Add files via upload

car_insurance.ipynb

@@ -1490,7 +1490,7 @@
     "%%capture\n",
     "from sklearn.model_selection import GridSearchCV\n",
     "\n",
-    "#grid search for decision tree regression\n",
+    "#grid search for Kneighbours classification\n",
     "tuningKNC=GridSearchCV(modelKNC,param_grid=ParamsKNC,scoring='neg_mean_squared_error',cv=3,verbose=3)\n",
     "tuningKNC.fit(X_train,y_train)\n"
    ]
@@ -1536,6 +1536,7 @@
     }
    ],
    "source": [
+    "#Decision tree classifier using the best parameters found\n",
     "modelDT = DecisionTreeClassifier(criterion=\"entropy\",max_depth=11,max_features=None,min_samples_leaf=9,min_weight_fraction_leaf=0.1,splitter=\"random\")\n",
     "modelDT = modelDT.fit(X_train, y_train)\n",
     "modelDT.predict(X_test)\n",
@@ -1559,12 +1560,21 @@
     }
    ],
    "source": [
+    "#KNeighbours Classifier using the best parameters found\n",
     "modelKNC = KNeighborsClassifier(algorithm=\"auto\",metric=\"manhattan\",n_neighbors=10,p=1,weights=\"uniform\")\n",
     "modelKNC = modelKNC.fit(X_train,y_train)\n",
     "modelKNC.predict(X_test)\n",
     "print (\"The accuricy for the train set is :\" + str(modelKNC.score(X_train,y_train)))\n",
     "print (\"The accuricy for the test set is :\" + str(modelKNC.score(X_test,y_test)))"
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "In both tuned models the score obtained in the training set were infiriors to the ones obtained using default parameteres,\n",
+    "but, the score obtained by the tuned models in the test set were better than using the default parameters"
+   ]
   }
  ],
  "metadata": {