python 3.x - how to add noise to data till the classifier can't classify the data anymore -


i'm working on machine supervised learning problem , i've been asked after classified data add different levels of gaussian noise till algorithms can't classify data more. data floating points , i'm using algorithms such knn, ann, gnb, random forest , others. whatever noise add data classification accuracy never went below 31%. question cause situation , how solve it.

the code i'm using add noise:

import pandas pd import numpy np  clean_pd = pd.read_csv("pddata.csv")     mu, sigma = 0, 0.02 # creating noise same dimension dataset  noise = np.random.normal(mu, sigma, [60000, 3])  print(noise)  nois_pd = clean_pd + noise  #save noise  nois_pd.to_csv("noisedata).csv", index=false) 

thanks


Comments

Popular posts from this blog

angular - Ionic slides - dynamically add slides before and after -

minify - Minimizing css files -

Add a dynamic header in angular 2 http provider -