$\backslash$table$[[,$Animal,Weight $($kg$),\backslash$table$[[$Top Speed$],[(k\frac{m}{h})$

$..$AI and ML Assignment

You are part of an organizing team to design a large open$-$range Zoo for the animals listed in the below table.

a

The zoo consists of sections; each section includes a group $($i$.$e$.$ cluster$)$ of animals. These sections will feature

expansive open$-$air houses, catering to either a single species or multiple species.

The z$0$o could include multiple water bodies, wooded areas etc. See Figure of an example zoo.

The ultimate goal is to craft a zoo that appropriate for the animals' physical and mental well$-$being while

is

simultaneously attractive to our visitors.

WARIE

TS & BEAR

ALISTIN ZOC

Your tasks are:

Cluster the animals into groups to assist in designing the Park.

Label each cluster with proper name i$.$e$.$ "large funny animals".

To achieve this, write a python program to find the optimum clustering for the animals listed in the

table. The python program performs the following tasks:

Read animal from the Table. Use Panda library.

Write a function for K$-$means algorithm.

O

Apply K$-$means algorithm repeatedly with different K$\text{'}$s$.$

O

Plot the elbow graph. Find optimum clustering. Give meaningful labels to the clusters

Write a report, and draw a map of the zoo, and show the sections and animal houses.

Important Notes:

For non$-$numerical feature$/$ trait, you use multiple methods:

$$ Data Transformation:

For example, One$-$ Hot Encoding: Convert data into binary vectors. Each category gets its

own column with a $1$ for belonging and $0$ for not belonging. This allows k$-$means to work

with the data as if it were numerical.

Feature Engineering: create new numerical features from the non$-$numerical data $($this might be

better$).$

Optionally, you can apply PCA prior to clustering.

Rules:

Deadline: $18/04$

This assignment is $15\%$ of your grade

is

Teams: $3$ students per team.

Submission include:

Python Code

O

$$ Report: assumptions$/$plots$/$labels$/$map

Grading is based on:

Correct program.

O

Good report.

O

The best clustering solution.

Demo: presentation and code modification by everyone in the team.