Using OpenML 9.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import openml\n",
    "import numpy as np\n",
    "from sklearn.svm import NuSVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Building the NuSVC model object with parameters found\n",
    "clf = NuSVC(cache_size=200, class_weight=None, coef0=0.0,\n",
    "   decision_function_shape=None, degree=3, gamma='auto', kernel='linear',\n",
    "   max_iter=-1, nu=0.3, probability=True, random_state=3, shrinking=True,\n",
    "   tol=3.2419092644286417e-05, verbose=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OpenML Classification Task\n",
       "==========================\n",
       "Task Type Description: https://www.openml.org/tt/1\n",
       "Task ID..............: 59\n",
       "Task URL.............: https://www.openml.org/t/59\n",
       "Estimation Procedure.: crossvalidation\n",
       "Evaluation Measure...: predictive_accuracy\n",
       "Target Feature.......: class\n",
       "# of Classes.........: 3\n",
       "Cost Matrix..........: Available"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Obtaining task used earlier\n",
    "t = openml.tasks.get_task(59)\n",
    "t"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OpenML Flow\n",
       "===========\n",
       "Flow Name.......: sklearn.svm.classes.NuSVC\n",
       "Flow Description: Nu-Support Vector Classification.\n",
       "\n",
       "Similar to SVC but uses a parameter to control the number of support\n",
       "vectors.\n",
       "\n",
       "The implementation is based on libsvm.\n",
       "Dependencies....: sklearn==0.21.3\n",
       "numpy>=1.6.1\n",
       "scipy>=0.9"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Running the model on the task\n",
    "# Internally, the model will be made into \n",
    "# an OpenML flow and we can choose to retrieve it\n",
    "r, f = openml.runs.run_model_on_task(model=clf, task=t, upload_flow=False, return_flow=True)\n",
    "f"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9866666666666667\n"
     ]
    }
   ],
   "source": [
    "# To obtain the score (without uploading)\n",
    "## r.publish() can be used to upload these results\n",
    "## need to sign-in to https://www.openml.org/\n",
    "score = []\n",
    "evaluations = r.fold_evaluations['predictive_accuracy'][0]\n",
    "for key in evaluations:\n",
    "    score.append(evaluations[key])\n",
    "print(np.mean(score))"
   ]
  }
 ],
 "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}