Dealing with classification problems, many researchers provide mean +/- std values for feature with respective classes. What actually the physical essence of those values? What exact inference we can draw from that? By looking those values can we say any feature information relate to classification?