Regressão múltipla

In [2]:

    

df <- read.csv('servidor.csv')
head(df)
str(df)


    

    






Duracao_media_ms
Perc_medio_CPU
Load_avg_minute
Requests_média

	
150      
10.580584
0.6547530
 72      
	
140      
 9.957563
0.5304159
120      
	
150      
10.718390
0.7739702
200      
	
155      
10.975842
0.5812908
 85      
	
140      
 9.824099
0.6504205
 56      
	
150      
10.552237
0.5846244
230      









    




'data.frame':	29 obs. of  4 variables:
 $ Duracao_media_ms: int  150 140 150 155 140 150 145 150 151 150 ...
 $ Perc_medio_CPU  : num  10.58 9.96 10.72 10.98 9.82 ...
 $ Load_avg_minute : num  0.655 0.53 0.774 0.581 0.65 ...
 $ Requests_média  : int  72 120 200 85 56 230 620 560 530 525 ...

In [3]:

    

summary(df)


    

    






 Duracao_media_ms Perc_medio_CPU    Load_avg_minute  Requests_média  
 Min.   : 140.0   Min.   :  6.867   Min.   :0.4983   Min.   :  56.0  
 1st Qu.: 150.0   1st Qu.: 10.582   1st Qu.:0.6504   1st Qu.: 230.0  
 Median : 520.0   Median : 20.151   Median :2.3333   Median : 720.0  
 Mean   : 811.6   Mean   : 47.975   Mean   :3.9274   Mean   : 631.8  
 3rd Qu.:1500.0   3rd Qu.: 98.171   3rd Qu.:7.3177   3rd Qu.: 890.0  
 Max.   :1950.0   Max.   :121.236   Max.   :9.5802   Max.   :1535.0  

In [4]:

    

model <- lm(df$Duracao_media_ms ~ df$Perc_medio_CPU + df$Load_avg_minute + df$Requests_média)
summary(model)


    

    






Call:
lm(formula = df$Duracao_media_ms ~ df$Perc_medio_CPU + df$Load_avg_minute + 
    df$Requests_média)

Residuals:
    Min      1Q  Median      3Q     Max 
-28.598 -12.485  -0.115  10.901  32.169 

Coefficients:
                    Estimate Std. Error t value Pr(>|t|)    
(Intercept)        2.079e+01  6.316e+00   3.292  0.00296 ** 
df$Perc_medio_CPU  2.873e-01  2.847e-01   1.009  0.32262    
df$Load_avg_minute 1.944e+02  3.828e+00  50.799  < 2e-16 ***
df$Requests_média  2.119e-02  9.605e-03   2.206  0.03681 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 15.5 on 25 degrees of freedom
Multiple R-squared:  0.9995,	Adjusted R-squared:  0.9995 
F-statistic: 1.755e+04 on 3 and 25 DF,  p-value: < 2.2e-16

In [5]:

    

plot(model)


    

In [6]:

    

library(car)


    

In [7]:

    

vif(model)


    

    






	
df$Perc_medio_CPU
		
17.7018743552122
	
df$Load_avg_minute
		
19.506189738464
	
df$Requests_média
		
1.64510975241899

In [8]:

    

ncvTest(model)


    

    






Non-constant Variance Score Test 
Variance formula: ~ fitted.values 
Chisquare = 0.8242639    Df = 1     p = 0.3639365 

In [9]:

    

durbinWatsonTest(model)


    

    






 lag Autocorrelation D-W Statistic p-value
   1      -0.2504768      2.479978   0.354
 Alternative hypothesis: rho != 0

In [ ]: