Build & Evaluate Classifier for Diabetes

Objectives

We will use iris dataset to create a classifier then evaluate it using various metrics

Here is a list of Tasks we’ll be doing:

  1. Use Pandas to load data sets.
  2. Identify the target and features.
  3. Use Logistic Regression to build a classifier.
  4. Use metrics to evaluate the model.
  5. Make predictions using a trained model.

Setup

We will be using the following libraries:

  • pandas for managing the data.
  • sklearn for machine learning and machine-learning-pipeline related functions.

Install Libraries

# All Libraries required for this lab are listed below. Already installed so this code block is commented out
pip install pandas==1.3.4
pip install scikit-learn==1.0.2

Suppress Warnings

To suppress warnings generated by our code, we’ll use this code block

# To suppress warnings generated by the code
def warn(*args, **kwargs):
    pass
import warnings
warnings.warn = warn
warnings.filterwarnings('ignore')

Import Libraries

import pandas as pd
from sklearn.linear_model import LogisticRegression

#import functions for train test split
from sklearn.model_selection import train_test_split


# functions for metrics

from sklearn.metrics import confusion_matrix
from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.metrics import f1_score

Data - Task 1

Load

# the data set is available at the url below.
URL = "https://cf-courses-data.s3.us.cloud-object-storage.appdomain.cloud/IBM-BD0231EN-SkillsNetwork/datasets/diabetes.csv"

# using the read_csv function in the pandas library, we load the data into a dataframe
df = pd.read_csv(URL)
# you can sample the data with df.sample(5)


Pregnancies

Glucose

BloodPressure

SkinThickness

Insulin

BMI

DiabetesPedigreeFunction

Age

Outcome

145

0

102

75

23

0

0.0

0.572

21

0

470

1

144

82

40

0

41.3

0.607

28

0

142

2

108

52

26

63

32.5

0.318

22

0

633

1

128

82

17

183

27.5

0.115

22

0

150

1

136

74

50

204

37.4

0.399

24

0 

df.shape
(768, 9)

Plot Data

  • Let’s plot the types and count of Outcome
  • There are 500 people without diabetes and 268 people with diabetes in this dataset.
df.Outcome.value_counts()
Outcome
0    500
1    268
Name: count, dtype: int64
df.Outcome.value_counts().plot.bar()

Define Targets/Features - Task 2

Target

In Classification models, the Target (Y) is the value our machine learning model needs to classify

So, in this example we are trying to classify the Outcome

Features

The feature(s) is/are the data columns we will provide our model as input from which our model learns from

In our example let’s provide the model with these Features, and see how accurate it will be in predicting the list is in the code

# Target is our Y
Y = df["Outcome"]

# Features is our X
X = df[['Pregnancies', 'Glucose', 'BloodPressure', 'SkinThickness', 'Insulin',
       'BMI', 'DiabetesPedigreeFunction', 'Age']]

Split Dataset - Task 3

  • We now split the data at 70/30 ratio, 70 training, 30 testing
  • The random_state variable controls the shuffling applied to the data before applying the split.
  • Pass the same integer for reproducible output across multiple function calls
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.30, random_state=40)

Build & Train Classifier - Task 4

Logistic Regression Model

Create a Logistic Regression model

classifier = LogisticRegression()

Train Logistic Regression Model

Response will be:

  • LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True, intercept_scaling=1, max_iter=100, multi_class=‘warn’, n_jobs=None, penalty=‘l2’, random_state=None, solver=‘warn’, tol=0.0001, verbose=0, warm_start=False)
classifier.fit(X_train, Y_train)
LogisticRegression()
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

Evaluate Model - Task 5

Now that the model has been trained on the data/features provided above, let’s evaluate it

Score

The higher the better

#Higher the score, better the model.
classifier.score(X_test, Y_test)
0.7229437229437229

Predict

In order to calculate the score metrics we need two values:

  • The original values from the test data set and we’ll compare to the predicted values
  • The predicted values which are the results of the model
original_values = Y_test
predicted_values = classifier.predict(X_test)

Precision

The higher the value the better the model

precision_score(original_values, predicted_values)
np.float64(0.6865671641791045)

Recall

The higher the value the better the model

recall_score(original_values, predicted_values)
np.float64(0.5168539325842697)

F1 Score

The higher the value the better the model

f1_score(original_values, predicted_values)
np.float64(0.5897435897435898)

Confusion Matrix

This can be used to manually calculate various metrics

confusion_matrix(original_values, predicted_values)
array([[121,  21],
       [ 43,  46]])