Robust Estimates for One-Parameter Exponential Regression Model

One-parameter exponential regression is one of the most common and widely used models in several fields, to estimate parameters one-parameter model use ordinary least square method but this not effective presence outlier values, so robust methods were treat values are using (Median-of-Means, Forward search, M-Estimation), simulation was compare between estimation with different sample sizes assuming four ratios from outliers data (10%, 20%, 30%, 40%). And mean error (MSE) made reach best for parameters, where results obtained showed that forward search because it gives lowest error. On practical side, expenditure revenue regression, tested, appeared have an distribution, boxplot (COOK) test detect present real data. The Goodness fit model, found did follow normal suffers problem heterogeneity variance. estimated advanced estimate.
 Paper type Research paper