Using "ifelse" with negative values

I did a sentiment analysis using VADER and now want to classify the values with negative, positive and neutral.

Positive when compound score is > 0.05

Negative when its < - 0.05 neutral when in between -0.05 and 0.05

df_polarity$VADER_Sent = ifelse(df_polarity$VADER_Sent > 0.05, "pos", 
                            ifelse (df_polarity$VADER_Sent < -0.05, "neg", 
                                    ifelse (between(df_polarity$VADER_Sent, -0.05, 0.05) , "neu", "NA")
                            ) 
)

When running this code, even values with - 0.4XXX will be classified as neutral and not as negative.

For some reason this won't work. There is anything I am missing... but I can figure out what it is...

I couldn't find any helpful tipps by googling it.

I hope someone of you can help me with this one!

Output from str(df_polarity):

$ VADER_Sent : chr "0.0" "-0.4939" "0.7717" "0.7096"

After further looking into my data, it seems that the "-" sign is not recognized in the context of a negative number.

Thanks to everyone who tried to help me! Really appreciated it!!!

CodePudding user response：

The problem is because the VADER_Sent column is character. The comparisons < and > are checking alphabetically instead of numerically.

Example:

> -0.4939 < -0.05
[1] TRUE

> "-0.4939" < "-0.05"
[1] FALSE

Try using as.numeric(df_polarity$VADER_Sent) in your ifelse() statements to get around this.

CodePudding user response：

Can't be sure without reproducible code, but you should be able to just give the 'neutral' category as the second option in the second ifelse() call.

df_polarity$VADER_Sent = ifelse(df_polarity$VADER_Sent > 0.05, "pos", 
                            ifelse(df_polarity$VADER_Sent < -0.05, "neg", "neutral"
                         )
)

CodePudding user response：

This should work:

df %>%
  mutate(X = if_else(VADER_sent < -0.5, "neg", 
                    if_else(VADER_sent <= 0.5 & VADER_sent >= -0.5, "neutral", "pos"))
  )
  VADER_sent       X
1       0.51     pos
2       0.10 neutral
3       2.00     pos
4      -0.60     neg
5       0.30 neutral
6      -1.20     neg

Data:

df <- data.frame(
  VADER_sent = c(0.51, 0.1, 2,-0.6, 0.3, -1.2)
)