APPLICATION OF POISSON, ZERO-INFLATED POISSON, AND NEGATIVE BINOMIAL MODEL TO COUNT DATA: CASE OF THE RECREATIONAL BOATING TRIPS TO LAKE SOMERVILLE, TEXAS
Recreation is a necessary activity in this modern lifestyle. The number of recreational trips made could be a benchmark to develop a recreational area. The number of recreational trips is an example of a count data which could be modeled by a Poisson model. However, a high frequency of zeros in a...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/42414 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Recreation is a necessary activity in this modern lifestyle. The number of recreational
trips made could be a benchmark to develop a recreational area. The number
of recreational trips is an example of a count data which could be modeled by a
Poisson model. However, a high frequency of zeros in a data set could cause an
overdispersion which could not be modeled by a Poisson model. Often, this data
set is considered to have excess zeros and a zero-inflated model is usually applied.
Another probability model which could be used to model an over-dispersed data is
a Negative Binomial distribution model. In this Final Project, a Poisson, a zeroinflated
Poisson, and a Negative Binomial model are applied to model the recreational
boating trips to Lake Somerville, Texas, United States of America. The
estimation of the model parameters is carried out using the maximum likelihood
estimation method. After comparing the three models using the AIC value, the BIC
value, and the Vuong test, it was found that the Negative Binomial regression model
is more appropriate compared to the other models, with the regression model be:
ln(^) = ????1; 1941 + 0; 7289 quality + 0; 5971 ski + 0; 0520 costC????0; 0969 costS +
0; 0390 costH. Thus, the factors which affect the number of recreational boating
trips to Lake Somerville are: the assessments by the visitors on the quality of the
Lake Somerville0s facilities; the visitors0 experience waterskiing at Lake Somerville;
and the expenditure spent when visiting Lake Somerville and the other alternative
lakes (Lake Conroe and Lake Houston). These results showed that although there
is a high frequency of zeros in the data, the data does not have to be modeled by a
zero-inflated model. |
|---|