Discussion of article "Data Science and Machine Learning (Part 05): Decision Trees"

New comment
 

New article Data Science and Machine Learning (Part 05): Decision Trees has been published:

Decision trees imitate the way humans think to classify data, let's see how to build a tree and use them to classify and predict some data. The main goal of the decision trees algorithm is to separate the data with impurity and into pure or close to nodes.

Decision Trees use multiple algorithms to decide to split a node into two or more subset nodes. The creation of subnodes increases the homogeneity of resultant sub-nodes. In other words, we can say that the purity of the node increases concerning the target variable. The decision tree algorithm splits nodes on all available variables and then selects the split that results in the most homogeneous sub-nodes.

decision tree example

The algorithm selection is based on the type of target variables 

The following are the Algorithms used in the Decision Tree 

  1. ID3 > Extension of D3
  2. C4.5 > Successor of ID3
  3. CART > Classification and Regression Tree
  4. CHAID > Chi-square Automatic Interaction Detection, performs multi-level splits when computing classification trees
  5. MARS > Multivariate Adaptive Regression Splines

In this article I am going to create a decision tree based on the ID3 algorithm, we'll discuss and use the other algorithms in the next Articles of this series 


Author: Omega J Msigwa

 

Hello Mr. Omega,

Thank you so much for the ID3 solution. it is very useful for me. However I provided and attached an excel sheet in this regards, that I think it's clear for your explains.

Many Thanks again,

F.Mahmoudian

Files:
ID3-DecisionTree.zip  39 kb
 
Fatemeh Haji Mahmoudian #:

Hello Mr. Omega,

Thank you so much for the ID3 solution. it is very useful for me. However I provided and attached an excel sheet in this regards, that I think it's clear for your explains.

Many Thanks again,

F.Mahmoudian

many thanks to it, I'm still trying to figure out how to let the script draw the tree itself

 
Omega J Msigwa #:

many thanks to it, I'm still trying to figure out how to let the script draw the tree itself

That would be great! 

Thanks so much

 

And by the way, yes, why everyone is so obsessed with MO, AI and DeepLearning? There is a well-forgotten old thing, which the topical starter reminded us about. There are expert systems and all sorts of weighted assessments. Of course the methods are 30-50 years old, not fashionable, but they stick to the physical model and cause and effect relationships and their results are interpretable. I'VE GOT TO DIG IN THERE.

it's the only thing that could potentially be a filter for already calculated signals. Other methods in this direction are fucked.

 
Maxim Kuznetsov #:

And by the way, yes, why everyone is so obsessed with MO, AI and DeepLearning? There is a well-forgotten old thing, which the topical starter reminded us about. There are expert systems and all sorts of weighted assessments. Of course the methods are 30-50 years old, not fashionable, but they stick to the physical model and cause and effect relationships and their results are interpretable. I'LL HAVE TO DIG IN THERE.

it's the only thing that could potentially be a filter for already calculated signals. Other methods in this direction have been fucked up.

Actually it's a basic algorithm, part of a more complex tree-based ) and the most retrained.
New comment