Intro to Machine Learning

构建模型步骤：

• • 　　Define:

　　　　　　　  What type of model will it be? A decision tree? Some other type of model? Some other parameters of the model type are specified too.

• • 　　Fit:

Capture patterns from provided data. This is the heart of modeling

• • 　　Predict:

　　　　　　　 Just what it sounds like

• • 　　Evaluate:

　　　　　　　  Determine how accurate the model’s predictions are

实现:

　　　　from sklearn.tree import DecisionTreeRegressor

　　　　melbourne_mode = DecisionTreeRegressor(random_state=1)

　　　　melbourne_mode.fit(X , y)

打印出开始几行：

　　　　print (X.head())

预测后的价格如下：

　　　　print (melbourne_mode.predict(X.head())

4.Model Validation

由于预测的价格和真实的价格会有差距，而差距多少，我们需要衡量

使用Mean Absolute Error

　　　　error= actual-predicted

在实际过程中，我们要将数据分成两份，一份用于训练，叫做training data, 一份用于验证叫validataion data

　　　　from sklearn.model_selection import train_test_split

　　　　train_X, val_X,  train_y, val_y =     train_test_split(X, y, random_state=0)

　　　　melbourne_model               =      DecisionTreeRegressor()

　　　　melbourne_model.fit(train_X, train_y)

　　　　val_predictions                  =      melbourne_model.predict(val_X)

　　　　print(mean_absolute_error(val_y, val_predictions))

5.Underfitting and Overfitting

• overfitting:     A model matches the data almost perfectly, but does poorly in validation and other new data.
• underfitting:   When a model fails to capture important distinctions and patterns in the data， so it performs poorly even in training data.

The more leaves we allow the model to make, the more we move from the underfitting area in the above graph to overfitting area.

　　from sklearn.metrics import mean_absolute_error

　　from sklearn.tree import DecsionTreeRegressor

　　def get_ame(max_leaf_nodes, train_X, val_X, train_y, val_y):

　　　　model = DecisionTreeRegressor(max_leaf_nodes = max_leaf_nodes, random_state = 0)

　　　　model.fit(train_X, train_y)

　　　　preds_val = model.predict(val_X)

　　　　mae = mean_absolute_error(val_y, preds_val)

　　　　return(mae)

我可以使用循环比较选择最合适的max_leaf_nodes

　　　　for max_leaf_nodes in [5,50,500,5000]:

　　　　　　my_ame = get_ame(max_leaf_nodes, train_X, val_X, train_y, val_y)

　　　　　　print(max_leaf_nodes, my_ame)

最后可以发现，当max leaf nodes 为 500时，MAE最小, 接下来我们换另外一种模型

6.Random Forests

The random forest uses many trees, and it makes a prediction by averaging the predictions of each component tree. It generally has much better predictive accuracy than a single decision tree and it works well with default parameters.

　　　　from sklearn.ensemble import RandomForestRegressor

　　　　from sklearn.metrics import mean_absolute_error

　　　　forest_model = RandomForestRegressor(random_state=1)

　　　　forest_model.fit(train_X,train_y)

　　　　melb_preds = forest_model.predict(val_X)

　　　　print(mean_absolute_error(val_y, melb_preds))

可以发现最后的误差，相对于决策树小。

one of the best features of Random Forest models is that they generally work reasonably even without this tuning.

7.Machine Learning Competitions

• Build a Random Forest model with all of your data

• Read in the “test” data, which doesn’t include values for the target. Predict home values in the test data with your Random Forest model.

• Submit those predictions to the competition and see your score.

• Optionally, come back to see if you can improve your model by adding features or changing your model. Then you can resubmit to see how that stacks up on the competition leaderboard.