Predicting electromagnetic environment parameters of power transmission line with data mining methods

Generally, the electromagnetic environment (EME) parameters are always estimated by empirical equations. This paper collects electromagnetic environment parameters, together with power transmission line related variables where principal factor analysis and regression analysis are conducted using information gain and decision tree for regression, respectively. Statistical method is explored to obtain two linear independent EME parameters to be analyzed. Principal factor to both parameters are obtained using information gain approach. As a result, radio interference and nominal field strength share the same principal factor, i.e., the number of bundled conductors. Decision tree for regression models, which achieves low mean square error in this case, for both of the linear independent variables are depicted to better illustrate the rule behind the EME factors.