Search code examples
wekanaivebayes

Interpret Weka naive Bayes output

Hi I need to understand the following output of Naive Bayes results. Can somone explain me this taking a small example out of the figure

ive Bayes output

Solution

  • There are 50 out of the 150 instances in each class, so under the class name it says (.33), which is one-third.

    Mean: This is the average within each class. For example, for Iris-setosa the average sepallength is 4.9913, for Iris-versicolor the average sepallength is 5.9379.

    std. dev.: The standard deviation of sepallength within Iris-setosa is 0.355.

    weight sum: This is the number of cases in the class (50 in each), since in your analysis each instance has a weight of 1, and there are 50 instances in each class.

    precision: This is a bit mysterious. According to http://weka.8497.n7.nabble.com/naive-bayes-output-interpretation-td28658.html , "The precision is the minimum standard deviation allowed for the attribute in question. It's obtained by a heuristic implemented in naïve Bayes that computes the average difference between adjacent values of the attribute."