The T-Statistic is the significance or amount of influence of a particular independent variable within a meter baseline model.  The higher the value, the stronger the indication that the independent variable really does have an impact on the utility usage.  Normally the energy analyst looks for a T-statistic greater than 2.0 as an indication of relevance.  If the value is less than 2.0, he/she would conclude that the independent variable has no effect on the meter and so would not select that variable to be included in the baseline model.



def_T-Statistic         ©2013 Managing Energy Inc.