\n",
"\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" genus\n",
"id \n",
"2783 0.847273\n",
"2175 0.875663\n",
"4517 0.879757\n",
"2831 0.732987\n",
"3556 0.828222"
]
},
"execution_count": 29,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"sf = pd.read_csv(\"data/SubmissionFormat.csv\", index_col=0)\n",
"\n",
"def make_predictions(submission_format_df,output_filepath=\"my_submission.csv\"):\n",
" prediction_df = submission_format_df.copy()\n",
" \n",
" # create features\n",
" test_features = create_feature_matrix(submission_format_df)\n",
" test_features_stand = ss.transform(test_features) \n",
" test_features_pca = pca.transform(test_features_stand)\n",
" \n",
" # predict with the best estimator from the grid search\n",
" preds = gs.best_estimator_.predict_proba(test_features_pca)\n",
" \n",
" # copy the predictions to the dataframe with the ids and\n",
" # write it out to a file\n",
" prediction_df.genus = preds[:, 1]\n",
" prediction_df.to_csv(output_filepath)\n",
" \n",
" return prediction_df\n",
" \n",
"pred_df = make_predictions(sf, output_filepath=\"raw_hog_daisy_rbf.csv\")\n",
"pred_df.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"| \n", " | genus | \n", "
|---|---|
| id | \n", "\n", " |
| 2783 | \n", "0.847273 | \n", "
| 2175 | \n", "0.875663 | \n", "
| 4517 | \n", "0.879757 | \n", "
| 2831 | \n", "0.732987 | \n", "
| 3556 | \n", "0.828222 | \n", "
![](\"https://s3.amazonaws.com/drivendata.org/bees-example/ss.png\")
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Yay! We can see that our cross-validation and holdout scores were pretty reliable in this case. Now it's up to you to [**make the performance even better**](http://www.drivendata.org/competitions/8/)!"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"
\n",
"
\n",
"\n",
"* Check out our [forum](http://community.drivendata.org/)\n",
"* Or, give us a shout out on twitter: **[@drivendataorg](https://twitter.com/drivendataorg)**\n"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 2",
"language": "python",
"name": "python2"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.9"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
\n",
"