Naive Bayes Classification Iteration

Hello,

I am using the Naive Bayes Classifier on a binary image dataset. The classifier seems to return different sets of data on the first and second runs. If I clear all the data and run it, I get one set of results, and if (without clearing), I run it a second time on the same set of data, I get a different (mostly overlapping) set of results.

Is there an explanation for this? Is the classifier using the information from the first time during the second run?