This presented thesis deals with applications of Tweedie compound Poisson model in non-life insurance pricing and claims reserving. Tweedie models are exponen- tial dispersion models with power mean-variance relationships and compound Poisson distribution is a particular Tweedie model. The interest in Tweedie com- pound Poisson model is motivated by its applications to generalized linear models (GLMs) and generalized estimation equations (GEE). The purpose of this thesis is to construct pricing and claims reserving models in which the response variables follow Tweedie compound Poisson model. Theoretical approaches are applied on the real datasets. 1

Most non-life insurers and many creditors use regressions, more specifically Generalized Linear Models (GLM), to price their liabilities. One limitation with GLMs is that interactions between predictors are handled manually, which makes finding interactions a tedious and time-consuming task. This increases the cost of rate making and, more importantly, actuaries can miss important interactions resulting in sub-optimal customer prices. Several papers have shown that Gradient Tree Boosting can outperform GLMs in insurance pricing since it handles interactions automatically. Insurers and creditors are however reluctant to use so-called ”Black-Box” solutions for both regulatory and technical reasons. Tree-based methods have been used to identify pricing errors in regressions, albeit only as ad-hoc solutions. The authors instead propose a systematic approach to automatically identify and evaluate interactions between predictors before adding them to a traditional GLM. The model can be used in three different ways: Firstly, it can create a table of statistically significant candidate interactions to add to a GLM. Secondly, it can automatically and iteratively add new interactions to an old GLM until no more statistically significant interactions can be found. Lastly, it can automatically create a new GLM without an existing pricing model. All approaches are tested on two motor insurance data sets from a Nordic P&C insurer and the results show that all methods outperform the original GLMs. Although the two iterative modes perform better than the first, insurers are recommended to mainly use the first mode since this results in a reasonable trade-off between automating processes and leveraging actuaries’ professional judgment. / De flesta skadeförsäkringsbolag och många långivare använder regressioner, mer specifikt generaliserade linjära modeller (GLM), för att prissätta sina skulder. En begräsning med GLM:er är att interaktioner mellan exogena variabler hanteras manuellt, vilket innebär att hanteringen av dessa är tidskrävande. Detta påverkar försäkringsbolags lönsamhet på flera sätt. För det första ökar kostnaderna för att skapa tariffer och för det andra kan aktuarier missa viktiga interaktioner, vilket resulterar i suboptimala kundpriser. Tidigare forskning visar att Gradient Boosting kan överträffa GLM:er inom försäkringsprissättning eftersom denna metod hanterar interaktioner automatiskt. Försäkringsbolag och kreditgivare är dock motvilliga till att använda så kallade ”Black-box-lösningar” på grund av både regulatoriska och tekniska skäl. Trädbaserade metoder har tidigare använts för att hitta felprissättningar i regressioner, dock endast genom situationsanpassade lösningar. Författarna föreslår i stället en systematisk metod för att automatiskt identifiera och evaluera interaktioner innan de inkluderas i en traditionell GLM. Modellen kan användas på tre olika sätt: Först och främst kan den användas för att skapa en tabell med statistiskt signifikanta interaktioner att addera till en existerande GLM. Utöver detta kan den iterativt och automatiskt lägga till sådana interaktioner tills inga fler återstår. Slutligen kan modellen också användas för att skapa en helt ny GLM från grunden, utan en existerande prissättningsmodell. Metoderna testas på två motorförsäkringsdataset från ett nordiskt skadeförsäkringsbolag och resultaten visar att alla överträffar originalregressionen. Även om de två iterativa metoderna överträffar den första metoden rekommenderas försäkringsbolag att använda den första metoden. Detta eftersom den resulterar i en rimlig avvägning mellan att automatisera processer och att nyttja aktuariers omdömesförmåga.

O presente trabalho tem por objetivo utilizar a abordagem dos modelos lineares generalizados e os modelos de dispersão no contexto do seguro agrícola. Os modelos lineares generalizados (MLG\'s) constituem uma extensão dos modelos lineares de regressão múltipla introduzida por Nelder e Wedderburn (1972), que inclui modelos cuja variável resposta pertence à família exponencial de distribuições. O MLG é formado por um componente aleatório, que possui distribuição pertencente à família exponencial, um componente sistemático, conectados por uma função de ligação. Jorgensen (1997) estende a utilização dos MLG para uma classe mais ampla de modelos probabilísticos, denominados modelos de dispersão. A estimação dos parâmetros foi baseada no método da máxima verossimilhança, e também, em função da amostra ser relativamente pequena, optou-se pelo método de bootstrap não-paramétrico. As duas abordagens foram aplicadas a dois conjuntos de dados de sinistros de 15 municípios do estado do Rio Grande do Sul. Os resultados mostraram que a precipitação acumulada tem influência na ocorrência de sinistros. Entretanto, na modelagem do montante do sinistro não foi encontrada nenhuma variável significativa. Usando o método de bootstrap, foi encontrada influência das variáveis precipitação acumulada e a temperatura média no numero de sinistros / The main objective of this work is to use the generalized linear models and dispersion models in the agricultural insurance context. The Generalized Linear Model (GLM) are an extension of the multiple regression linear models presented by Nelder e Wedderburn (1972). This approach include situations in which the response variable can be included in exponencial the family. The GLM is composed of a randomized component, a sistematic component and the link functions. JÁrgensen (1997) extend the application of the GLM for a more general class of probability models, called dispersion models. Both approaches were applied in two insurance datasets for 15 citys in Rio Grande do Sul. The parameters estimation was based in the maximum likelihood method, in addition, because of the relatively small sample, the non-parametric Bootstrap method was used. This study show, using GLM, that only the accumulated rainfall was statistically significant . However, any of the covariates was significant when modelling the amount of claims. In the analysis using Bootstrap method the accumulated rainfall and average temperature were significant when modelling the number of insurance clains.

O presente trabalho tem por objetivo utilizar a abordagem dos modelos lineares generalizados e os modelos de dispersão no contexto do seguro agrícola. Os modelos lineares generalizados (MLG\'s) constituem uma extensão dos modelos lineares de regressão múltipla introduzida por Nelder e Wedderburn (1972), que inclui modelos cuja variável resposta pertence à família exponencial de distribuições. O MLG é formado por um componente aleatório, que possui distribuição pertencente à família exponencial, um componente sistemático, conectados por uma função de ligação. Jorgensen (1997) estende a utilização dos MLG para uma classe mais ampla de modelos probabilísticos, denominados modelos de dispersão. A estimação dos parâmetros foi baseada no método da máxima verossimilhança, e também, em função da amostra ser relativamente pequena, optou-se pelo método de bootstrap não-paramétrico. As duas abordagens foram aplicadas a dois conjuntos de dados de sinistros de 15 municípios do estado do Rio Grande do Sul. Os resultados mostraram que a precipitação acumulada tem influência na ocorrência de sinistros. Entretanto, na modelagem do montante do sinistro não foi encontrada nenhuma variável significativa. Usando o método de bootstrap, foi encontrada influência das variáveis precipitação acumulada e a temperatura média no numero de sinistros / The main objective of this work is to use the generalized linear models and dispersion models in the agricultural insurance context. The Generalized Linear Model (GLM) are an extension of the multiple regression linear models presented by Nelder e Wedderburn (1972). This approach include situations in which the response variable can be included in exponencial the family. The GLM is composed of a randomized component, a sistematic component and the link functions. JÁrgensen (1997) extend the application of the GLM for a more general class of probability models, called dispersion models. Both approaches were applied in two insurance datasets for 15 citys in Rio Grande do Sul. The parameters estimation was based in the maximum likelihood method, in addition, because of the relatively small sample, the non-parametric Bootstrap method was used. This study show, using GLM, that only the accumulated rainfall was statistically significant . However, any of the covariates was significant when modelling the amount of claims. In the analysis using Bootstrap method the accumulated rainfall and average temperature were significant when modelling the number of insurance clains.

No description available.

非負値行列因子分解(NMF)は，全要素が非負であるデータ行列に対する行列分解法である．本論文では，実在するデータ行列に頻繁に見られる特徴や解釈容易性の向上を考慮に入れ，探索的にデータ分析を行うためのNMFの拡張について論じている．具体的には，零過剰行列や外れ値を含む行列を扱うための確率分布やダイバージェンス，さらには分解結果である因子行列の数や因子行列への直交制約について述べている． / Nonnegative matrix factorization (NMF) is a matrix decomposition technique to analyze nonnegative data matrices, which are matrices of which all elements are nonnegative. In this thesis, we discuss extensions of NMF for exploratory data analysis considering common features of a real nonnegative data matrix and an easy interpretation. In particular, we discuss probability distributions and divergences for zero-inflated data matrix and data matrix with outliers, two-factor vs. three-factor, and orthogonal constraint to factor matrices. / 博士(文化情報学) / Doctor of Culture and Information Science / 同志社大学 / Doshisha University

Face checking in decorative maple veneered plywood panels is a significant problem for hardwood plywood manufacturers, furniture makers, cabinetmakers, and consumers. Efforts made by panel producers and researchers to minimize checking conducted to-­‐date have been limited, and produced contradictory results. In this study the impact of four manufacturing factors believed to contribute to check development in decorative maple veneer panels were determined. The factors investigated were face veneer thickness and preparation, lathe-­‐check orientation, adhesive and core type. An efficient, automated, optical technique based on digital image correlation principles was developed and used to detect and measure checks as they develop. The novel new method for characterizing check severity and development was effective in efficiently measuring checking for a substantial number of samples. The results of the factor screening analysis reveal intricate four way interactions between factor levels contribute to check development, and that some combinations are likely to exhibit much more checking than others. / Graduation date: 2013