LGPIF 留学生R代码代编程代写

日期：2018-04-25 10:45

I. Introduction (what is the project about?)

The Wisconsin Local Government Property Insurance Fund (LGPIF) is a fund to make property insurance available for local government units, such as counties, cities, towns, villages, school districts, and library boards, etc.

In this project, we examine the insurance coverage for Wisconsin building and contents in previous years within LGPIF and develop our own aggregate loss models by using individual risk approach, collective risk approach, and convolution simulation.

II. Problem Description (why is the project important and interesting?)

The LGPIF is a unique fund in Wisconsin. It is designed to moderate the budget effects of uncertain insurable events for local government entities, in a sense, to separate budgetary responsibilities for the Wisconsin government. We are interested in how well the fund has operated in previous years and prospect its future performances. We are also wondering if this fund is meaningful and should be applied to more states.

III. Objectives (what are the goals of the project?)

Our goal is to develop our own aggregate loss models for LGPIF based on the existing data by using individual risk approach, collective risk approach, and convolution simulation. By tracking the insurance coverage from Year 2006 to Year 2011, we wish to predict the insurance coverage that is likely to happen after Year 2012 by using our aggregate loss models.

IV. Expected Results (what will be the expected outcome of your project?)

We are guessing that since there were no mega-events (in sense of water/fire disaster) happen in Wisconsin from Year 2006 to Year 2011, there was no distinct coverage change during that period. But due to inflation and price increases, the annual coverage amount will be in a slightly increasing behavior. We hope to calculate that increasing rate and build the corresponding aggregate loss models that can work for future.

V. Approach/Methodology (how are you going to conduct the project?)

We will divide our original data set into 2 subsets, one for analysis and one for test.

When analyze the original data, we will first show the distribution of each component by plotting histograms and empirical cdf. Then we will demonstrate the two-part framework to construct each component (water claims, fire claims and other perils).

By using the distributions that we found, we can develop our individual risk model (record losses for each contract and then add them up), as well as the collective risk model (record losses as claims are made and then add them up). After that, we apply the convolution simulation to evaluate the distribution of both risk models.

Lastly, we will then examine the models on test data and see which one is a better fit.