I have used the following code to obtain mean decrease in accuracy for random forest
AAA<-randomForest(CPercentage~., data=data, importance= T)
BBB<-as.data.frame(importance(AAA))
I have created the following dataframe by the above process
%IncMSE IncNodePurity
Campaigntype 3.4815273 207.5336052
Email -1.1606079 2042.5660103
get 4.9073550 35.1237017
free 2.8777972 14.5362957
new 8.4464445 93.3491610
buy 5.9636483 23.9926669
just 4.1262164 21.5611278
month 4.0817729 16.6345631
I am able to obtain the second and third column by BBB$%IncMSE and
BBB$IncNodePurity. I want to subset thsi based on the first column which appears unnamed. I am unable to do this. When writing this dataframe to a csv file, it works and all three columns are listed separately. However, I am unable to separate the first two columns. Is there any way i can do this and rename the first column. Will be grateful to anyone who helps
It seems that your "first column" may actually be the index. You can create a column in your data frame from the index, and then reset your index so they're row numbers instead of names.
Try this:
BBB$col_one <- rownames(BBB)
rownames(BBB) <- 1:nrow(rownames)
Related
## my data frame
crime = read.csv("url")
## specific columns that need to be represented
property_crime = crime$Burglary + crime$Theft + crime$`Motor Vehical Theft`
## the rows that I am looking for have the name "harris" within the column named "county_name"
## my attempt
with(crime, hist(harris))
## Error in hist(harris) : object 'harris' not found
Not sure why I am getting object 'harris' not found as that is the name under the county_name column. I'm new to R, could someone walk me through the process of displaying a histogram only including the values of specific columns and specific rows?
the rows that I am looking for have the name "harris" within the column named "county_name"
You have to tell R the same logic that you are telling us.
There are several ways of making this in R but I am going to put here the base R way.
We can access the desired rows of object crime column county_name by indexing like data.frame[rows, columns]. So, in your case, crime[harris_rows, "county_name"] should work. To get harris_rows, we can make a boolean index like so crime$county_name == harris. If we put all of this together and call hist():
hist(crime[crime$county_name == "harris", "county_name"])
You don't provide a reproducible example, but you can check a similar logic with the mtcars dataset. Here, I am making the histogram of the cars with mpg > 15
hist(mtcars[mtcars$mpg >15, "mpg"])
# this is another option that produces the same result
# hist(mtcars$mpg[mtcars$mpg >15])
I'm trying to add a conditional column to a dataframe, but not getting the results I'm expecting.
I have a dataframe with values recorded for the column "steps" across 5-minute intervals over various days. I'm trying to impute missing values in the 'steps' column by using the mean number of steps for a given 5-minute interval on the days that do have measurements. n.b. I tried using the MICE package for this but it just crashed my computer so I opted for a more manual workaround.
As an intermediate stage, I have bound an additional column to the existing dataframe with the mean number of steps for that interval. What I want to do next is create a column that returns that mean if the raw number of steps is NULL, and just uses the raw value if not null. Here's my code for that part:
activityTimeAvgs$stepsImp <- if(is.na(activityTimeAvgs$steps)){
activityTimeAvgs$avgsteps
} else {
activityTimeAvgs$steps
}
What I expected to happen is that the if statement would evaluate as TRUE if 'steps' is NA and consequently give 'avgsteps'; in cases where 'steps' is not NA I would expect it to just use the raw value for 'steps'. However, the output just gives the value for 'avgsteps' in every row, which is not much use. I also get the following warning:
Warning message:
In if (is.na(activityTimeAvgs$steps)) { :
the condition has length > 1 and only the first element will be used
Any ideas where I'm going wrong?
Thanks in advance.
The if statement is not suitable for this. You need to use ifelse:
activityTimeAvgs$stepsImp <- ifelse(is.na(activityTimeAvgs$steps), activityTimeAvgs$avgsteps, activityTimeAvgs$steps)
This is how I've retrieved my dataset, everything is good so far.
> mantis<-read.csv("mantis.csv")
> attach(mantis)
The dataset provides numerical data on body mass/length/claw strength/etc. of FEMALE and MALE mantises. The object is to create a histogram showing the body masses of ONLY female mantises. I created a subset;
> mantis_sub<-subset(mantis, Sex=="f",select="Body.Mass.g")
Then I tried;
> hist(mantis_sub)
Error in hist.default(mantis_sub) : 'x' must be numeric
I've searched this link;
Plot a histogram of subset of a data
...and I cannot figure out how to properly create this histogram. I am unfortunately not fluent enough in R to understand the solution and the textbook I'm using does not cover this.
It is because mantis_sub is a dataframe (ie a table of body masses, lengths, claw strengths, ..), not a set of numbers, so hist is unsure which column you wish to plot.
You need to extract the column you want to do a histogram of. To do this you put mantis_sub${column name}. The dollar sign extracts the appropriate column from the mantis_sub table.
e.g. to do a histogram of the column named "BodyMass"
hist(mantis_sub$BodyMass)
If you want to do histograms of many columns automatically, then you'll have to loop through them, e.g.
for (column in c("BodyMass", "ClawStrength")) {
hist(mantis_sub[[column]])
}
Question:
I have a data.frame (hlth) that consists of 49 vectors - a mix of numeric(25:49) and factor(1:24). I am trying to randomly select 50 rows, then calculate column means only for the numeric columns (dropping the other values), and then place the random row mean(s) into a new data.frame (beta). I would then like to iterate this process 1000 times.
I have attempted this process but the values that get returned are identical and the new means will not enter the new data.frame
Here is a few rows and columns of the data.frame(hlth)
DateIn adgadj Sex VetMedCharges pwtfc
1/01/2006 3.033310 STEER 0.00 675.1151
1/10/1992 3.388245 STEER 2540.33 640.2261
1/10/1995 3.550847 STEER 572.78 607.6200
1/10/1996 2.893707 HEIFER 549.42 425.5217
1/10/1996 3.647233 STEER 669.18 403.8238
The code I have used thus far:
set.seed[25]
beta<-data.frame()
net.row<-function(n=50){
netcol=sample(1:nrow(hlth),size=n ,replace=TRUE)
rNames <- row.names(hlth)
subset(hlth,rNames%in%netrow,select=c(25:49))
colMeans(s1,na.rm=TRUE,dims=1)
}
beta$net.row=replicate(1000,net.row()); net.row
The two issues, that I have detected, are:
1) Returns the same value(s) each iteration
2) "Error during wrap-up: object of type 'closure' is not subsettable" when the beta$netrow
Any suggestions would be appreciated!!!
Just adding to my comment (and firstly pasting it):
netcol=sample(1:nrow(hlth),size=n ,replace=TRUE) should presumably by netrow = ... and the error is a scoping problem - R is trying to subset the function beta, presumably again, because it can't find netRowMeans in the data.frame you've defined, moves on to the global environment and throws an error there.
There are also a couple of other things. You don't assign subset(hlth,rNames%in%netrow,select=c(25:49)) to a variable, which I think you mean to assign to s1, so colMeans is probably running on something you've set in the global environment.
If you want to pass a variable directly in to the data frame beta in that manner, you'll have to initialise beta with the right number of columns and number of rows - the column means you've passed out will be a vector of (1 x 25), so won't fit in a single column. You would probably be better of initalising a matrix called mat or something (to avoid confusion with scoping errors masking the actual error messages) with 25 columns and 1000 rows.
EDIT: Question has been edited slightly since I posted this, but most points still stand.
For a marketing class I have to write a function that calculates the retention rate of the customers (probability that a customer still is a customer). I've come so far that I isolated the ids of the individual customers and stored them in the matrix first.transactions.data. I then split them into cohorts (group of customers by time) with split() and stored them in the list cohort.
Now comes my problem: I calculated another sub-matrix from the full data set called final.period.data where I will calculate the retention rate. However, therefore I have to isolate the ids in final.period.data for each cohort. My instructor told me that I should create an additional column in final.period.data that shows TRUE or FALSE depending on whether the cohort's id and final.period.data's id are the same. For this I tried to use exists, but I always receive error messages. I tried the following:
final.period.data <- if(exists(cohort$'1'$id, where = final.period.data$id) final.period.data$same = TRUE)
but always receive error messages such as: unexpected symbol or invalid first argument. I also tried to convert the list cohort into a matrix but this didn't help either. How do I have to change the exist command or is there a simpler way to locate cohort's ids in final.period.data?
Thank you for your help.
You can just create a function that does what you want:
funct <-(final.period.data){
if (final.period.data$cohort =='1' & final.period.data$id ==<condition2>){
#Change the number for the TRUE condition}
else{ #If it doesn't fit the two conditions
#Change the number for the FALSE condition}
}
vector <- c(nrow(final.period.data))
final.period.data <- cbind(vector)
And use it as the apply function. Here will you find more information about apply
But I usually do it with a for loop, first creating the new column and then adding it to the data frame.