如何将回归方程放在每月时间序列数据的图上？

Question

I'd like to analyse monthly rainfall data (make time series plot + regression equation for time series). I've written code in R and plot the monthly time series data and I've tried to make different regression equations (linear and non-linear) and show these equation on the same graph of time series plot but unfortunately I cannot. May be because I'm new user of R / Rstudio statistical packages.

Data style

Date   monthly rainfall (mm)
jan94     12
Feb94      11
.
.
. Dec14    1x

The code

# plotting of time series rainfall data (option1)
# step1: read files

MR<-read.table("C:\\Users\\Salam\\Desktop\\trend Kufa\\CSV2 Habel\\Monthly rainfall.csv", header=T,sep=",")

summary(MR)
names(MR)
MR

# step2: plot observed discharge

MR1<-MR[c(1:252),2];

summary (MR1)
MR1
class(MR1)

require(zoo)

x <- yearmon(1994 + seq(0, 251)/12)
x

y<-MR1
y

pcp<-y~x

plot(pcp,type="l", xlab="Month",ylab="Monthly Rainfall(mm)", axes=T) 

grid(nx=250, ny=250, col="lightgray", lty="solid")
lines(pcp,lwd=2, col="blue")

box(which='plot')
title("Monthly Observed rainfall(mm)")

##  Regression



S1 <- lm(y ~ z, data=MR)
abline(S1,col='red',lwd=3)
summary(S1)


S2<-lm( y~poly(x,3), data=MR)
summary(S2)
abline(S2,col='green',lwd=3)

S3 <- nls(y ~ exp(a + b / x),start = list(a = 0, b = 0))
summary(S3)

S4 <- nls(y ~ (a + b *log( x)), start = list(a = 0, b = 0))
summary(S4) 

Answer 1

You can use the text function to put the equations on the plots.

text(x, y, "S1 <- lm(y ~ z, data=x)",
     cex = .8)

the x and y are the coordinates on the plot where you would like the equation

put the equation in quotes

data is your data frame

cex controls the font size

for more info & options on text use ?text

Answer 2

Check the following example and you can modify yours easily.

library(ggplot2)

# example dataset
dt = data.frame(date = 1:10,
                value = c(10,11,15,13,16,17,18,19,16,22))

# plot everything in one graph
ggplot(dt, aes(date, value)) +
 geom_point() + # plot the points
 stat_smooth(method="lm",se=F,level=0.95, col="red") + # linear reg
 stat_smooth(method="lm", formula = y~poly(x,2,raw=T), se=F,level=0.95, col="blue") + # quadratic reg
 stat_smooth(method="lm", formula = y~poly(x,3,raw=T), se=F,level=0.95, col="green") # cubic reg

# plot everything in separately
library(gridExtra)

plot1 = ggplot(dt, aes(date, value)) +
        geom_point() +
        stat_smooth(method="lm",se=T,level=0.95, col="red") 

plot2 = ggplot(dt, aes(date, value)) +
        geom_point() +
        stat_smooth(method="lm", formula = y~poly(x,2,raw=T), se=T,level=0.95, col="blue") 

plot3 = ggplot(dt, aes(date, value)) +
        geom_point() +
        stat_smooth(method="lm", formula = y~poly(x,3,raw=T), se=T,level=0.95, col="green")


grid.arrange(plot1,plot2,plot3)

I hope you are familiar with the ggplot2 package as it is the most important in this case. You can then investigate ways to add titles, change colours, change confidence intervals, etc.