Prostate cancer prediction model.ipynb

Kernel: Python [default]

import pandas as pd
from pandas import Series,DataFrame
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
import numpy as np

data = pd.read_csv('Prostate_Cancer.csv',index_col=0)

data.head()

data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 100 entries, 1 to 100
Data columns (total 9 columns):
diagnosis_result     100 non-null object
radius               100 non-null int64
texture              100 non-null int64
perimeter            100 non-null int64
area                 100 non-null int64
smoothness           100 non-null float64
compactness          100 non-null float64
symmetry             100 non-null float64
fractal_dimension    100 non-null float64
dtypes: float64(4), int64(4), object(1)
memory usage: 7.8+ KB

Since the diagnosis_result column is in text , we need to convert it to binary so that it can be fed into the algorithm.

new_df=pd.get_dummies(data,columns=['diagnosis_result'],drop_first=True)

new_df.head()

#using StandardScaler
from sklearn.preprocessing import StandardScaler
scaler=StandardScaler()
scaler.fit(new_df.drop('diagnosis_result_M',axis=1))
#removing the last column of the dataframe as this table will be used for the feature matrix
scaled_features=scaler.transform(new_df.drop('diagnosis_result_M',axis=1))
new_data=pd.DataFrame(scaled_features,columns=new_df.columns[:-1])

new_data.head()

#feature matrix
X=new_data

#target matrix
y=new_df['diagnosis_result_M']

from sklearn.cross_validation import train_test_split

#allocating 33% of the dataset for testing
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33)

from sklearn.neighbors import KNeighborsClassifier

# random value of n_neighbors, we will find a better value of k later.
knn = KNeighborsClassifier(n_neighbors=1)

knn.fit(X_train,y_train)

KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
           metric_params=None, n_jobs=1, n_neighbors=1, p=2,
           weights='uniform')

pred=knn.predict(X_test)

from sklearn.metrics import classification_report,confusion_matrix

print(classification_report(y_test,pred))

             precision    recall  f1-score   support

        0.0       0.79      0.69      0.73        16
        1.0       0.74      0.82      0.78        17

avg / total       0.76      0.76      0.76        33

#since a precision of only 76%, we'll try to use another value for k

error_rate=[]
for i in range(1,40):
    knn=KNeighborsClassifier(n_neighbors=i)
    knn.fit(X_train,y_train)
    pred_i=knn.predict(X_test)
    error_rate.append(np.mean(pred_i!=y_test))

plt.figure(figsize=(10,6))
plt.plot(range(1,40),error_rate)

[<matplotlib.lines.Line2D at 0xdfe6400>]

It can be seen from the above plot that the error is least when k=4

knn=KNeighborsClassifier(n_neighbors=4)

knn.fit(X_train,y_train)

KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
           metric_params=None, n_jobs=1, n_neighbors=4, p=2,
           weights='uniform')

new_pred=knn.predict(X_test)

print(classification_report(y_test,new_pred))

             precision    recall  f1-score   support

        0.0       0.92      0.75      0.83        16
        1.0       0.80      0.94      0.86        17

avg / total       0.86      0.85      0.85        33

The model is now 86% precise, compared to 76% when a random value of K was chosen