8 kNN Model - Joooy | Zeng

1 Similarity Based Learning

Compute the distance matrices between objects

2 k Nearest Neighbor (kNN) Model

2.1 Pros and Cons of kNN

Pros	Cons
Simple and Effective	Does not produce a model, limiting the ability to understand how the features are related to the class
Makes no assumption about the underlying data distribution Non-parametric	Requires selection of an appropriate value of ‘k’

2.2 Example

3 kNN Model Assessment

4 Data Normalization: Standardization & Scaling

Suppose we have 2 data items

• Height: varies from 4 – 7 feet
• Net Worth: 100B

If we use both the variables in a model

• Net Worth will dominate because it contains large values

Solution

• Standardize
• Scale

4.1 Data Standardization and Scaling

5 kNN in Python

03_kNN_Optimum_k_Iris_Sci_L_ConMat.ipynb