report_phase2.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fifa Position Predictor\n",
    "### Phase 2\n",
    "- Ali Elmancy 201901993\n",
    "- Ahmed Abdou 202005605\n",
    "- Ahmed Terchoun 201802794"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Note\n",
    "#### The entire data processing and data visualization is done in the first phase, this phase is dedicated to the machine learning part."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Abstract\n",
    "In phase 2 of the project we have utilized the same dataset used in phase 1, the cleaned, the processed and the final data. We have also added a tensorflow ANN model to predict the position of a player based on his attributes. Based on the outputs of all models used so far, we picked the one that gives the best accuracy and started hypertuning using gridcv. The complete data analysis made in phase 1 helped us in this phase to pick the best features to use in the model using advanced techniques infeature selections. Some of the features were dropped manually and some automatically based on correlation matrices and data properties. After the models have been tested, we came to the conclusion that no further processing needed. In this report we are going to describe the process of hypertuning and the results of the models to acheive 83% accuracy.\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initial Results\n",
    "The results of the initial models are as follows:\n",
    "- Random Forest (10 estimators): 80%\n",
    "- Random Forest using PCA data (10 estimators): 78%\n",
    "- Logistic Regression: 76%\n",
    "- Multilayer Perceptron FCM (manually tuned): 76%\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Choosing the final model\n",
    "We decided to on the random forest classifier since it produced the best results and had potential to improve with hypertuning. the multilayer classifier was also a good candidate but it was already manually hypertuned and didnot have much room for improvement.\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hypertuning\n",
    "We used gridcv to hypertune the model. We used the following parameters:\n",
    "- n_estimators: [25, 50, 100, 150]\n",
    "- max_depth: [6, 9, 12, 24]\n",
    "- max_features: [None, 'sqrt', 'log2']\n",
    "- max_leaf_nodes: [None, 66, 9, 12, 24]\n",
    "\n",
    "We chose gridcv because it is the most efficient way to hypertune a model. since it brute forces all the possible combinations of the parameters and returns the best one. It also allows us to pick the number of concurrent jobs to run in parallel, which makes it blazingly faster than the typical for loop hypertuning.\n",
    "\n",
    "Another reason to pick gridcv is the automatic cross validation it does. It splits the data into 5 folds and trains the model on 4 folds and tests on the remaining one. It does this 5 times and returns the average accuracy of the 5 folds. This is a very good way to test the model and make sure it is not overfitting.\n",
    "\n",
    "The best parameters were:\n",
    "- max_depth: 24\n",
    "- max_features: 'sqrt'\n",
    "- max_leaf_nodes: 24\n",
    "- n_estimators: 150\n",
    "\n",
    "With these parameters, the model accuracy jumped from 80% to 83.2%.\n"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The real accuracy\n",
    "\n",
    "Eventhough the model accuracy was 83.2%, the model displayed excellent results when plotted on a heatmap\n",
    "\n",
    "![image.png](attachment:image.png)\n",
    "\n",
    "The results here shows that, the propability of the model classifying a striker as a defender is very very low. Infact, the onyl reason the accuracy is low is because it only predicts a singl position for a player, which is not the case in real life. A player can play in multiple positions of same category ex: (ST, LW, RW), and the model is not able to predict that. If we were to consider the model as a multi-label classifier, the accuracy would be much higher."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "cmps460",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.10"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}