Building a Regression Model

One of the great things about PySpark ML module is that most algorithms can be tried and tested without changing much code. Random Forest Regression is a fairly simple ensemble model, using bagging to fit. Another tree based ensemble model is Gradient Boosted Trees which uses a different approach called boosting to fit. In this exercise let's train a GBTRegressor.

แบบฝึกหัดนี้เป็นส่วนหนึ่งของหลักสูตร

Feature Engineering with PySpark

ดูคอร์ส

คำแนะนำการฝึกหัด

Import GBTRegressor from pyspark.ml.regression which you will notice is the same module as RandomForestRegressor.
Instantiate GBTRegressor with featuresCol set to the vector column of our features named, features, labelCol set to our dependent variable, SALESCLOSEPRICE and the random seed to 42
Train the model by calling fit() on gbt with the imported training data, train_df.

แบบฝึกหัดเชิงโต้ตอบแบบลงมือทำ

ลองทำแบบฝึกหัดนี้โดยเติมโค้ดตัวอย่างนี้ให้สมบูรณ์

from ____ import ____

# Train a Gradient Boosted Trees (GBT) model.
gbt = ____(featuresCol=____,
                           labelCol=____,
                           predictionCol="Prediction_Price",
                           seed=____
                           )

# Train model.
model = gbt.fit(train_df)

แก้ไขและรันโค้ด