Giter Club home page Giter Club logo

Comments (4)

svaante avatar svaante commented on August 16, 2024

I am happy to track down the bug, but could you specify which data set you where using to save me some time.

from decision-tree-id3.

shenxiangzhuang avatar shenxiangzhuang commented on August 16, 2024

I used the data set in the doc.

Code:

from id3 import Id3Estimator, export_text
import numpy as np

feature_names = ["age",
                 "gender",
                 "sector",
                 "degree"]

X = np.array([[45, "male", "private", "m"],
              [50, "female", "private", "m"],
              [61, "other", "public", "b"],
              [40, "male", "private", "none"],
              [34, "female", "private", "none"],
              [33, "male", "public", "none"],
              [43, "other", "private", "m"],
              [35, "male", "private", "m"],
              [34, "female", "private", "m"],
              [35, "male", "public", "m"],
              [34, "other", "public", "m"],
              [34, "other", "public", "b"],
              [34, "female", "public", "b"],
              [34, "male", "public", "b"],
              [34, "female", "private", "b"],
              [34, "male", "private", "b"],
              [34, "other", "private", "b"]])

y = np.array(["(30k,38k)",
              "(30k,38k)",
              "(30k,38k)",
              "(13k,15k)",
              "(13k,15k)",
              "(13k,15k)",
              "(23k,30k)",
              "(23k,30k)",
              "(23k,30k)",
              "(15k,23k)",
              "(15k,23k)",
              "(15k,23k)",
              "(15k,23k)",
              "(15k,23k)",
              "(23k,30k)",
              "(23k,30k)",
              "(23k,30k)"])
# is_repeating=True
clf1 = Id3Estimator(is_repeating=True)
clf1.fit(X, y, check_input=True)
print('is_repeating=True:')
print(export_text(clf1.tree_, feature_names))


# is_repeating=False
clf2 = Id3Estimator(is_repeating=False)
clf2.fit(X, y, check_input=True)
print('is_repeating=False:')
print(export_text(clf1.tree_, feature_names))

The output:

is_repeating=True:

degree b
|   sector private: (23k,30k) (3) 
|   sector public
|   |   age <=47.50: (15k,23k) (3) 
|   |   age >47.50: (30k,38k) (1) 
degree m
|   age <=44.00
|   |   sector private: (23k,30k) (3) 
|   |   sector public: (15k,23k) (2) 
|   age >44.00: (30k,38k) (2) 
degree none: (13k,15k) (3) 

is_repeating=False:

degree b
|   sector private: (23k,30k) (3) 
|   sector public
|   |   age <=47.50: (15k,23k) (3) 
|   |   age >47.50: (30k,38k) (1) 
degree m
|   age <=44.00
|   |   sector private: (23k,30k) (3) 
|   |   sector public: (15k,23k) (2) 
|   age >44.00: (30k,38k) (2) 
degree none: (13k,15k) (3)

There are two age and sector in different levels when I set is_repeating=False. (Actually, I am not sure about whether or not this situation could be considered as repeating:-)

from decision-tree-id3.

svaante avatar svaante commented on August 16, 2024

After giving the code a peek after +1 year I figured out the behavior of is_repeating. The parameter doesn't constrain global repeating features. The flag is_repeating constrains a decision node to pick a feature not used by any parent of said node. So the following tree is not a bug per se with is_repeating = True:

degree b
|   sector private: (23k,30k) (3) 
|   sector public
|   |   age <=47.50: (15k,23k) (3) 
|   |   age >47.50: (30k,38k) (1) 
degree m
|   age <=44.00
|   |   sector private: (23k,30k) (3) 
|   |   sector public: (15k,23k) (2) 
|   age >44.00: (30k,38k) (2) 
degree none: (13k,15k) (3)

By definition this only limit numerical features from repeating.

If this behavior is by design or incorrectly implemented I actually can't remember, but the docs is either way surly lacking in the case of is_repeating.

I will give this a think. Thanks for bringing it up!

from decision-tree-id3.

shenxiangzhuang avatar shenxiangzhuang commented on August 16, 2024

Well, I think your explanation for the parameter is_repeating is right and the implementation of the decision tree algorithm is right too.(I'm the one that misunderstanding the theory of ID3:-) Anyway, as you have said, we may need a more appropriate demo to make the parameter's meaning clear. Thanks a lot for your awesome work and helpful reply.

from decision-tree-id3.

Related Issues (6)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.