{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A single layer ANN exercise"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this python notebook you should:\n",
    "1. Import a dataset from a text file\n",
    "2. Split the dataset into input and ouput data\n",
    "3. Build a single layer ANN using tensorflow and keras\n",
    "4. Visualize the accuracy of the ANN\n",
    "5. What can you try to improve the accuracy of the ANN?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Dataset description"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset we are interested in is stored in the file called 'Data_Exercise_1.txt'.\\\n",
    "The file contains 1000 points where the three first columns represent the input variable while the 3 last columns represents the ouput variables."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### ANN description"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The ANN we want to build contains 3 input and 3 output variables.\\\n",
    "The activation functions should be defined as rectified linear units (relu).\\\n",
    "The output layer should be defined by a linear activation function.\\\n",
    "As a starting point the number of neurons should be set to 50.\\\n",
    "The number of epochs should be set to 50."
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}