Forum Discussion

Gracie's avatar
Gracie
Helper II
9 years ago

How to put x as dates for R linear regression

I am writing R script to do linear regression. My dateset is date and price. Date is from Date table and Price is from Sales table. These two tables have no relationship setup. How can I put date as x axis ? And how to cancel #Remove duplicated rows because I don't want to remove duplicated rows?

 

Here is my code:

#Create dataframe
#dataset <- data.frame(Date,Price)

#Remove duplicated rows
#dataset <- unique(dataset)

strDates <- dataset[,1]
x <- as.Date(strDates, "%m/%d/%Y")
y <- as.numeric(dataset[,2])
plot(y~x, xlim = c(as.POSIXct("1/1/2010", format = "%m/%d/%Y"),as.POSIXct("1/1/2010", format = "%m/%d/%Y")))
lm.ab<-lm(y ~ x)
abline(lm.ab)
summary(lm.ab)
Need Help

3 Replies

  • v-huizhn-msft's avatar
    v-huizhn-msft
    Microsoft Employee
    9 years ago

    Hi Gracie,

    Please create a index column in Date and Sales table separately.  Then merge the two table to one according to index field.

    Then you load the Index, Date and Price to create a dataset.

    Change your corresponding part of code to the following code. 

    strDates <- dataset[,3]
    y <- as.numeric(dataset[,3])

    Then run your R code and change if it works fine.

    Best Regards,
    Angelia

    • Gracie's avatar
      Gracie
      Helper II
      9 years ago

      Hi v-huizhn-msft

       

      It doesn't work. I guess your point is to add the index to avoid removing duplicated rows. But the x-axis still can't show as dates.

       

       

      • v-huizhn-msft's avatar
        v-huizhn-msft
        Microsoft Employee
        9 years ago

        Hi Gracie,

        >>But the x-axis still can't show as dates.

        Could you please share a screenshot for further analysis?

        Thanks,
        Angelia