Citas bibligráficas
Villar, G., (2023). Modelos de elección discreta logit y datos sintéticos generados con el algoritmo rose aplicados a valorar un bien público [Universidad Nacional Agraria La Molina]. https://hdl.handle.net/20.500.12996/5959
Villar, G., Modelos de elección discreta logit y datos sintéticos generados con el algoritmo rose aplicados a valorar un bien público []. PE: Universidad Nacional Agraria La Molina; 2023. https://hdl.handle.net/20.500.12996/5959
@mastersthesis{renati/241992,
title = "Modelos de elección discreta logit y datos sintéticos generados con el algoritmo rose aplicados a valorar un bien público",
author = "Villar Morales, Gesabel",
publisher = "Universidad Nacional Agraria La Molina",
year = "2023"
}
The process of estimating the economic value of a public good, such as environmental services or national defense, is based on welfare theory. One of the best-known methods of economic valuation is called contingent valuation. In their application, respondents answer a question about their willingness to pay (WTP), which reflects the maximum amount of money an individual would pay to obtain a public good. This value is estimated using a binary logistic regression model. However, this technique has a serious limitation related to the possibility of hypothetical bias, due to the lack of honesty in the answers, which produces an imbalance in the observations of the groups defined by the dependent variable that indicates the response to the proposal to make a payment for access to a public good. This imbalance causes problems in the estimation and evaluation stages of the accuracy of the classification model. This study used contingent valuation data from the Reserved Forest of the Universidad Nacional Agraria de la Selva (BRUNAS), located 1.5 km from the town of Tingo María in Huánuco, for which WTP was calculated using different models, with the aim of valuing a public good using binary logistic regression models estimated with balanced groups using the ROSE algorithm. In the first model, a variable selection method using the Akaike Information Criterion (AIC) was applied, considering the original data set with unbalanced groups. The second model was estimated after applying the ROSE algorithm, which allows synthetic data to balance the groups and has approximately the same number of negative and positive responses. After applying the ROSE algorithm, the obtained model achieved a more realistic estimate of the DAP and its standard error, resulting in confidence intervals with less amplitude than the initial model.
Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons