Unit 5: e-Portfolio Activity: Jaccard Coefficient Calculations

Unit 5: e-Portfolio Activity: Jaccard Coefficient Calculations on Github

The calculation of the Jaccard coefficient helps a assess the similarity between test results of Jack, Mary, and Jim. The Jaccard formula quantifies the similarity by evaluating the ratio of the intersection of two sets to their union. This points out the overlap in their test results. The resulting Jaccard coefficient has values ranging from 0 (no similarity) to 1 (complete similarity). This analysis is useful in contexts such as clustering, recommendation systems, or social network analysis.

One-hot encoding was necessary to convert the categorical data: ‘Y’, ‘N’, ‘P’, and ‘A’ into numerical vectors. After this was done, I used Scikit-learn to calculate the coefficients. Based on the results, Jack and Mary’s results are more similar.

---

Unit 5: Jaccard Coefficient Calculations¶

Import from scikit learn

In [ ]:

from sklearn.metrics import jaccard_score
from sklearn.preprocessing import OneHotEncoder
import numpy as np

Test results for jack, mary, jim

In [ ]:

jack = ['Y','N','P','N','N','A']
jim = ['Y','N','P','A','P','N']
mary = ['Y','P','N','N','N','A']

One hot encode - convert categorical value to vector format

In [ ]:

data = np.array([jack,jim,mary])
# Create OneHotEncoder instance
encoder = OneHotEncoder()
# Fit and transform the data
encoded_data = encoder.fit_transform(data)

jack_vector = encoded_data[0]
jim_vector = encoded_data[1]
mary_vector = encoded_data[2]

Get jaccard score using scikit

In [ ]:

jack_mary_jaccard_coefficient = jaccard_score(jack_vector, mary_vector, average='macro')
jack_jim_jaccard_coefficient = jaccard_score(jack_vector, jim_vector, average='macro')
jim_mary_jaccard_coefficient = jaccard_score(jim_vector, mary_vector, average='macro')

jack_mary_jaccard_coefficient, jack_jim_jaccard_coefficient, jim_mary_jaccard_coefficient

Out[ ]:

(0.36363636363636365, 0.2727272727272727, 0.09090909090909091)