A3_STUDENTNUMBER.tex

\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage[a4paper, margin=2.5cm]{geometry}
\usepackage{float}
\usepackage{graphicx}

\title{MLCourse-LU}
\author{YOUR STUDENT NUMBER HERE}
\date{Assignment 3}

\begin{document}

% REMINDER: you don't have to write a long text, just concise answers/analyses of the tasks and your findings. There is no page limit but that's only to give you room for your figures.

% You may find the overleaf documentation on multifigures useful:
% https://www.overleaf.com/learn/latex/How_to_Write_a_Thesis_in_LaTeX_(Part_3):_Figures,_Subfigures_and_Tables

% REMINDER: if you use external sources, be sure to cite them!

\maketitle

\section{Visualizing clustering results}
% TODO: side by side visualization of KMeans, KMedoids and ground truth
% TODO: ensure the centroids are marked in the plots of the models
% TODO: ensure your plots have all the things good plots should have
% TODO: brief analysis of what you can make out from the plot about how well the models are doing


\section{$k-$means vs. $k-$medoids}
% TODO: scatterplots with the best results you achieved with KMeans and KMedoids on this dataset
% TODO: report homogeneity, completeness and the k you used for both models
% TODO: analyze which model works best
% TODO: analyze WHY you think it works best
% TODO: come up with a situation for both models when that model would work better than the other


\section{Using the silhouette method to compare $k-$means vs. $k-$medoids on unlabeled data}
% TODO: two pairs of plots, each showing silhouette scores + clusterings for KMeans and KMedoids
% TODO: report on how they score
% TODO: analyze which one is doing better (think about the scatterplot)


\section{The elbow method}
% TODO: your elbow plots
% TODO: your choice of $k$ for both datasets and why you think that's the best choice

\section{Generating difficult clusters}
% TODO: a figure with two plots, one with ground truth and one with KMedoids clusters
% TODO: brief explanation of how you went about generating the dataset
% TODO: analysis why this dataset is easier for humans than KMedoids to cluster


\end{document}