One common method for detecting and removing outliers is the z-score method. This technique identifies how far a data point is from the mean in terms of standard deviations. If a data point lies beyond a certain threshold (commonly ±3), it is considered an outlier.

What Is a Z-Score?

A z-score (also known as a standard score) is calculated using the formula:

Where:

• $X$: the original data point;
• $\mu$: the mean of the dataset;
• $\sigma$: the standard deviation of the dataset.

Calculating Z-Scores

You can either compute z-scores manually by following the formula:

mean_cgpa <- mean(df$cgpa)
sd_cgpa <- sd(df$cgpa)
df$cgpa_zscore <- (df$cgpa - mean_cgpa) / sd_cgpa

Or you can use the built-in function:

df$cgpa_zscore <- scale(df$cgpa)

Identifying Outliers

After calculating the z-scores, you can choose a threshold (±3 in this case) and apply a simple filtering operation to select all entries outside of the range:

thresh_hold <- 3
outliers <- df[df$cgpa_zscore > thresh_hold | df$cgpa_zscore < -thresh_hold, ]

Or you can select all entries inside the range to create an outlier-free dataset:

df2 <- df[df$cgpa_zscore < thresh_hold & df$cgpa_zscore > -thresh_hold, ]