Aim

This project intends to show the capabilities of machine learning using the K Nearest Neighbor Classification Model on Iris Dataset. The slides are located in the file 'ML workshop.pdf'. They authors of the slides are: Joe, Shreya and Rushik, Lily and Minjun of UNSW DataSoc.

To start with Iris Classifier:

1. Download iris_classifer.ipynb file.
2. Install the scikit-learn library using pip install -U scikit-learn
3. Run the cells
4. Go to the section with the title: "Classify flowers using their Measurements"
5. Classify Mystery Flower 1, Mystery Flower 2 and Mystery Flower 3 using KNN model :)
6. Check if the model classifies the Flowers correctly by double-clicing on the images and seeing the image names.

To start with Pokemon Classifier:

Function

Classifies the Pokemon into into Bug or Rock types.

Instruction

1. Download Pokemon.ipynb file.
2. Install the scikit-learn library using pip install -U scikit-learn
3. Run all cells
4. Go to the markdown with the title: "Testing Model with Real Pokemon (Features: weight and height)".
5. Run the python cells and enter the weight and height of a pokemon to see if it can classify the type correctly :)
6. Go to the markdown with the title: "Testing Model with Real Pokemon (Features: Attack and HP)".
7. Run the python cells and enter the Attack and HP of a pokemon to see if it can classify the type correctly :)

Note

Model 1 that uses weight and height has an accuracy of 63%. Hence, it can get things wrong sometimes

• Really light or small pokemon tend to be classified as bug-type

Note

Model 2 that uses attack and hp has an accuracy of 83%. Hence, it can get things wrong sometimes

• Really high hp pokemon tend to be classified as rock-type