Homec4science
Diffusion Laser-DED Self supervised learning-Coaxial Imaging (master) Browse

R12521/BYOL/Anomaly/Utils_one_class_SVM.py
c4cea181e418master

Last Change
Raw File

Utils_one_class_SVM.py
View Options

# -*- coding: utf-8 -*-
"""
Created on Thu Jun 9 13:37:13 2022
@author: srpv
"""
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.font_manager
import pandas as pd
import os
from sklearn.model_selection import train_test_split# implementing train-test-split
from sklearn import svm
from sklearn.manifold import TSNE
def anomaly_detection(X_train,X_test,y_train,y_test):
Featurespace=np.concatenate((X_train,X_test))
classspace=np.concatenate((y_train,y_test))
X_Features,X_labels,Y_Features,Y_labels = anomaly_data_prep(Featurespace, classspace)
anomaly_one_class_SVM(X_Features,X_labels,Y_Features,Y_labels)
def anomaly_one_class_SVM(X_Features,X_labels,Y_Features,Y_labels):
folder = os.path.join('Figures/', 'Anomaly')
try:
os.makedirs(folder, exist_ok = True)
print("Directory created....")
except OSError as error:
print("Directory already exists....")
print(folder)
folder=folder+'/'
Outliers=Y_Features.to_numpy()
X_Features=X_Features.to_numpy()
X_train, X_test, y_train, y_test = train_test_split(X_Features, X_labels, test_size=0.25, random_state=66)
# fit the model
clf = svm.OneClassSVM(nu=0.1, kernel="rbf", gamma='auto')
clf.fit(X_train)
y_pred_train = clf.predict(X_train)
y_pred_test = clf.predict(X_test)
y_pred_outliers = clf.predict(Outliers)
n_error_train = y_pred_train[y_pred_train == -1].size
n_error_test = y_pred_test[y_pred_test == -1].size
n_error_outliers = y_pred_outliers[y_pred_outliers == 1].size
print(
"error train: %d/%d ; errors novel regular: %d/%d ; "
"errors novel abnormal: %d/%d"
% (n_error_train,len(X_train), n_error_test,len(X_test), n_error_outliers,len(Outliers)))
fig = plt.figure(figsize=(9,9), dpi=800)
plt.rcParams["legend.markerscale"] = 3
plt.rcParams['xtick.labelsize']=25
plt.rcParams['ytick.labelsize']=25
s = 60
c = plt.scatter(Outliers[:, 0], Outliers[:, 1], c="orange", s=s, edgecolors="k",alpha=0.2)
b1 = plt.scatter(X_train[:, 0], X_train[:, 1], c="cyan", s=s, edgecolors="k")
b2 = plt.scatter(X_test[:, 0], X_test[:, 1], c="lightblue", s=s, edgecolors="k")
plt.axis("tight")
plt.xlabel ('Dimension 1', labelpad=5,fontsize=30)
plt.ylabel ('Dimension 2', labelpad=5,fontsize=30)
# plt.ticklabel_format(axis="x", style="sci", scilimits=(0,0))
plt.legend(
[ b1, b2, c],
[
"Training (conduction)",
"Testing (conduction)",
"Outlier observations (LoF,Keyhole)",
],
loc="upper right",
bbox_to_anchor=(0.90, -0.15),
prop=matplotlib.font_manager.FontProperties(size=25),
)
plotname=folder+'anomaly_classification_accuracy.png'
plt.savefig(plotname, bbox_inches='tight',dpi=800)
plt.show()
def anomaly_data_prep(X, y):
y = np.ravel(y)
RS=np.random.seed(1974)
perplexity=20
tsne = TSNE(n_components=3, random_state=RS, perplexity=perplexity)
X = tsne.fit_transform(X)
print(X.shape)
X = pd.DataFrame(X)
labels = pd.DataFrame(y)
labels.columns = ['Categorical']
labels=labels['Categorical'].replace(0,-1)
labels = pd.DataFrame(labels)
labels=labels['Categorical'].replace(2,-1)
labels = pd.DataFrame(labels)
labels=labels['Categorical'].replace(3,-1)
labels = pd.DataFrame(labels)
df=pd.concat([X,labels], axis=1)
new_columns = list(df.columns)
new_columns[-1] = 'target'
df.columns = new_columns
df.target.value_counts()
df = df.sample(frac=1.0)
df_1 = df[df.target == 1]
print(df_1.shape)
df_2 = df[df.target != 1]
print(df_2.shape)
X_labels=df_1.iloc[:,-1]
X_labels = pd.DataFrame(X_labels)
X_Features = df_1.drop(labels='target', axis=1)
Y_labels=df_2.iloc[:,-1]
Y_labels = pd.DataFrame(Y_labels)
Y_Features = df_2.drop(labels='target', axis=1)
return X_Features,X_labels,Y_Features,Y_labels
c4science · Help