nb_model_tSNE.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "260031f3",
   "metadata": {},
   "source": [
    "## Load"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7018f2ef",
   "metadata": {},
   "outputs": [],
   "source": [
    "task_name = 'task1_sandwich_openface'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "21c8a6f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_path = './models/{0}/'.format(task_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b75d0af1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import warnings\n",
    "import pickle\n",
    "\n",
    "from tqdm import tqdm\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "55a91a2a",
   "metadata": {},
   "outputs": [],
   "source": [
    "save_path = './data/tmp_analysis/{0}'.format(task_name)\n",
    "save_train_path = save_path + '_train.csv'\n",
    "save_test_path = save_path + '_test.csv'\n",
    "save_valid_path = save_path + '_valid.csv'\n",
    "\n",
    "\n",
    "Train = pd.read_csv(save_train_path)\n",
    "Test = pd.read_csv(save_test_path)\n",
    "Valid = pd.read_csv(save_valid_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c8d8d0d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.asarray(Train.to_numpy()[:,3:]).astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b1d18d3b",
   "metadata": {},
   "outputs": [],
   "source": [
    "del Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "d57e73af",
   "metadata": {},
   "outputs": [],
   "source": [
    "Xtest = np.asarray(Test.to_numpy()[:,3:]).astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "32cfdade",
   "metadata": {},
   "outputs": [],
   "source": [
    "del Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "3d795865",
   "metadata": {},
   "outputs": [],
   "source": [
    "Xvalid = np.asarray(Valid.to_numpy()[:,3:]).astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "a3d6c461",
   "metadata": {},
   "outputs": [],
   "source": [
    "del Valid"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09970af1",
   "metadata": {},
   "source": [
    "## Direct TSNE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "45fc8c6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.manifold import TSNE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "92f99d3e",
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-13-b96ecea0d44b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mtsne\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mTSNE\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mn_components\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mtsne\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/analysisEnv/lib/python3.6/site-packages/sklearn/manifold/_t_sne.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, X, y)\u001b[0m\n\u001b[1;32m    948\u001b[0m         \u001b[0my\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mIgnored\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    949\u001b[0m         \"\"\"\n\u001b[0;32m--> 950\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    951\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/analysisEnv/lib/python3.6/site-packages/sklearn/manifold/_t_sne.py\u001b[0m in \u001b[0;36mfit_transform\u001b[0;34m(self, X, y)\u001b[0m\n\u001b[1;32m    930\u001b[0m             \u001b[0mEmbedding\u001b[0m \u001b[0mof\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mtraining\u001b[0m \u001b[0mdata\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mlow\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mdimensional\u001b[0m \u001b[0mspace\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    931\u001b[0m         \"\"\"\n\u001b[0;32m--> 932\u001b[0;31m         \u001b[0membedding\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_fit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    933\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0membedding_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0membedding\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    934\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0membedding_\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/analysisEnv/lib/python3.6/site-packages/sklearn/manifold/_t_sne.py\u001b[0m in \u001b[0;36m_fit\u001b[0;34m(self, X, skip_num_points)\u001b[0m\n\u001b[1;32m    842\u001b[0m                           \u001b[0mX_embedded\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mX_embedded\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    843\u001b[0m                           \u001b[0mneighbors\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mneighbors_nn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 844\u001b[0;31m                           skip_num_points=skip_num_points)\n\u001b[0m\u001b[1;32m    845\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    846\u001b[0m     def _tsne(self, P, degrees_of_freedom, n_samples, X_embedded,\n",
      "\u001b[0;32m~/analysisEnv/lib/python3.6/site-packages/sklearn/manifold/_t_sne.py\u001b[0m in \u001b[0;36m_tsne\u001b[0;34m(self, P, degrees_of_freedom, n_samples, X_embedded, neighbors, skip_num_points)\u001b[0m\n\u001b[1;32m    881\u001b[0m         \u001b[0mP\u001b[0m \u001b[0;34m*=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mearly_exaggeration\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    882\u001b[0m         params, kl_divergence, it = _gradient_descent(obj_func, params,\n\u001b[0;32m--> 883\u001b[0;31m                                                       **opt_args)\n\u001b[0m\u001b[1;32m    884\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mverbose\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    885\u001b[0m             print(\"[t-SNE] KL divergence after %d iterations with early \"\n",
      "\u001b[0;32m~/analysisEnv/lib/python3.6/site-packages/sklearn/manifold/_t_sne.py\u001b[0m in \u001b[0;36m_gradient_descent\u001b[0;34m(objective, p0, it, n_iter, n_iter_check, n_iter_without_progress, momentum, learning_rate, min_gain, min_grad_norm, verbose, args, kwargs)\u001b[0m\n\u001b[1;32m    361\u001b[0m         \u001b[0mkwargs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'compute_error'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcheck_convergence\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mn_iter\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    362\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 363\u001b[0;31m         \u001b[0merror\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgrad\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mobjective\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    364\u001b[0m         \u001b[0mgrad_norm\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlinalg\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnorm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgrad\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    365\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/analysisEnv/lib/python3.6/site-packages/sklearn/manifold/_t_sne.py\u001b[0m in \u001b[0;36m_kl_divergence_bh\u001b[0;34m(params, P, degrees_of_freedom, n_samples, n_components, angle, skip_num_points, verbose, compute_error, num_threads)\u001b[0m\n\u001b[1;32m    265\u001b[0m                                       \u001b[0mdof\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdegrees_of_freedom\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    266\u001b[0m                                       \u001b[0mcompute_error\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcompute_error\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 267\u001b[0;31m                                       num_threads=num_threads)\n\u001b[0m\u001b[1;32m    268\u001b[0m     \u001b[0mc\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m2.0\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mdegrees_of_freedom\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1.0\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mdegrees_of_freedom\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    269\u001b[0m     \u001b[0mgrad\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mgrad\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mravel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "tsne = TSNE(n_components=3)\n",
    "tsne.fit(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97c7a650",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(X)\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9d1d0cfb",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(Xtest)\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1680673c",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(Xvalid)\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "844bd3f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "pickle.dump(tsne,open(model_path+'tsne_3.p','wb'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "05ef90d3",
   "metadata": {},
   "source": [
    "## PCA - TSNE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b9aa276e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "# scikit-learn v0.24.1\n",
    "PCA_n_32 = pickle.load(open(model_path + 'pca_n_32.p', 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9874c541",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.manifold import TSNE\n",
    "tsne = TSNE(n_components=3)\n",
    "tsne.fit(PCA_n_32.transform(X))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4f98efca",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(PCA_n_32.transform(X))\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4bacc392",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(PCA_n_32.transform(Xtest))\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6cbd3090",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(PCA_n_32.transform(Xvalid))\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0d305eff",
   "metadata": {},
   "outputs": [],
   "source": [
    "pickle.dump(tsne,open(model_path+'tsne_3_pca_32.p','wb'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d468b2e0",
   "metadata": {},
   "source": [
    "## Autoencoder - TSNE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f4154751",
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "Encoder_32 = keras.models.load_model('./models/all_data/alt_encoder_32.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3127bfec",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.manifold import TSNE\n",
    "tsne = TSNE(n_components=3)\n",
    "tsne.fit(Encoder_32.predict(X))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "05377b39",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(Encoder_32.predict(X))\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e6964dd6",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(Encoder_32.predict(Xtest))\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f18697bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hat = tsne.transform(Encoder_32.predict(Xvalid))\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(projection='3d')\n",
    "ax.scatter(X_hat[:,0], X_hat[:,1], X_hat[:,2], s = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "64311280",
   "metadata": {},
   "outputs": [],
   "source": [
    "pickle.dump(tsne,open(model_path+'tsne_3_encoder_32.p','wb'))"
   ]
  }
 ],
 "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": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}