CODEWITHSUNDEEP
pythonlogistic-regression
boo
boo

Reputation: 35

Why does my accuracy go over 100% on my logistic regression model?

I am working on a dataset that is a collection of several medical predictor variables and one target variable, used to classify whether a patient has diabetes or not. I am building my model without using scikit learn / sklearn library. I have attached the link to dataset below.

https://www.kaggle.com/uciml/pima-indians-diabetes-database

I have trained and tested mode but I keep getting over 100% accuracy. I am very beginner in this field, therefore I apologize if I have made silly mistakes. Below is my code ( and I only use Glucose and DiabetesPedigreeFunction) to classify.

    import numpy as np
    import pandas as pd
    import matplotlib.pyplot as plt
    %matplotlib inline
    df = pd.read_csv('diabetes.csv')
    df.head()

    df.drop(['BloodPressure', 'SkinThickness', 'Insulin', 'BMI', 
    'Pregnancies', 'Age'], axis = 1, inplace=True)
    df

    positive = df[df['Outcome'].isin([1])]
    negative = df[df['Outcome'].isin([0])]

    fig, ax = plt.subplots(figsize=(12,8))
    ax.scatter(positive['DiabetesPedigreeFunction'],positive['Glucose'], 
    s=50, c='b', marker='o', label='Diabetes')
    ax.scatter(negative['DiabetesPedigreeFunction'],negative['Glucose'], 
    s=50, c='r', marker='x', label='Not Diabetes')
    ax.legend()

    def sigmoid(x):
      return 1/(1 + np.exp(-x))

    
    nums = np.arange(-10, 10, step=1)
    fig, ax = plt.subplots(figsize=(12,8))
    ax.plot(nums, sigmoid(nums), 'r')

    def cost(theta, X, y):
        theta = np.matrix(theta)
        X = np.matrix(X)
        y = np.matrix(y)
        first = np.multiply(-y, np.log(sigmoid(X * theta.T)))
        second = np.multiply((1 - y), np.log(1 - sigmoid(X * theta.T)))
        return np.sum(first - second) / (len(X))

        X.shape, theta.shape, y.shape
        
        cost(theta, X, y)

        def gradient(theta, X, y):
            theta = np.matrix(theta)
            X = np.matrix(X)
            y = np.matrix(y)

            parameters = int(theta.ravel().shape[1])
            grad = np.zeros(parameters)

            error = sigmoid(X * theta.T) - y

            for i in range(parameters):
                term = np.multiply(error, X[:,i])
                grad[i] = np.sum(term) / len(X)

            return grad

    gradient(theta, X, y)
    import scipy.optimize as opt
    result = opt.fmin_tnc(func=cost, x0=theta, fprime=gradient, args=(X, 
    y))

    cost(result[0], X, y)

    def predict(theta, X):
        probability = sigmoid(X * theta.T)
        return [1 if x >= 0.5 else 0 for x in probability]

    theta_min = np.matrix(result[0])
    predictions = predict(theta_min, X)
    correct = [1 if ((a == 1 and b == 1) or (a == 0 and b == 0)) else 0 
    for (a, b) in zip(predictions, y)]
    accuracy = (sum(map(int, correct)) % len(correct))
    print ('accuracy = {}%'.format(accuracy))

my accuracy is 574%. I need some feedback. Thanks in advance.

Upvotes: 0

Views: 229

Answers (1)

akocz
akocz

Reputation: 138

You used mod instead of division.

Accuracy should be computed like this:

accuracy = sum(correct) / len(correct)

Upvotes: 1

Related Questions