powered by
confidence interval for a point (using log transformation and geometric mean)
iconfianza.logx(datos, nivel = 0.95, ic = T, colas = 2)