Random Forest Accuracy

OsamaAnmar

Why the accuracy of random forest (majority voting) different between each experiment Although I use the same parameters in each experiment (Knn imputing, Binning,  Forward Selection) 

first experiment: 91.97

second experiment : 92.03

third experiment : 91.85

hughesfleming68

Most learning operators use randomness in some form. You can get repeatable results by ticking the local random seed box. Is this a good idea? There are pros and cons. A valid alternative is to take the average of multiple runs but in many cases setting the random seed is the best way to go especially if other people need to reproduce your results.