Unsupervised Learning - Kmeans Week 11 Challenge
- How does similarity is calculated if data is categorical in nature
Similarity of categorical variables can be find my using following 3 correlations
1->Tetrachoric correlation:
its used to calculate the correlation between binary variables. Its value ranges between
-1 to +1.
-1 - Negative correlation,
0 - no correlation and
1 - positive correlation.
Eg: gender vs election party preference2->Polychoric correlation:
It is used to calculate the correlation between 2 ordinal categorical variables. Ordinal refers to the variables having
a values have a natural order.
Its value ranges from -1 to +1
Eg:Movie name and ratings3->Cramer's V:
Its used to calculate the correlation between nominal variables(which doesn't have any natural order within them)
Value ranges from 0 to 1.
0 - no relationship between the variables
1 - strong relationship between the variables - Implement K-means for “Car dataset” and come up with the business insightsKindly find the files. Clustering was done in "Automobile_1985_KMEANS_Clustering_Class" file and analysis clusters were performed in 2 ways
- with outliers : in file "Automobile_1985_Insights_from_clusters_derived_from_KMeans_with_outliers"
- without outliers : in file "Automobile_1985_Insights_from_clusters_derived_from_KMeans_without_outliers"
- For conclusion refer to the text file attached.