From the course: Machine Learning in Telecommunication: From Basics to Real-World Cases
Join today to access over 24,800 courses taught by industry experts.
Variance reduction and feature importance
From the course: Machine Learning in Telecommunication: From Basics to Real-World Cases
Variance reduction and feature importance
(upbeat music) - [Instructor] As we need to reduce the variance for better predictions, let's understand how the variance reduction works. Initially, the data points are spreaded out having a higher variance, but after each split of the data, they becomes more uniform and reducing the variance. And this step-by-step reduction helps the tree to capture overall patterns effectively, which improves the prediction accuracy. So variance starts high and start decreasing after each split because the subsets become more homogeneous. So why variance matters in regression trees? Variance matter because it directly impacts the prediction quality. By minimizing the variance, the tree reduces the uncertainty in the predictions. Lower variance means model generalize better to new data, which avoids overfitting and ensure reliable results. Imagine predicting house prices if you have a tree splitting data by number of bedrooms and their pricing in each subset, for example, two bedroom homes, it…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.