Context
In today's business landscape, many companies especially in the insurance and finance sectors grapple with the complexities of dealing with large volumes of data. This can often strain their ability to balance computational efficiency and time management. Our mission is to simplify their operations by improving data models that handle big data more efficiently, reducing the time spent on computations. As well as offering estimators that are efficient in their asymptotic performance.
My research focus is on enhancing these models, making them more relevant and effective. Ultimately, our goal is to help companies tackle a wide range of challenges more easily and efficiently.
Methodology
In the context of my thesis research, we employ Le Cam's One-step estimation procedure as a fundamental approach. This procedure is applied to categorical and numerical variables separately, with an additional focus on investigating interactions within the categorical variable domain.
As we progress to the third segment of my research, we delve into the realm of copula models. Copulas are designed to characterize multivariate uniform distributions, offering an analytical tool for assessing the association or dependence among multiple variables. Copulas are particularly advantageous in scrutinizing the interdependence of returns from two or more assets within a portfolio. Notably, these models offer an alternative to traditional correlation-based methods, as financial market distributions often deviate from the normal distribution assumption, necessitating a more nuanced approach.
Concluding our research endeavor, we complement our findings with a comprehensive analysis that includes Monte Carlo simulations. Additionally, our study is enriched with empirical results derived from authentic data generously supplied by the Covéa insurance company.