Integrantes

#"Maricely Zelada Alvarado", "Rodrigo Mendivil Rodriguez", "Liz Hinostroza Vidal", "Piero Morales Romero"

Reading Data

data2  <- data.frame(ÁrbolC =arbol, Red_N = redn, Regresion = regresion, SActual = actual)
data2
##    ÁrbolC Red_N Regresion SActual
## 1   23.81 23.24     16.13   17.09
## 2   22.13 20.08     17.84   15.77
## 3   22.64 18.01     18.28   18.45
## 4   21.69 23.28     15.61   16.55
## 5   23.58 19.23     17.62   22.23
## 6   22.14 21.22     16.12   22.11
## 7   18.73 21.47     17.29   18.26
## 8   21.59 20.60     16.13   18.04
## 9   20.36 21.11     16.64   19.66
## 10  20.53 21.27     15.03   19.76
## 11  20.11 21.03     18.16   18.74
## 12  20.34 17.34     16.82   19.02
## 13  19.19 22.80     17.44   18.54
## 14  22.92 21.85     16.76   16.70
## 15  18.65 17.85     17.26   17.57
## 16  20.60 23.15     15.55   19.89
## 17  19.83 19.57     17.49   19.06
## 18  20.09 19.56     18.42   18.70
## 19  19.43 20.79     17.54   19.39
## 20  22.06 18.04     17.13   19.68
## 21  21.15 20.95     15.50   19.20
## 22  19.26 21.83     16.80   16.85
## 23  18.08 18.17     18.47   19.91
## 24  20.24 22.66     18.42   19.82
## 25  18.75 18.29     18.43   18.08
## 26  20.69 18.89     15.56   19.38
## 27  21.62 19.49     16.03   20.30
## 28  23.69 19.19     15.39   21.60
## 29  23.93 26.47     15.12   23.39
## 30  23.19 25.25     17.77   19.33

Modelo actual

summary(data2$SActual)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   15.77   18.12   19.13   19.10   19.80   23.39
hist(data2$SActual)

cv_SActual <- sd(data2$SActual)/mean(data2$SActual)
cv_SActual
## [1] 0.0905464

Modelos propuestos

Árbol de clasificación

summary(data2$ÁrbolC)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   18.08   19.89   20.64   21.03   22.14   23.93
hist(data2$ÁrbolC)

cv_ÁrbolC <- sd(data2$ÁrbolC)/mean(data2$ÁrbolC)
cv_ÁrbolC
## [1] 0.08108817
Prop1 <- t.test(data2$SActual, data2$ÁrbolC, conf.level = 0.95)
Prop1
## 
##  Welch Two Sample t-test
## 
## data:  data2$SActual and data2$ÁrbolC
## t = -4.3555, df = 57.989, p-value = 5.479e-05
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -2.819433 -1.043900
## sample estimates:
## mean of x mean of y 
##  19.10233  21.03400

Redes Neuronales

summary(data2$Red_N)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   17.34   19.20   20.87   20.76   21.84   26.47
hist(data2$Red_N)

cv_Red_N <- sd(data2$Red_N)/mean(data2$Red_N)
cv_Red_N
## [1] 0.1071663
Prop2 <- t.test(data2$SActual, data2$Red_N, conf.level = 0.95)
Prop2
## 
##  Welch Two Sample t-test
## 
## data:  data2$SActual and data2$Red_N
## t = -3.2145, df = 54.681, p-value = 0.002194
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -2.684758 -0.622575
## sample estimates:
## mean of x mean of y 
##  19.10233  20.75600

Métodos de regresión

summary(data2$Regresion)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   15.03   16.05   16.98   16.89   17.73   18.47
hist(data2$Regresion)

cv_Regresion <- sd(data2$Regresion)/mean(data2$Regresion)
cv_Regresion
## [1] 0.06531693
Prop3 <- t.test(data2$SActual, data2$Regresion, conf.level = 0.95)
Prop3
## 
##  Welch Two Sample t-test
## 
## data:  data2$SActual and data2$Regresion
## t = 5.902, df = 49.248, p-value = 3.26e-07
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  1.458045 2.963288
## sample estimates:
## mean of x mean of y 
##  19.10233  16.89167

Conclusión

Se implementará el Metodo de regresión por lo siguiente:
p-value < 0.05
IC 95% = (1.46 , 2.96)
Media del Método Actual = 19.10
Media del Método Regresión = 16.89
El método de regresión reduce el tiempo promedio de fallas en aproximadamente 2.21 minutos, respecto al sistema actual, lo que representa una mejora cercana al 11.6%.

Nota: Las alternativas de Árbol de clasificación y Redes neuronales presentan tiempos de fallas significativamente mayores que el sistema actual, por lo que no se recomienda su implementación.
```