Skip to content

Commit

Permalink
formatting
Browse files Browse the repository at this point in the history
  • Loading branch information
@edublancas
edublancas committed Sep 18, 2024
1 parent 373cac8 commit c4047eb
Show file tree
Hide file tree
Showing 40 changed files with 2,860 additions and 2,503 deletions.
174 changes: 105 additions & 69 deletions _kaggle/_render/breast-cancer-diagnostic-classification/nb.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,8 @@
# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory

import os
for dirname, _, filenames in os.walk('/kaggle/input'):

for dirname, _, filenames in os.walk("/kaggle/input"):
for filename in filenames:
print(os.path.join(dirname, filename))

Expand All @@ -35,7 +36,7 @@
import os
import warnings

warnings.filterwarnings('ignore')
warnings.filterwarnings("ignore")
import time as t
import pandas as pd
import numpy as np
Expand All @@ -45,7 +46,17 @@
from sklearn.preprocessing import StandardScaler
from sklearn.utils import resample
from imblearn.over_sampling import SMOTE
from sklearn.metrics import confusion_matrix, accuracy_score, f1_score, recall_score, precision_score, classification_report, roc_curve, auc, roc_auc_score
from sklearn.metrics import (
confusion_matrix,
accuracy_score,
f1_score,
recall_score,
precision_score,
classification_report,
roc_curve,
auc,
roc_auc_score,
)
from sklearn.linear_model import LogisticRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
Expand All @@ -62,8 +73,7 @@


# %%
def data_load(
): #check for the availability of the dataset and change cwd if not found
def data_load(): # check for the availability of the dataset and change cwd if not found
df = pd.read_csv("../input/breast-cancer-prediction/data.csv")
return df

Expand All @@ -73,15 +83,15 @@ def data_clean(df):


def X_y_split(df):
X = df.drop(['diagnosis'], axis=1)
y = df['diagnosis']
X = df.drop(["diagnosis"], axis=1)
y = df["diagnosis"]
return X, y


def data_split_scale(X, y, sampling):
#Splitting dataset into Train and Test Set
# Splitting dataset into Train and Test Set
X_tr, X_test, y_tr, y_test = train_test_split(X, y, test_size=0.3)
#Feature Scaling using Standardization
# Feature Scaling using Standardization
ss = StandardScaler()
X_tr = ss.fit_transform(X_tr)
X_test = ss.fit_transform(X_test)
Expand All @@ -91,8 +101,7 @@ def data_split_scale(X, y, sampling):
samp_sel = int(input("Now enter your selection for sampling strategy: \t"))
samp = [sampling.upsample, sampling.downsample, sampling.smote]
temp = samp[samp_sel - 1]
X_train, y_train = temp(X_train=pd.DataFrame(X_tr),
y_train=pd.DataFrame(y_tr))
X_train, y_train = temp(X_train=pd.DataFrame(X_tr), y_train=pd.DataFrame(y_tr))
return pd.DataFrame(X_train), pd.DataFrame(X_test), y_train, y_test


Expand All @@ -107,29 +116,28 @@ def data_split_scale(X, y, sampling):
class sampling:

def upsample(X_train, y_train):
#combine them back for resampling
# combine them back for resampling
train_data = pd.concat([X_train, y_train], axis=1)
# separate minority and majority classes
negative = train_data[train_data.diagnosis == 0]
positive = train_data[train_data.diagnosis == 1]
# upsample minority
pos_upsampled = resample(positive,
replace=True,
n_samples=len(negative),
random_state=30)
pos_upsampled = resample(
positive, replace=True, n_samples=len(negative), random_state=30
)
# combine majority and upsampled minority
upsampled = pd.concat([negative, pos_upsampled])
# check new class counts
#print(upsampled.diagnosis.value_counts())
# print(upsampled.diagnosis.value_counts())
print(upsampled.diagnosis.value_counts())
upsampled = upsampled.sample(frac=1)
X_train = upsampled.iloc[:, 0:-2]
y_train = upsampled.iloc[:, -1]
#graph barplot counts
# graph barplot counts
return X_train, y_train

def downsample(X_train, y_train):
#combine them back for resampling
# combine them back for resampling
train_data = pd.concat([X_train, y_train], axis=1)
# separate minority and majority classes
negative = train_data[train_data.diagnosis == 0]
Expand All @@ -139,23 +147,24 @@ def downsample(X_train, y_train):
negative,
replace=True, # sample with replacement
n_samples=len(positive), # match number in minority class
random_state=30) # reproducible results
random_state=30,
) # reproducible results
# combine minority and downsampled majority
downsampled = pd.concat([positive, neg_downsampled])
downsampled = downsampled.sample(frac=1)
X_train = downsampled.iloc[:, 0:-2]
y_train = downsampled.iloc[:, -1]
# check new class counts
print(downsampled.diagnosis.value_counts())
#graph
# graph
return X_train, y_train

def smote(X_train, y_train):
sm = SMOTE(random_state=30)
X_train, y_train = sm.fit_resample(X_train, y_train)
y_train = pd.DataFrame(y_train, columns=['diagnosis'])
y_train = pd.DataFrame(y_train, columns=["diagnosis"])
print(y_train.diagnosis.value_counts())
#graph
# graph
return X_train, y_train


Expand Down Expand Up @@ -207,10 +216,20 @@ def feat5():
df = data_load() # Loading Dataset into Dataframe
df = data_clean(df)
drop_cols = [
'radius_worst', 'texture_worst', 'perimeter_worst', 'area_worst',
'symmetry_worst', 'fractal_dimension_worst', 'perimeter_mean',
'perimeter_se', 'area_mean', 'area_se', 'concavity_mean',
'concavity_se', 'concave points_mean', 'concave points_se'
"radius_worst",
"texture_worst",
"perimeter_worst",
"area_worst",
"symmetry_worst",
"fractal_dimension_worst",
"perimeter_mean",
"perimeter_se",
"area_mean",
"area_se",
"concavity_mean",
"concavity_se",
"concave points_mean",
"concave points_se",
]
df_sf = df.drop(drop_cols, axis=1)
X, y = X_y_split(df_sf)
Expand All @@ -221,9 +240,7 @@ def feature():
"'\t The number '1' stands for 'ALL- FEATURES'. \n \t The number '2' stands for 'MEAN- FEATURES' . \n \t The number '3' stands for 'SQUARED- ERROR FEATURES'. \n \t The number '4' stands for 'WORST- FEATURES'. \n \t The number '5' stands for 'SELECTED- FEATURES'.'"
)
selection = input("\t Enter your choice of feature selection: \t")
feat_options = [
feat.feat1, feat.feat2, feat.feat3, feat.feat4, feat.feat5
]
feat_options = [feat.feat1, feat.feat2, feat.feat3, feat.feat4, feat.feat5]
return feat_options[int(selection) - 1]()


Expand Down Expand Up @@ -274,7 +291,7 @@ def rfc(dat):
def knn(dat):
# K-Nearest Neighbors
start = t.time()
knn = KNeighborsClassifier(n_neighbors=5, metric='minkowski', p=2)
knn = KNeighborsClassifier(n_neighbors=5, metric="minkowski", p=2)
model_knn = knn.fit(dat[0], dat[2])
pred = model_knn.predict(dat[1])
pred_prob = model_knn.predict_proba(dat[1])
Expand All @@ -284,7 +301,7 @@ def knn(dat):
def svc_l(dat):
# Linear SVM
start = t.time()
svc_l = SVC(kernel='linear', random_state=0, probability=True)
svc_l = SVC(kernel="linear", random_state=0, probability=True)
model_svc_l = svc_l.fit(dat[0], dat[2])
pred = model_svc_l.predict(dat[1])
pred_prob = model_svc_l.predict_proba(dat[1])
Expand All @@ -294,7 +311,7 @@ def svc_l(dat):
def svc_r(dat):
# Kernel SVM
start = t.time()
svc_r = SVC(kernel='rbf', random_state=0, probability=True)
svc_r = SVC(kernel="rbf", random_state=0, probability=True)
model_svc_r = svc_r.fit(dat[0], dat[2])
pred = model_svc_r.predict(dat[1])
pred_prob = model_svc_r.predict_proba(dat[1])
Expand All @@ -320,8 +337,13 @@ def gnb(dat):
def train_n_test():
ft = feat.feature()
modelsss = [
models.lr, models.dtc, models.rfc, models.knn, models.svc_l,
models.svc_r, models.gnb
models.lr,
models.dtc,
models.rfc,
models.knn,
models.svc_l,
models.svc_r,
models.gnb,
]
print(
"'\t The number '1' stands for 'LOGISTIC REGRESSION'. \n \t The number '2' stands for 'Decision Tree' . \n \t The number '3' stands for 'Random Forest Classifier'. \n \t The number '4' stands for 'KNN'. \n \t The number '5' stands for 'Liner SVM'. \n \t The number '6' stands for 'Kernal SVM'. \n \t The number '7' stands for 'Guassian NB'.'"
Expand All @@ -339,9 +361,13 @@ def train_n_test():
def performance():
out, y_test, mdl_selection = train_n_test()
models = [
"Logistic Regression", "Desicion Tree Classifier",
"Random Forest Classifier", "KNN", "Liner SVM", "Kernal SVM",
"Guassian NB"
"Logistic Regression",
"Desicion Tree Classifier",
"Random Forest Classifier",
"KNN",
"Liner SVM",
"Kernal SVM",
"Guassian NB",
]
cm_lr = confusion_matrix(y_test, out[2])
sns.heatmap(cm_lr, annot=True, cmap="Reds")
Expand All @@ -350,50 +376,60 @@ def performance():
rs = recall_score(y_test, out[2])
fs = f1_score(y_test, out[2])
ps = precision_score(y_test, out[2])
#Report Bar Plot
report = pd.DataFrame(
classification_report(y_test, out[2], output_dict=True))
# Report Bar Plot
report = pd.DataFrame(classification_report(y_test, out[2], output_dict=True))
rg = report.drop(report.index[3]).drop(report.columns[2:], axis=1)
plt.style.use('seaborn')
rg.plot(kind='bar', color=["red", "salmon"])
plt.style.use("seaborn")
rg.plot(kind="bar", color=["red", "salmon"])
plt.title("Classification Report of {}".format(models[mdl_selection - 1]))
plt.legend(report.columns,
ncol=2,
loc="lower center",
bbox_to_anchor=(0.5, -0.3))
plt.legend(report.columns, ncol=2, loc="lower center", bbox_to_anchor=(0.5, -0.3))
plt.yticks(np.arange(0, 1.05, step=0.05))
print(
'\n\t The accuracy score of {} with given parameters is: {}%.'.format(
models[mdl_selection - 1], acs * 100))
print('\n\t The recall score of {} with given parameters is: {}%.'.format(
models[mdl_selection - 1], rs * 100))
"\n\t The accuracy score of {} with given parameters is: {}%.".format(
models[mdl_selection - 1], acs * 100
)
)
print(
"\n\t The recall score of {} with given parameters is: {}%.".format(
models[mdl_selection - 1], rs * 100
)
)
print(
"\n\t The precision score of {} with given parameters is: {}%.".format(
models[mdl_selection - 1], ps * 100
)
)
print(
'\n\t The precision score of {} with given parameters is: {}%.'.format(
models[mdl_selection - 1], ps * 100))
print('\n\t The F1 score of {} with given parameters is: {}%.'.format(
models[mdl_selection - 1], fs * 100))
"\n\t The F1 score of {} with given parameters is: {}%.".format(
models[mdl_selection - 1], fs * 100
)
)
print(
'\n\t The training and testing time taken by {} with given parameters is: {} seconds.'
.format(models[mdl_selection - 1], out[1]))
"\n\t The training and testing time taken by {} with given parameters is: {} seconds.".format(
models[mdl_selection - 1], out[1]
)
)
prob = out[3]
prob = prob[:, 1]
#ROC
# ROC
false_pos, true_pos, thresh = roc_curve(y_test, prob, pos_label=1)
auc_score = roc_auc_score(y_test, prob)
rand_pr = [0 for i in range(len(y_test))]
p_fpr, p_tpr, _ = roc_curve(y_test, rand_pr, pos_label=1)
plt.figure()
plt.style.use('seaborn')
plt.plot(false_pos,
true_pos,
linestyle='--',
color='orange',
label=models[mdl_selection - 1])
plt.plot(p_fpr, p_tpr, linestyle='--', color='green')
plt.title('ROC Curve')
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.legend(loc='best')
plt.style.use("seaborn")
plt.plot(
false_pos,
true_pos,
linestyle="--",
color="orange",
label=models[mdl_selection - 1],
)
plt.plot(p_fpr, p_tpr, linestyle="--", color="green")
plt.title("ROC Curve")
plt.xlabel("False Positive Rate")
plt.ylabel("True Positive Rate")
plt.legend(loc="best")

return out[0], out[2], auc_score

Expand Down
Loading

0 comments on commit c4047eb

Please sign in to comment.