0

here is a vector: [0,0,0,1,0,0,1,4,2,5,0,0,0,0,0,0,0,0,0,....,2,0] and is there any statistic test and p value to tell me such vector is holding too many zero ? I just check https://www.statsmodels.org/stable/api.html but they didn't fit my requirement. I am not statistician but I prefer to use statistic test, because of academic support, implemented by Python so as to find p value to convince my advisor that vector is facing zero inflated issue

Thank you so much

Sincerely

KnowNothing JohnSnow
  • 11
  • 1
    Does this answer your question? Diagnostic plots for count regression I'm assuming that you want to evaluate zero inflation against a Poisson distribution. In particular, see Point 5 of this answer. – EdM Jul 25 '22 at 16:49
  • Thanks ... But previously I have checked a lot of answers that are similar with your link. Generally, those zero inflation issues are designed for Y ~ X1 + X2 + ... Xn, but I would like to inspect X1, X2 or Xn singgally. Here is the tutorial I found at YouTubelink. Is there any R code or Python code ? – KnowNothing JohnSnow Jul 26 '22 at 04:50
  • by the way.. Is this link is suitable to answer my question ??? How can I convince my boss to apply it ?? Thanks – KnowNothing JohnSnow Jul 26 '22 at 05:05
  • Yes, the page you link to in the second comment is fine (more specific to your question than the one I suggested) and provides a simple test for evaluating zero inflation versus Poisson for individual variables. Be careful: unless you know that each Xi is supposed to be Poisson-distributed, there might be nothing wrong with any zero inflation that you find. Sometimes real count data just have a lot of 0 values. So unless your boss accepts that the values should be Poisson distributed, it will be hard to convince. – EdM Jul 26 '22 at 13:23
  • Thank you very much~~ My problem has been resolved – KnowNothing JohnSnow Aug 29 '22 at 17:14

0 Answers0