling as well as liquidity risk modelling using an approach developed by Kalkbrener and Willing. When modelling the liquidity volumes the data set was divided depending on account and customer type into six groups, for four out of these the models had lower in and out of set prediction errors using SARIMA models for only two of the six models were there improvements made to the in and out of set prediction error using SARIMAX models. Finally, the resulting minimization of liquidity volume forecasting 5 years in the future gave reasonable and satisfactory results. / Sedan finanskrisen 2008 har intresset kring ämnesområdet gällande modellering av inlåningsvolymer utan en kontrakterad förfallodag ökat snabbt. Området har analyserats i stor utsträckning från perspektivet av en traditionell bank där kunder har framförallt transaktions- och lönekonton. De senaste åren har den Svenska banksektorn blivit mer digitaliserad. Detta har öppnat upp möjligheter för nischbanker att etablera sig på marknaden. Därför ämnar denna studie att undersöka hur teorier som har utvecklats och tidigare använts på traditionella banker för att modellera likviditetsvolymer kan användas på en nischbank som är fokuserad på sparande och investeringar. I denna studie modelleras korträntor med Vasicek's modell, likviditetsvolymer med SARIMA och SARIMAX modeller och likviditetsrisk med en modell utvecklad av Kalkbrener och Willing. För modelleringen av likviditetsvolymer delades likviditetsdatan upp i sex grupper baserat på konto- och kund typ. För fyra av dessa data set gav SARIMA-modeller lägre prediktionsfel och endast för två av de sex grupperna gav SARIMAX-modeller bättre resultat. Slutligen så gav den resulterande minimeringen av nödvändiga likviditetsvolymer på en 5 årig horisont rimliga och tillfredsställande resultat.