Algorithm 1: Replacement Noise Algorithm

Input: Dataset D with N samples and F features

Output: D’, a corrupted dataset with N samples and F features

1: Define noise term p, the proportion of values to replace

2: For each $f \in F:$

3:   $d_f=column f of Dataset D$

4:   ${\mu }_f=\frac{sum\left(d_f\right)}{N}$

5:   ${\sigma }_f=\sqrt{\frac{({\sum }_{n=0}^N{d}_{n,f}-{\mu }_f) }{N-1}}$

6:   $v=p\ast N$

7:   Randomly remove v values from column $d_f$

8:   Replace each value with a randomly sampled value from N$({\mu }_f,{\sigma }_f^2)$

9: End for loop