Register | Guestbook | Forums | Downloads | 1 active visitor 
MixFit Home MixFit Home
If it fits, use it!


 MENU
 :: As-If ( A Simulation Intructional Facility)
 :: MixFit Home
 

 Login
   
    
Register
Forgot password?
 

 Downloads
Stuart Klugman's FIT program (288)
The programs FIT and DFIT find maximum likelihood estimates for 20 continuous and 16 discrete distributions respectively. They also produce estimates of the covariance matrix of the estimators as well as graphs that illustrate the quality of the fit. The chi-square goodness-of-fit test is also performed. The program CR evaluates probabilities for the collective risk model. Frequencies used can be any of the models used in DFIT while severities can be any of the models from FIT or a model entered by the user. In all cases the probabilities are obtained using recursive formulas. Prior to the recursion, individual or aggregate deductibles, limits, or coinsurances may be applied.

Toward a unified approach to fitting loss models (476)

This is a paper being submitted for publication in the NAAJ. It includes the formulas and models implemented in the MixFit application also available for download here.

 

Abstract

There are two components to fitting models: selecting a set of candidate distributions and then determining which member fits best. It is important to have the candidate set be small to avoid overfitting. Finite mixture models using a small number of base distributions pro-vide an ideal set. Because actuaries fit models for a variety of situ-ations, particularly with regard to data modifications, it is useful to have a single approach. Though not optimal or exact for a particular model or data structure, the method should be reasonable for most all settings. Such a method is proposed in this article. To aid the user, a computer program implementing these models and techniques is provided.

 

MixFit (272)

This is the installation package for MixFit 1.0.0

Please take note of the requirements:

Get MixFit.msi here:

Copyright 2002 © MixFit website - all rights reserved