model.py

import pandas as pd
import matplotlib.pyplot as plt
import pickle



pd.set_option('display.max_column',None) ## Displaying all columns ##


# ## Data Importing

df = pd.read_csv('./data.csv')  ## Importing Breast cancer data ##
df.drop(['id'], axis = 1, inplace=True)
df['diagnosis']=df['diagnosis'].map({'M':1 ,'B':0 })


mean_features = list(df.columns[1:11])
se_features = list(df.columns[11:21])
worst_features = list(df.columns[21:31])

mean_features.append('diagnosis')
se_features.append('diagnosis')
worst_features.append('diagnosis')

corr_mean = df[mean_features].corr()
corr_se = df[se_features].corr()
corr_worst = df[worst_features].corr()



# ## Random Forest Model

## taking in account features having corr 60% or higher (considering mean)
## taking in account features having corr 50% or higher (considering se)
## taking in account features having corr 70% or higher (considering worst)

prediction_vars = ['texture_mean', 'area_mean', 'smoothness_mean', 'concavity_mean',
       'area_se', 'concavity_se', 'fractal_dimension_se', 'smoothness_worst',
       'concavity_worst', 'symmetry_worst']



from sklearn.model_selection import train_test_split

train,test = train_test_split(df , test_size=0.3 , random_state=42)

train_X = train[prediction_vars]
train_y = train['diagnosis']
test_X = test[prediction_vars]
test_y = test['diagnosis']




from sklearn.ensemble import RandomForestClassifier

model=RandomForestClassifier()


model.fit(train_X,train_y)

from sklearn.metrics import confusion_matrix

confusion = confusion_matrix(test_y,model.predict(test_X))
class_label = ["malignant", "benign"]
df_cm = pd.DataFrame(confusion, index=class_label,columns=class_label)


from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.metrics import accuracy_score

precision=precision_score(test_y,model.predict(test_X))

recall=recall_score(test_y,model.predict(test_X))

accuracy=accuracy_score(test_y,model.predict(test_X))

print("precision score :%5.3f "%(precision))
print("recall score :%5.3f "%(recall))
print("accuracy score :%5.3f "%(accuracy))

pickle.dump(model, open(r".\model.pkl", "wb"))