K-NN Algorithm and Linear Regression Model

Getting Started

Installing Python

To use the script, you will need to install Python 3.6.x and add to path:

• Python 3.6.x

Installing Dependencies

After cloning the project, go to the root directory:

Install the dependent libraries by typing the following in cmd/terminal:

$ pip install -r requirements.txt

Starting the Script

To run the script, go to the root directory and run python in cmd/terminal and type the following in the python console:

>>> from knn_main import *

Note: ensure that python refers to Python 3.6.x

Running K-NN on Regression Dataset (Q1)

To run the k-nn algorithm on regression dataset [mauna_loa,rosenbrock,pumadyn32nm], type the following in the python console:

>>> run_Q1(k_range=[1,31])

Note: k_range takes a list containing lower bound and upper bound of k values

Running K-NN on Classification Dataset (Q2)

To run the k-nn algorithm on classification dataset [iris,mnist_small], type the following in the python console:

>>> run_Q2(k_range=[1,31])

Note: k_range takes a list containing lower bound and upper bound of k values

Running K-NN with KD Tree and Compare Performance (Q3)

Type the following in the python console:

>>> run_Q3(d=list(range(2,10)))

Note: d takes a list containing values of dimension numbers

Running Linear Regression with SVD on All Dataset (Q4)

Type the following in the python console:

>>> run_Q4()

Built With

• numpy - all variables are numpy arrays

• sklearn - kd tree data structure