Analise de Regressao Linear Simples

Matheus Henrique de Morais

29/04/2022

Leitura dos dados

Grafico de dispersao - Letra A

Compensacoes Pagas por Volume de Chuva

##
## Call:
## lm(formula = Compensacoes.Pagas ~ Volume.chuva, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -148659 -107542 -51844 27486 1996466
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 159332.647 66992.331 2.378 0.0181 *
## Volume.chuva -4.264 137.159 -0.031 0.9752
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 214000 on 264 degrees of freedom
## Multiple R-squared: 3.661e-06, Adjusted R-squared: -0.003784
## F-statistic: 0.0009664 on 1 and 264 DF, p-value: 0.9752

## 2.5 % 97.5 %
## (Intercept) 27425.3834 291239.9106
## Volume.chuva -274.3279 265.8002
Compensacoes Pagas por Descargas

##
## Call:
## lm(formula = Compensacoes.Pagas ~ Descargas.atm, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -354538 -96974 -49050 28626 2002681
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.277e+05 1.461e+04 8.742 2.73e-16 ***
## Descargas.atm 1.733e+01 4.202e+00 4.123 5.01e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 207400 on 264 degrees of freedom
## Multiple R-squared: 0.0605, Adjusted R-squared: 0.05694
## F-statistic: 17 on 1 and 264 DF, p-value: 5.01e-05
## 2.5 % 97.5 %
## (Intercept) 98919.55323 156436.20855
## Descargas.atm 9.05254 25.60157
Compensacoes Pagas por Vento

## function (nx = NULL, ny = nx, col = "lightgray", lty = "dotted",

## lwd = par("lwd"), equilogs = TRUE)
## {
## if (is.null(nx) || (!is.na(nx) && nx >= 1)) {
## log <- par("xlog")
## if (is.null(nx)) {
## ax <- par("xaxp")
## if (log && equilogs && ax[3L] > 0)
## ax[3L] <- 1
## at <- axTicks(1, axp = ax, log = log)
## }
## else {
## U <- par("usr")
## at <- seq.int(U[1L], U[2L], length.out = nx + 1)
## at <- (if (log)
## 10^at
## else at)[-c(1, nx + 1)]
## }
## abline(v = at, col = col, lty = lty, lwd = lwd)
## }
## if (is.null(ny) || (!is.na(ny) && ny >= 1)) {
## log <- par("ylog")
## if (is.null(ny)) {
## ax <- par("yaxp")
## if (log && equilogs && ax[3L] > 0)
## ax[3L] <- 1
## at <- axTicks(2, axp = ax, log = log)
## }
## else {
## U <- par("usr")
## at <- seq.int(U[3L], U[4L], length.out = ny + 1)
## at <- (if (log)
## 10^at
## else at)[-c(1, ny + 1)]
## }
## abline(h = at, col = col, lty = lty, lwd = lwd)
## }
## }
## <bytecode: 0x000000002218a4a0>
## <environment: namespace:graphics>

##
## Call:
## lm(formula = Compensacoes.Pagas ~ Vento, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -148472 -106666 -51397 27609 1997101
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 152828 54024 2.829 0.00503 **
## Vento 3067 36018 0.085 0.93221
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 214000 on 264 degrees of freedom
## Multiple R-squared: 2.746e-05, Adjusted R-squared: -0.00376
## F-statistic: 0.007251 on 1 and 264 DF, p-value: 0.9322
## 2.5 % 97.5 %
## (Intercept) 46455.17 259200.62
## Vento -67851.26 73985.18
Compensacoes Pagas por Umidade

##
## Call:
## lm(formula = Compensacoes.Pagas ~ umidade, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -161307 -102667 -53135 27862 1988773
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 340668 211638 1.610 0.109
## umidade -2611 3008 -0.868 0.386
##
## Residual standard error: 213700 on 264 degrees of freedom
## Multiple R-squared: 0.002847, Adjusted R-squared: -0.0009305
## F-statistic: 0.7537 on 1 and 264 DF, p-value: 0.3861
## 2.5 % 97.5 %
## (Intercept) -76045.853 757382.155
## umidade -8532.519 3310.781
Compensacoes Pagas por Temperatura

##
## Call:
## lm(formula = Compensacoes.Pagas ~ temperatura, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -183951 -102739 -48401 29266 1963233
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -292415 286209 -1.022 0.308
## temperatura 18634 11847 1.573 0.117
##
## Residual standard error: 213000 on 264 degrees of freedom
## Multiple R-squared: 0.009284, Adjusted R-squared: 0.005531
## F-statistic: 2.474 on 1 and 264 DF, p-value: 0.1169
## 2.5 % 97.5 %
## (Intercept) -855957.985 271127.21
## temperatura -4692.644 41960.86
Compensacoes Pagas por AG1

##
## Call:
## lm(formula = Compensacoes.Pagas ~ AG1, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -259791 -80581 -42875 31124 1955976
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 157121 12196 12.883 < 2e-16 ***
## AG1 78681 12203 6.448 5.38e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 198900 on 264 degrees of freedom
## Multiple R-squared: 0.136, Adjusted R-squared: 0.1328
## F-statistic: 41.57 on 1 and 264 DF, p-value: 5.384e-10
## 2.5 % 97.5 %
## (Intercept) 133107.73 181134.5
## AG1 54652.86 102709.3
Compensacoes Pagas por AE1

##
## Call:
## lm(formula = Compensacoes.Pagas ~ AE1, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -363080 -68334 -20771 19382 1890329
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 156727 11418 13.726 <2e-16 ***
## AE1 105430 11462 9.198 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 186200 on 264 degrees of freedom
## Multiple R-squared: 0.2427, Adjusted R-squared: 0.2398
## F-statistic: 84.61 on 1 and 264 DF, p-value: < 2.2e-16
## 2.5 % 97.5 %
## (Intercept) 134244.86 179209.9
## AE1 82862.03 127998.6
Compensacoes Pagas por AE2

##
## Call:
## lm(formula = Compensacoes.Pagas ~ AE2, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -269854 -63188 -26869 18936 1947780
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 156933 11421 13.740 <2e-16 ***
## AE2 105100 11439 9.188 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 186300 on 264 degrees of freedom
## Multiple R-squared: 0.2423, Adjusted R-squared: 0.2394
## F-statistic: 84.42 on 1 and 264 DF, p-value: < 2.2e-16
## 2.5 % 97.5 %
## (Intercept) 134444.88 179421.8
## AE2 82577.18 127623.3
Compensacoes Pagas por VC1

##
## Call:
## lm(formula = Compensacoes.Pagas ~ VC1, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -173174 -103549 -53336 26314 1978325
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 157250 13097 12.007 <2e-16 ***
## VC1 -12932 13114 -0.986 0.325
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 213600 on 264 degrees of freedom
## Multiple R-squared: 0.00367, Adjusted R-squared: -0.000104
## F-statistic: 0.9724 on 1 and 264 DF, p-value: 0.325
## 2.5 % 97.5 %
## (Intercept) 131462.38 183037.49
## VC1 -38753.12 12889.29
Compensacoes Pagas por DS1

##
## Call:
## lm(formula = Compensacoes.Pagas ~ DS1, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -274126 -66882 -31418 15194 1938060
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 156892 11900 13.184 < 2e-16 ***
## DS1 90049 11931 7.548 7.21e-13 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 194100 on 264 degrees of freedom
## Multiple R-squared: 0.1775, Adjusted R-squared: 0.1744
## F-statistic: 56.97 on 1 and 264 DF, p-value: 7.213e-13
## 2.5 % 97.5 %
## (Intercept) 133461.82 180323.0
## DS1 66556.93 113540.4
Compensacoes Pagas por DS2

##
## Call:
## lm(formula = Compensacoes.Pagas ~ DS2, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -337034 -61929 -9937 21990 1947623
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 156821 10385 15.10 <2e-16 ***
## DS2 130524 10403 12.55 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 169400 on 264 degrees of freedom
## Multiple R-squared: 0.3735, Adjusted R-squared: 0.3712
## F-statistic: 157.4 on 1 and 264 DF, p-value: < 2.2e-16
## 2.5 % 97.5 %
## (Intercept) 136372.4 177268.7
## DS2 110040.2 151007.1
Compensacoes Pagas por AR1

##
## Call:
## lm(formula = Compensacoes.Pagas ~ AR1, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -421569 -64283 -32469 17491 1888865
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 156907 11928 13.155 < 2e-16 ***
## AR1 89047 11957 7.447 1.36e-12 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 194500 on 264 degrees of freedom
## Multiple R-squared: 0.1736, Adjusted R-squared: 0.1705
## F-statistic: 55.46 on 1 and 264 DF, p-value: 1.357e-12
## 2.5 % 97.5 %
## (Intercept) 133420.93 180392.5
## AR1 65503.04 112591.1
Compensacoes Pagas por Grupos 2

##
## Call:
## lm(formula = Compensacoes.Pagas ~ grupos2, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -179296 -97308 -45535 35303 1959664
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 59151 39876 1.483 0.13917
## grupos2 67806 26056 2.602 0.00978 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 211300 on 264 degrees of freedom
## Multiple R-squared: 0.02501, Adjusted R-squared: 0.02132
## F-statistic: 6.772 on 1 and 264 DF, p-value: 0.009784
## 2.5 % 97.5 %
## (Intercept) -19365.61 137666.8
## grupos2 16501.11 119110.3
Compensacoes Pagas por Unidades Consumidoras

##
## Call:
## lm(formula = Compensacoes.Pagas ~ Unidades.Consumidoras, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -193436 -83463 -53600 20765 1988619
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 8.477e+04 2.068e+04 4.098 5.54e-05 ***
## Unidades.Consumidoras 2.326e+00 5.246e-01 4.434 1.36e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 206400 on 264 degrees of freedom
## Multiple R-squared: 0.06931, Adjusted R-squared: 0.06578
## F-statistic: 19.66 on 1 and 264 DF, p-value: 1.359e-05
## 2.5 % 97.5 %
## (Intercept) 44042.412050 1.254897e+05
## Unidades.Consumidoras 1.293067 3.358864e+00

Olhando em primeira mã o para o modelo vemos a plotagem das variá veis em busca de suas
correlaçõ es. Em todos os grá ficos, nã o se há uma certeza de uma correlaçã o devido a
dispersã o dos dados. Porém, em algumas, podemos assumir um grau um pouco mais alto
de correlaçã o. Em relaçã o à s compensaçõ es pagas, o Grupo2 nã o demonstra nenhuma
correlaçã o. As variá veis Descargas e AR1 também uma baixissima correlaçã o. As variá veis
Vento, AG1 e DS2 possuem a correlaçã o um pouco maior que as anteriores, mas ainda bem
pouca e certamente inconclusiva. Já as variá veis AE1, AE2 e DS1 demonstram um grau de
correlaçã o mais claro, seguido das Unidades Consumidoras, Temperatura, Umidade e
Chuva que já demonstram uma correlaçã o e tendo menos dispersã o dos dados.

Ajuste do Modelo
modelo <- lm(Compensacoes.Pagas ~ Volume.chuva, data=dados)
summary(modelo)

##
## Call:
## lm(formula = Compensacoes.Pagas ~ Volume.chuva, data = dados)
##
## Residuals:
## Min 1Q Median 3Q Max
## -148659 -107542 -51844 27486 1996466
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 159332.647 66992.331 2.378 0.0181 *
## Volume.chuva -4.264 137.159 -0.031 0.9752
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 214000 on 264 degrees of freedom
## Multiple R-squared: 3.661e-06, Adjusted R-squared: -0.003784
## F-statistic: 0.0009664 on 1 and 264 DF, p-value: 0.9752

# Intervalo de confianca para os parametros do modelo

confint(modelo)

## 2.5 % 97.5 %
## (Intercept) 27425.3834 291239.9106
## Volume.chuva -274.3279 265.8002

Interpretacao dos resultados

Interpretando o coeficiente de determinacao ( R2)
dados$fit <- predict(modelo)
SQT <- with(dados, sum((Compensacoes.Pagas -
mean(Compensacoes.Pagas))^2))
SQres <- with(dados, sum((Compensacoes.Pagas - fit)^2))

(R2 <- 1 - SQres/SQT)

## [1] 3.660562e-06

letra B
Ranqueamento das variaveis preditoras em escala original
require(exploreR)

## Carregando pacotes exigidos: exploreR

saida <- masslm(data = dados[,c("Compensacoes.Pagas",

"AG1","AE1","AE2","VC1","DS1","DS2","AR1")], dv.var = "Compensacoes.Pagas")
print(saida)

## IV Coefficient P.value R.squared

## 1 AG1 78680 5.384e-10 0.136042278
## 2 AE1 105400 1.138e-17 0.242706700
## 3 AE2 105100 1.224e-17 0.242291294
## 4 VC1 -12930 3.250e-01 0.003669943
## 5 DS1 90050 7.213e-13 0.177481846
## 6 DS2 130500 1.234e-28 0.373545846
## 7 AR1 89050 1.357e-12 0.173600749

Ranqueamento das variaveis preditoras em escala logarítmica

## Variáveis Coeficiente P - Valor R Quadrado
## 1 Codigo -1.684e-02 NaN 1.000000e+00
## 2 DSC_Conj_N 2.290e-01 NaN 1.000000e+00
## 22 Compensacoes.Pagas 1.400e-06 1.481e-47 5.490023e-01
## 13 DS2 2.593e-01 2.268e-32 4.128483e-01
## 12 DS1 2.572e-01 1.170e-31 4.055422e-01
## 10 AE2 2.500e-01 1.294e-29 3.840913e-01
## 9 AE1 2.440e-01 8.918e-28 3.641480e-01
## 14 AR1 2.218e-01 2.293e-22 3.017056e-01
## 17 Unidades.Consumidoras 8.228e-06 1.083e-17 2.429889e-01
## 8 AG1 1.650e-01 3.631e-12 1.675231e-01
## 4 Descargas.atm 4.035e-05 4.639e-07 9.192137e-02
## 15 lnDEC 8.863e-02 2.994e-04 4.840498e-02
## 18 DEC.ANEEL 1.224e-02 3.046e-04 4.828851e-02
## 16 grupos2 1.138e-01 2.186e-02 1.975207e-02
## 20 FEC.ANEEL 1.287e-02 1.637e-01 7.333747e-03
## 3 Volume.chuva 3.346e-04 1.964e-01 6.313233e-03
## 23 fit -7.846e-05 1.964e-01 6.313233e-03
## 5 Vento -7.606e-02 2.636e-01 4.731870e-03
## 21 FEC.Limite -1.574e-02 2.952e-01 4.149514e-03
## 6 umidade -5.135e-03 3.669e-01 3.084980e-03
## 11 VC1 -2.451e-03 9.214e-01 3.693681e-05
## 7 temperatura -1.545e-03 9.453e-01 1.787904e-05
## 19 DEC.Limite 4.052e-04 9.477e-01 1.632241e-05

/*```{r} dados <- dados[,c(“Compensacoes.Pagas”,

“AG1”,“AE1”,“AE2”,“VC1”,“DS1”,“DS2”,“AR1”)]
require(exploreR)
saida <- masslm(dados, dv.var = “Compensacoes.Pagas”) saida <-
saida[order(saida$R.squared, decreasing = TRUE),] as.data.frame(saida)
DT::datatable(saida, rownames = FALSE) ```