Changes the usage of numerical_limits<double>::min() to lowest() in cases where negatives values data may be permitted. In this case using min() to initialise the max value may lead to the max value being 0.00 rather than the maximum negative value of the data.