Introduction To Decision Trees – Part 1

Intro To Decision Trees :

Let’s start with a proper definition then go to an example :

Decision Tree is a type of Supervised Learning Algorithm wherein the data is continuously split on the basis of certain parameters.

To understand the decision tree in a better way let’s take an example given by the Udacity Course Instructor, Recommending Apps on Google Play Store.

Here is the sample data of Google Play Store which describes the user’s download history on the basis of Gender and Age. With the help of the decision tree algorithm, we’ll try to recommend the apps to the user on the basis of this data.

If the female is of age 15, she downloads Pokemon Go, and if Female is the age of 25 she downloads Whatsapp.

If the male is of age 32, he downloads Snapchat and so on…

So, we have data but between Gender and Age, which one seems more decisive for predicting what app the users will download?

Think, see the data…

So, we can use age as our parameter but why?

Because if we use gender as the parameter then we can see that female and male both downloads the Pokemon Go. So we can’t decide Pokemon Go should be recommended to every male or female? No…

That’s why we used age as our parameter in which we can identify that the people of age less than 20 irrespective of the gender is likely to download Pokemon Go and we recommend Pokemon Go the users of age less than 20.

And If the age is greater than 20 then we use gender as our parameter to check further, if the age is greater than 20 and the gender is female it is likely to download Whatsapp and if male it is likely to download Snapchat and we recommend on the basis of this decision tree algorithm.

We will be using this algorithm to recommend apps to the new user, and no tree or algorithm perfectly fits the data, it is us who tweak the parameters to find the best algorithm which fits on the data.

Stay Tuned for the next articles on decision tree and links are mentioned below for the next articles.

Thanks.

Sandeep Yadav