I am dealing with a very hard-to-work data set: fish larval density. It is a semicontinuous data, with 90% of zeros and a right-skewed distribution, with few very large values.

One problem is that my zeros have two sources: true and false, or structural and sampling.

I saw models that can deal with these two types of zeros, called Zero inflated models, but only for count data.

I didn't see models for semicontinuous data that can deal with false zeros. These models are two-part models with log-normal or gamma distribution, and all of them assume true zeros. I am searching for some model that can deal with 2 sources of zero.

Here is a part of my data set: My response is fish larval density (Dprochilodus). And I have a lot of zeros, but I can't assume that all of them are trully zeros. What I want to tell is: when I have zero, it doesn't mean there were no larvae. Can be because I could't sample/detect them. Here there are some explanations about the different models for true and false zeros

ponto            Dprochilodus      periodo    Dif_his.y            temp      
 Length:574         Min.   :  0.0000   I :278   Min.   :0.01571   Min.   :21.00  
 Class :character   1st Qu.:  0.0000   II:296   1st Qu.:0.09738   1st Qu.:24.00  
 Mode  :character   Median :  0.0000            Median :0.22505   Median :27.00  
                    Mean   :  0.5949            Mean   :0.25405   Mean   :26.87  
                    3rd Qu.:  0.0000            3rd Qu.:0.40296   3rd Qu.:29.01  
                    Max.   :134.7469            Max.   :0.89746   Max.   :33.00  


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.