{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "20fce00f",
   "metadata": {},
   "source": [
    "### Deep Learning\n",
    "#### Artificial neuron \n",
    "An **artificial neuron** gets an input vector $\\boldsymbol{x}$ and produces a scalar value $y$, which is modeled by the following equation:\n",
    "<br> $\\large y=\\phi(\\boldsymbol{x}^T\\boldsymbol{w}+b)=\\phi(\\boldsymbol{w}^T\\boldsymbol{x}+b)$\n",
    "<br> Where $b$ is the bias. The weight vector is $\\boldsymbol{w}=[w_1,w_2,...,w_m]^T$. The input vector is $\\boldsymbol{x}=[x_1,x_2,...,x_m]^T$. The $\\phi$ is an activation function. \n",
    "- Common activation functions are: *Logistic*, *ReLU*, *Tanh*, and *Linear*.\n",
    "\n",
    "<hr>\n",
    "In the following, we implement an arificial neuron based on the formula mentioned above.\n",
    "<hr>\n",
    "https://github.com/ostad-ai/Machine-Learning\n",
    "<br> Explanation: https://www.pinterest.com/HamedShahHosseini/Machine-Learning/Data-Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "22ebfdcc",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import required module\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "57cfa2b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the neuron with y=φ(x^Tw+b)\n",
    "class ArtificialNeuron:\n",
    "    def __init__(self, n_inputs, activation='logistic'):\n",
    "        self.weights = np.random.randn(n_inputs)\n",
    "        self.bias = np.random.randn()\n",
    "        self.activation = activation\n",
    "\n",
    "    def __call__(self, x):\n",
    "        z = np.dot(x, self.weights) + self.bias\n",
    "        if self.activation == 'logistic':\n",
    "            return 1 / (1 + np.exp(-z))\n",
    "        elif self.activation == 'relu':\n",
    "            return np.maximum(0, z)\n",
    "        else:  # Linear\n",
    "            return z"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b6d79cba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input is: [ 0.3 -1.2  0.7]\n",
      "Output is:0.9308216406202954\n",
      "Weights are: [ 1.41314181 -1.01381938 -0.12499518];\n",
      "and Bias is:1.0463504203241891\n"
     ]
    }
   ],
   "source": [
    "# Example\n",
    "neuron = ArtificialNeuron(n_inputs=3)\n",
    "x = np.array([0.3, -1.2, 0.7])\n",
    "y = neuron(x)\n",
    "print(f'Input is: {x}')\n",
    "print(f'Output is:{y}')  # Output after activation\n",
    "print(f'Weights are: {neuron.weights};\\nand Bias is:{neuron.bias}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7cb5407a",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}