svm-ozone-rbf.py

from sklearn.decomposition import PCA
from sklearn import svm
from sklearn.model_selection import train_test_split
import numpy as np
from sklearn import metrics
import pandas as pd
from sklearn.metrics import f1_score, precision_score, recall_score, classification_report, confusion_matrix

#from sklearn.preprocessing import LabelEncoder

#read data
df = pd.read_csv('ozone-data.csv')
#print(df.head())

#data select
X = df.iloc[:,1:73]
#print(X)
#target select [class]
y = df.iloc[:,73:74]
#print(y)
#target type of Dataframe to type of Numpy Array
yyy = np.array(y).ravel()

#train,test of creating
X_train, X_test, y_train, y_test = train_test_split(X,yyy,test_size=0.2880820836621942,random_state=49) # 80% training and 20% test


#creationg pca model
model_ozone = PCA(n_components=72).fit(X_train)
model_ozone_2d = model_ozone.transform(X_train)

#print(model_ozone_2d)e


#creating of model the SVM
model = svm.SVC(kernel='rbf', gamma=0.05, C=3)

#The model is training in equation
model.fit(model_ozone_2d, y_train)

#prediction equation is the model.
y_test_pred = model.predict(X_test)
y_train_pred = model.predict(X_train)

#accuracy is model
print("Accuracy:",metrics.accuracy_score(y_test,y_test_pred))
#Accuracy: 0.9602739726027397

from sklearn.metrics import confusion_matrix
c_matrix = confusion_matrix(y_train, y_train_pred)
print(c_matrix)
#[[1676    0]
# [ 128    0]]

from sklearn.metrics import auc

fpr, tpr, _ = metrics.roc_curve(y_train, y_train_pred)

auc_test_roc_curve= (fpr,tpr)
print("Auc Roc Curve Score: ",auc_test_roc_curve)
#Auc Roc Curve Score:  0.5


print("Precision Score: ", precision_score(y_test, y_test_pred, average="macro").mean()*100)
print("Recall Score: ", recall_score(y_test, y_test_pred, average="macro").mean()*100)
print("F1 Score: ",f1_score(y_test, y_test_pred, average="macro").mean()*100)

#Precision Score:  47.80821917808219
#Recall Score:  50.0
#F1 Score:  48.87955182072829


#data visualition
import pylab as pl
for i in range(0, model_ozone_2d.shape[0]):
    if y_train[i] == 0:
        c1 = pl.scatter(model_ozone_2d[i,0],model_ozone_2d[i,1],color='r',edgecolors='y',marker='*',linewidths=1)

    elif y_train[i] == 1:
        c2 = pl.scatter(model_ozone_2d[i,0],model_ozone_2d[i,1],color='g',edgecolors='y',marker='o',linewidths=1)
import matplotlib.pyplot as plt
pl.legend([c1, c2], ['Ozone Day', 'Normal Day'])
plt.title('Ozone and Normal Day Classification')
pl.show()