dnn_PyTorch.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Brain Age Prediction using Deep Neural Network with PyTorch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import random as rn\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "#from keras.callbacks import TensorBoard\n",
    "\n",
    "from plotnine import * # a Grammar of Graphics for Python, aka ggplot2 in R\n",
    "\n",
    "import time\n",
    "from sklearn import preprocessing\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn import datasets, linear_model\n",
    "\n",
    "import torch\n",
    "import torch.nn.functional as F\n",
    "import torch.nn as nn\n",
    "from torch.autograd import Variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data\n",
    "dat = np.load('./full_reduced/full/feaDAT.npy')\n",
    "age = np.load('./full_reduced/full/ageDAT.npy')\n",
    "sex = np.load('./full_reduced/full/sexDAT.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 'cmd /' to comment on/off\n",
    "X = preprocessing.scale(dat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2026, 5124)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Splitting into Training set and Test set, X, Y pair\n",
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, age, test_size = 0.2, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train1 = Variable(torch.from_numpy(X_train))\n",
    "Y_train1 = Variable(torch.from_numpy(Y_train))\n",
    "X_train1 = X_train1.float()\n",
    "Y_train1 = Y_train1.float()\n",
    "\n",
    "X_test1 = Variable(torch.from_numpy(X_test))\n",
    "X_test1 = X_test1.float()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.manual_seed(999)    # reproducible\n",
    "# model setup\n",
    "n_hidden = 50\n",
    "class Model(torch.nn.Module): \n",
    "    \n",
    "    def __init__(self):\n",
    "        super(Model, self).__init__()\n",
    "        self.layer1 = nn.Linear(X.shape[1], 500)\n",
    "        self.layer2 = nn.Linear(500, n_hidden)\n",
    "        self.layer3 = nn.Linear(n_hidden, n_hidden)\n",
    "        self.layer4 = nn.Linear(n_hidden, 1)\n",
    "        \n",
    "        self.sigmoid = torch.nn.Sigmoid()\n",
    "        # self.dropout = nn.Dropout(0.5)  \n",
    "            \n",
    "    def forward (self, x):\n",
    "        out1 = self.sigmoid(self.layer1(x))\n",
    "        out2 = self.sigmoid(self.layer2(out1))\n",
    "        out3 = self.sigmoid(self.layer3(out2))\n",
    "        y_pred = self.layer4(out3) # linear output\n",
    "        return y_pred\n",
    "    \n",
    "net = Model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "criterion = nn.MSELoss()# Mean Squared Loss\n",
    "l_rate = 0.001\n",
    "optimizer = torch.optim.SGD(net.parameters(), \n",
    "                            lr = l_rate, \n",
    "                            weight_decay=1e-6) \n",
    "#                             momentum=0.9, \n",
    "#                             nesterov=True) #Stochastic Gradient Descent\n",
    "\n",
    "epochs = 500 # "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 100, loss 297.25\n",
      "epoch 200, loss 112.18\n",
      "epoch 300, loss 41.85\n",
      "epoch 400, loss 23.34\n",
      "epoch 500, loss 15.03\n"
     ]
    }
   ],
   "source": [
    "inputs = X_train1\n",
    "labels = Y_train1.view(Y_train.shape[0],1)\n",
    "# Collect errors to evaluate performance\n",
    "errorlist = [];\n",
    "for epoch in range(epochs):\n",
    "\n",
    "    #increase the number of epochs by 1 every time\n",
    "    epoch +=1\n",
    "    \n",
    "    #clear grads \n",
    "    optimizer.zero_grad()\n",
    "    #forward to get predicted values\n",
    "    outputs = net(inputs) # model predict,  outputs = net.forward(inputs)\n",
    "    loss = criterion(outputs, labels)\n",
    "    loss.backward()# back props\n",
    "    optimizer.step()# update the parameters \n",
    "    \n",
    "    errorlist = np.append(errorlist,loss.item()) \n",
    "    \n",
    "    if epoch % 100 == 0:\n",
    "        print('epoch {}, loss {:.2f}'.format(epoch,loss.data))\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train data correlation 0.984 \n"
     ]
    }
   ],
   "source": [
    "y_train_pred = net(X_train1)\n",
    "y_train_pred = y_train_pred.detach().numpy()\n",
    "corr_train = np.corrcoef(Y_train, y_train_pred.flatten())[0,1]\n",
    "print(\"Train data correlation %.3f \" % corr_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test data correlation 0.928 \n"
     ]
    }
   ],
   "source": [
    "X_prediction = net(X_test1)\n",
    "X_prediction = X_prediction.detach().numpy()\n",
    "corr_test = np.corrcoef(Y_test, X_prediction.flatten())[0,1]\n",
    "print(\"Test data correlation %.3f \" % corr_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5.752950768700962"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_absolute_error(Y_test, X_prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(errorlist,'--.g')\n",
    "plt.title(\"Mean Squarred Error\")\n",
    "plt.xlabel(\"Batch Iteration Number\")\n",
    "plt.ylabel(\"Error\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[min, median, max] age: [6 25 88] years\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/numpy/core/fromnumeric.py:2223: FutureWarning: Method .ptp is deprecated and will be removed in a future version. Use numpy.ptp instead.\n",
      "  return ptp(axis=axis, out=out, **kwargs)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<ggplot: (334191315)>\n"
     ]
    }
   ],
   "source": [
    "# bias shown in brain age gap: over/under-estimate for lower/upper ends\n",
    "idx = np.argsort(Y_test)\n",
    "age_gap = X_prediction.flatten() - Y_test\n",
    "dframe = pd.DataFrame()\n",
    "dframe['age_gap'] = age_gap\n",
    "dframe['Y_test'] = Y_test\n",
    "p2 = ggplot(dframe, aes('Y_test','age_gap')) + geom_point() +\\\n",
    "    geom_hline(yintercept=[0], size=1, linetype='--') +\\\n",
    "    xlab('chronological age') +\\\n",
    "    ylab('brain age gap') +\\\n",
    "    stat_smooth(method='lm') \n",
    "  \n",
    "print(\"[min, median, max] age: [%d %d %d] years\" % (min(Y_test), np.median(Y_test), max(Y_test)))\n",
    "print(p2) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}