With the boost of vehicle industry in China, automobile insurance has become an important pillar of China's insurance industry. However, this industry was once hindered by the governments excessive regulations and the cut-throat competitions in the market, which turns the scientific ratemaking problem into an urgent issue. How to set a reasonable premium which can win the customers and make profits at the same time is crucial to the business of automobile insurance.

In this thesis, with real data from an insurance company in China, we develop a risk-based pricing system called R-Pricing for ratemaking of the automobile insurance. This system mainly apply the theory of generalized linear models, which is one of the most popular statistical models used in automobile insurance pri...[ Read more ]

With the boost of vehicle industry in China, automobile insurance has become an important pillar of China's insurance industry. However, this industry was once hindered by the governments excessive regulations and the cut-throat competitions in the market, which turns the scientific ratemaking problem into an urgent issue. How to set a reasonable premium which can win the customers and make profits at the same time is crucial to the business of automobile insurance.

In this thesis, with real data from an insurance company in China, we develop a risk-based pricing system called R-Pricing for ratemaking of the automobile insurance. This system mainly apply the theory of generalized linear models, which is one of the most popular statistical models used in automobile insurance pricing.

We will briefly introduce the theory of GLMs and then focus on the design of our system R-Pricing. Finally, we evaluate our system from both the model angle and the business angle. And the results demonstrate that R-Pricing has a good performance close to a commercial pricing system.