kanser_analizi.py

"""
data set and problem desc
-data set name
-feature
-class
-class distribition
-missing value and outlier,
-feature expession(kanser hücresinin 32 değişkeninin olmasına rağmen bunu sıkıştırıp 10 adete indiriyo)

"""

import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
from matplotlib.colors import ListedColormap

from sklearn.neighbors import LocalOutlierFactor
#from sklearn.f import train_test_split, GridSearchCV
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.metrics import accuracy_score, confusion_matrix
from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsClassifier
from sklearn.neighbors import NeighborhoodComponentsAnalysis


#warning classlarını engellemek için warnings kütüphanesinden araralanıyoruz
import warnings
warnings.filterwarnings("ignore")

data = pd.read_csv("data.csv")                                                                                                                                                       
data.drop(['Unnamed: 32','id'], inplace=True,axis=1)
# Yeni bir isimlendirme yaptık
data = data.rename(columns={"diagnosis": "target"})
#M ve B olan (kötü huy iyi huy özelliklerini int çevirdik o kötü huy 1 iyi huy demektir)
data["target"]=[0 if i.strip()=="M"else 1 for i in data.target]
#sns.countplot(data["target"])
sns.countplot(x="target", data=data)
plt.show()
print("data target deger sayısı:",data.target.value_counts())
#M ve B olan (kötü huy iyi huy özelliklerini int çevirdik o kötü huy 1 iyi huy demektir)
#data["target"]=[0 if i.strip()=="M"else 1 for i in data.target]

print("data shape:",data.shape)
data.info()
"""
ecploarotary data analyisc
standardization
missing value:none

"""

corr_matrix=data.corr()

sns.clustermap(corr_matrix, annot=True,fmt=".2f")
plt.title("standartazation figure")
plt.show()

threshold=0.5
filtre=np.abs(corr_matrix["target"])>threshold
corr_features=corr_matrix.columns[filtre].tolist()
sns.clustermap(data[corr_features].corr(), annot=True,fmt=".2f")
plt.title("filtrelenmiş figür")
plt.show()

#bazı ilişkiler
#box plotlar
data_melted=pd.melt(data,id_vars="target",
                    var_name="features",
                    value_name="value")
plt.figure()
sns.boxplot(x="features",y="value",hue="target",data=data_melted)
plt.xticks(rotation=90)
plt.title("Benzer özellikler figürü")
plt.show()
"""
normalizasyon

"""
#pair plot
sns.pairplot(data[corr_features],diag_kind="kde",markers="+",hue="target")
plt.show()
# outlier
y=data.target
x=data.drop(["target"],axis=1)
columns=x.columns.tolist()
#LOf

clf=LocalOutlierFactor()
y_pred=clf.fit_predict(x)
x_score=clf.negative_outlier_factor_

outlier_score=pd.DataFrame()
outlier_score["score"]=x_score
#threshold işlemini yaptık
threshold=-2.5
filtre_2=outlier_score["score"]<threshold
outlier_index=outlier_score[filtre_2].index.tolist()


plt.figure()
plt.scatter(x.iloc[outlier_index,0],x.iloc[outlier_index,1],color="blue",s=50,label="outliers")
plt.scatter(x.iloc[:,0],x.iloc[:,1],color="k",s=3,label="Data Points")

radius=(x_score.max()-x_score)/(x_score.max()-x_score.min())
outlier_score["radius"]=radius
plt.scatter(x.iloc[:,0],x.iloc[:,1],s=1000*radius,edgecolors="r",facecolors="none",label="Outlier Scores")
plt.legend()
plt.show()
#drop outliers
x=x.drop(outlier_index)
y=y.drop(outlier_index).values
#train-test-split
test_size=0.3
x_train,x_test,y_train,y_test=train_test_split(x,y,test_size=test_size,random_state=42)

"""
standardization(ready for training)=datamızı etkileyen faktörler arasında ki farklılıklardan dolayı bu işleme ihtiyaç duymaktayız


"""
scaler=StandardScaler()
x_train=scaler.fit_transform(x_train)
x_test=scaler.transform(x_test)

x_train_data_frame=pd.DataFrame(x_train,columns=columns)
x_train_data_frame_describe=x_train_data_frame.describe()
x_train_data_frame["target"]=y_train
#box plotlar
data_melted=pd.melt(x_train_data_frame,id_vars="target",
                    var_name="features",
                    value_name="value")

plt.figure()
sns.boxplot(x="features",y="value",hue="target",data=data_melted)
plt.title("box plot")
plt.xticks(rotation=90)
plt.show()

#KNN algoritması
knn=KNeighborsClassifier(n_neighbors=2)
knn.fit(x_train,y_train)
y_pred=knn.predict(x_test)
cm=confusion_matrix(y_test,y_pred)
acc=accuracy_score(y_test,y_pred)
score=knn.score(x_test,y_test)
print("başarı sonucu:",score)
print("basit knn algoritması:",acc)
print(cm)
#en iyi k değerini bulmak için fonksiyon yapıyoz
def knn_en_iyi_parametre(x_train,x_test,y_train,y_test):
    k_range=list(range(1,31))
    weight_options=["uniform","distance"]
    print()
    param_grid=dict(n_neighbors=k_range,weights=weight_options)
    knn=KNeighborsClassifier()
    grid=GridSearchCV(knn, param_grid,cv=10,scoring="accuracy")
    grid.fit(x_train,y_train)
    print("en iyi deneme skoru:{} with parametres:{}".format(grid.best_score_,grid.best_params_))
    print()
    
    knn=KNeighborsClassifier(**grid.best_params_)
    knn.fit(x_train,y_train)
    
    y_pred_test=knn.predict(x_test)
    y_pred_train=knn.predict(x_train)
    
    cm_test=confusion_matrix(y_test,y_pred_test)
    cm_train=confusion_matrix(y_train,y_pred_train)

    acc_test=accuracy_score(y_test,y_pred_test)
    acc_train=accuracy_score(y_train,y_pred_train)
    print("test skoru:{},train skoru:{}".format(acc_test,acc_train))
    print()
    print("cm test:",cm_test)
    print("cm train:",cm_train)
    print()
    return grid
    
grid=knn_en_iyi_parametre(x_train,x_test,y_train,y_test)
    
#PCA
scaler_1=StandardScaler()
x_scaled=scaler_1.fit_transform(x)  
pca=PCA(n_components=2)
pca.fit(x_scaled)
x_radyus_pca=pca.transform(x_scaled)
pca_data=pd.DataFrame(x_radyus_pca,columns=["p1","p2"])
pca_data["target"]=y
sns.scatterplot(x="p1",y="p2",hue="target",data=pca_data)
plt.title("PCA GRAFİĞİ:P1 ve P2")
    
x_train_pca,x_test_pca,y_train_pca,y_test_pca=train_test_split(x_radyus_pca,y,test_size=test_size,random_state=42)

grid_pca=knn_en_iyi_parametre(x_train_pca, x_test_pca, y_train_pca, y_test_pca)
#visualize
cmap_light=ListedColormap(['orange','cornflowerblue'])
cmap_bold=ListedColormap(['darkorange','darkblue'])

h=.05#adım boyutları
X=x_radyus_pca
x_min,x_max=X[:,0].min()-1,X[:,0].max()+1    
y_min,y_max=X[:,1].min()-1,X[:,1].max()+1    
xx,yy=np.meshgrid(np.arange(x_min,x_max,h),np.arange(y_min,y_max,h))
    
Z=grid_pca.predict(np.c_[xx.ravel(),yy.ravel()])
Z=Z.reshape(xx.shape)
plt.figure()
plt.pcolormesh(xx,yy,Z,cmap=cmap_light)
plt.scatter(X[:,0],X[:,1],c=y,cmap=cmap_bold,edgecolors="k",s=20)
plt.xlim(xx.min(),xx.max())
plt.ylim(yy.min(),yy.max())
plt.title("%i-class classicifation(k=%i,weights='%s')"%
          (len(np.unique(y)),grid_pca.best_estimator_.n_neighbors,grid_pca.best_estimator_))   
 
#  NCA
nca=NeighborhoodComponentsAnalysis(n_components=2,random_state=42) 
nca.fit(x_scaled,y)

x_radyus_nca=nca.transform(x_scaled)
nca_data=pd.DataFrame(x_radyus_nca,columns=["p1","p2"])
nca_data["target"]=y
sns.scatterplot(x="p1",y="p2",hue="target",data=nca_data)
plt.title("NCA:p1 ve p2 grafiği")

x_train_nca,x_test_nca,y_train_nca,y_test_nca=train_test_split(x_radyus_nca,y,test_size=test_size,random_state=42)

grid_nca=knn_en_iyi_parametre(x_train_nca, x_test_nca, y_train_nca, y_test_nca)

#nca visualize
cmap_light=ListedColormap(['orange','cornflowerblue'])
cmap_bold=ListedColormap(['darkorange','darkblue'])

h=.2#adım boyutları
X=x_radyus_nca
x_min,x_max=X[:,0].min()-1,X[:,0].max()+1    
y_min,y_max=X[:,1].min()-1,X[:,1].max()+1    
xx,yy=np.meshgrid(np.arange(x_min,x_max,h),np.arange(y_min,y_max,h))
    
Z=grid_nca.predict(np.c_[xx.ravel(),yy.ravel()])
Z=Z.reshape(xx.shape)
plt.figure()
plt.pcolormesh(xx,yy,Z,cmap=cmap_light)
plt.scatter(X[:,0],X[:,1],c=y,cmap=cmap_bold,edgecolors="k",s=20)
plt.xlim(xx.min(),xx.max())
plt.ylim(yy.min(),yy.max())
plt.title("%i-class classicifation(k=%i,weights='%s')"%
          (len(np.unique(y)),grid_nca.best_estimator_.n_neighbors,grid_nca.best_estimator_))