The improvement of energy efficiency in wireless cellular networks has become a popular topic in recent years due to its positive effects on the environment and economical benefits for network operators. An effective way to improve energy efficiency is to deploy energy-efficient base stations and turn as many base stations off as possible. In this paper, we consider an LTE network composed by the Light Radio Cubes, and through data mining, we find that a significant proportion of power is saved if two conditions are considered jointly. The optimal parameters involved in the conditions are derived by applying a Genetic Algorithm (GA).