Diebetes-Predicting-ML

Objective

To predict whether the person is suffering from Diabetes Mellitus(DM) or not

Data Collection

We are using dataset from the National Institute of Diabetes and Kidney Disease. All data are female and > 21 years old of PIMA.

Why Pima?

Pima, North American Indians who traditionally lived along the Gila and Salt rivers in Arizona, U.S., in what was the core area of the prehistoric Hohokam culture. They have the highest prevalence ever recorded in the world¹.

Approach

Logistic regression: 0 for not suffering from DM; 1 for suffering from DM

Code

import pandas as pd 
import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np
import sklearn as skl
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn import preprocessing

Make header and load data

column_names = ["Pregnant", "Glucose", "BP", "Skin", "Insulin", "BMI", "Pedigree", "Age", "DM"]
data_set = pd.read_csv("pima_diabetes.csv", header = None, names=column_names, skiprows=(0,0))
data_set.head()

Convert data from string to number

covert_col = ["Pregnant", "Insulin", "BMI", "Age", "BP", "Pedigree"]
for col in covert_col:
    data_set[col] = pd.to_numeric(data_set[col])

Feature selection

Select our independent value (x) and dependent value (y)

feature_col = ["Pregnant", "Insulin", "BMI", "Age", "Glucose", "BP" ,"Pedigree"]
X = data_set[feature_col]
y = data_set.DM

Check correlation

plt.figure(figsize=(40, 30))
coor_range = corr[(corr >= 0.3) | (corr <= -0.1)]
sns.heatmap(coor_range, vmax=0.8, linewidths=0.01, square=True, annot=True, cmap='GnBu', linecolor="white", cbar_kws={'label': 'Feature Correlation Color'})
plt.title('Correlation between features of Pima Datasets')
plt.ylabel("Feature Values on Y axis")
plt.xlabel("Feature Values on X axis")

Partition dataset to test and training set

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20, random_state = 42) 

Applying logistic regression algorithm

logistic_function=LogisticRegression()
logistic_function.fit(X_train, y_train)
y_prediction = logistic_function.predict(X_test)

Model Evaluation using Confusin Matrix

from sklearn import metrics
cnf_matrix_evaluation = metrics.confusion_matrix(y_test, y_prediction)
cnf_matrix_evaluation

Data Visualization

class_names = [0, 1]
fig, ax = plt.subplots()
tick_marks = np.arange(len(class_names))
plt.xticks(tick_marks, class_names)
plt.yticks(tick_marks, class_names)

sns.heatmap(pd.DataFrame(cnf_matrix_evaluation), annot=True, cmap="YlGnBu", fmt='g')
ax.xaxis.set_label_position("top")
plt.tight_layout()
plt.title("Confusion Matrix: Diabetes Patient", y=1.1)
plt.ylabel("Acutal DMs")
plt.xlabel("Predicted DMs")

ROC curve to check performance

y_prediction_probability = logistic_function.predict_proba(X_test)[::,1]
fpr, tpr, _ = metrics.roc_curve(y_test, y_prediction_probability)
auc = metrics.roc_auc_score(y_test, y_prediction_probability)
plt.plot(fpr, tpr, label="data 1, auc=" + str(auc))
plt.legend(loc=4)
plt.show()

Reference

¹ Gohdes D: Diabetes in North American Indians and Alaska natives. In Diabetes in America. Washington, DC, U.S. Govt. Printing Office, 1995, p. 1683–1701 (NIH publ. no. 95-1468). Google Scholar

Credits

1. Pima Indians Diabetes - EDA & Prediction (0.906)