Feature Importance In Decision Tree | Sklearn | Scikit Learn | Python | Machine Learning | Codegnan
In this video, you will learn more about Feature Importance in Decision Trees using Scikit Learn library in Python. You will also learn how to visualise it.
Decision trees are a type of supervised Machine Learning. There are 2 types of Decision trees – classification(categorical) and regression(continuous data types).
Decision trees split data into smaller subsets for prediction, based on some parameters. By making splits using Decision trees, one can maximize the decrease in impurity. The condition is represented as leaf and possible outcomes are represented as branches.
Decision trees can be useful to check the feature importance. We can look for the important features and remove those features which are not contributing much for making classifications.
The importance of a feature, also known as the Gini importance, is the normalized total reduction of the criterion brought by that feature.
Get the feature importance of each variable along with the feature name sorted in descending order of their importance. Then you can drop variables that are of no use in forming the decision tree.
The decreasing order of importance of each feature is useful. You can plot this as well with feature name on X-axis and importances on Y-axis on a bar graph.
This graph shows the mean decrease in impurity against the probability of reaching the feature.
For lesser contributing variables(variables with lesser importance value), you can decide to drop them based on business needs.
Learn Machine Learning from our Tutorials: http://bit.ly/CodegnanMLPlaylist
Learn Python from our Tutorials: http://bit.ly/CodegnanPythonTuts
Subscribe to our channel and hit the bell icon and never miss the update: https://bit.ly/SubscribeCodegnan
++++++++++++++Follow us ++++++++++++++++
Visit our website: https://codegnan.com
CodeGnan offers courses in new technologies and niches that are gaining cult reach. This is to ensure that students understand the workflow from each and every perspective in a Real-Time environment. Students can train themselves and enrich their skillset in the best way possible.
We always used to believe in student-centric methods. Every student, if trained in a Real-Time environment can achieve more in their careers. Hence, CodeGnan offers courses where students can access live environments and nourish themselves in the best way possible in order to increase their “CodeGnan”.
With Codegnan, you get an industry-recognized certificate with worldwide validity. You get to reach the heights of your career in a shorter period of time. With your skillset, you can find a place at any top companies in India and worldwide. This will help you to improve your skillset like never before and get access to the top-level placement opportunities that are currently available.
CodeGnan offers courses in new technologies and makes sure students understand the flow of work from each and every perspective in a Real-Time environment.
#Featureselection #FeatureSelectionTechnique #DecisionTree #FeatureImportance #Machinelearninng #python