الفهرس 
قوائم البحثيوجد فقط 14 صفحة متاحة للعرض العام
 |  | from | 271 |  |  |
| | | from | 271 | | |
المستخلص
Forecasting traditional approaches are commonly used to estimate loss ratio. These traditional approaches may be simple models, such as linear regression models or ad-hoc models such as autoregressive integrated moving-average (ARIMA) models. These approaches have many difficulties: (1) They do not work well when applying on small time series data sets, and (2) they give inaccurate results when the data is unstable. Loss ratio of Petroleum insurance branch has a special nature such as fluctuated, unstable, and has a limited number of observation which leads to inaccurate estimate.
Thus, the main objective of this study is to introduce forecasting approaches based on the fuzzy sets theory: fuzzy regression models, fuzzy autoregressive regression, and fuzzy time series to estimate loss ratio of Petroleum branch loss ratio using 15 observations only from year 1995 to 2009. This data set has been used to predict the loss ratio for the Petroleum insurance of the five years afterwards, from 2010 to 2014.
The estimation results of both traditional approaches and those based on fuzzy sets are compared in terms of Mean Absolute Percentage Error (MAPE). The results of study show that the fuzzy time series gives better estimates if data has outliers or not. Accurate estimates of Fuzzy Time Series approach is due to the ability of this approach to deal with the random walk and unexpected structure, which is the case in loss ratio of Petroleum insurance field. The traditional approaches work well only if outliers affects are isolated.
Finally, the estimation approaches based on Fuzzy Sets Theory are recommended to estimate the loss ratio especially that has fluctuated and unstable nature. Also, it is recommended to use this kind of approaches in other fields of insurance industry, such as pricing, technical provisions and investment.