Home Backend Development Python Tutorial How to Extract Decision Rules from a Scikit-Learn Decision Tree?

How to Extract Decision Rules from a Scikit-Learn Decision Tree?

Oct 28, 2024 am 02:26 AM

How to Extract Decision Rules from a Scikit-Learn Decision Tree?

Extracting Decision Rules from Scikit-Learn Decision Tree

Decision trees, a widely used machine learning algorithm, provide insights by modeling decision-making processes as a hierarchical structure of rules. However, extracting these decision rules explicitly can be challenging. This article outlines a comprehensive approach to extracting textual decision rules from a trained Scikit-Learn decision tree.

Python Code for Decision Rule Extraction

The following Python code snippet utilizes the underlying data structures of Scikit-Learn decision trees to traverse and generate human-readable decision paths:

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

from sklearn.tree import _tree

 

def tree_to_code(tree, feature_names):

    tree_ = tree.tree_

    feature_name = [

        feature_names[i] if i != _tree.TREE_UNDEFINED else "undefined!"

        for i in tree_.feature

    ]

    print("def tree({}):".format(", ".join(feature_names)))

 

    def recurse(node, depth):

        indent = "  " * depth

        if tree_.feature[node] != _tree.TREE_UNDEFINED:

            name = feature_name[node]

            threshold = tree_.threshold[node]

            print("{}if {} <= {}:".format(indent, name, threshold))

            recurse(tree_.children_left[node], depth + 1)

            print("{}else:  # if {} > {}".format(indent, name, threshold))

            recurse(tree_.children_right[node], depth + 1)

        else:

            print("{}return {}".format(indent, tree_.value[node]))

 

    recurse(0, 1)

Copy after login

Creating a Valid Python Function

This code traverses the tree recursively, printing out each conditional split and threshold. The result is a valid Python function that effectively emulates the decision-making process of the trained decision tree.

Example Output

For instance, consider a tree that attempts to return its input, a number between 0 and 10. The generated Python function would look like this:

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

def tree(f0):

  if f0 <= 6.0:

    if f0 <= 1.5:

      return [[ 0.]]

    else:  # if f0 > 1.5

      if f0 <= 4.5:

        if f0 <= 3.5:

          return [[ 3.]]

        else:  # if f0 > 3.5

          return [[ 4.]]

      else:  # if f0 > 4.5

        return [[ 5.]]

  else:  # if f0 > 6.0

    if f0 <= 8.5:

      if f0 <= 7.5:

        return [[ 7.]]

      else:  # if f0 > 7.5

        return [[ 8.]]

    else:  # if f0 > 8.5

      return [[ 9.]]

Copy after login

Benefits and Cautions

This method provides a clear and testable representation of the tree's decision rules. However, note that the code assumes that each node in the tree is a binary decision node. If your decision tree contains non-binary decision nodes, you will need to adapt the code accordingly.

The above is the detailed content of How to Extract Decision Rules from a Scikit-Learn Decision Tree?. For more information, please follow other related articles on the PHP Chinese website!

Statement of this Website
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn

Hot AI Tools

Undresser.AI Undress

Undresser.AI Undress

AI-powered app for creating realistic nude photos

AI Clothes Remover

AI Clothes Remover

Online AI tool for removing clothes from photos.

Undress AI Tool

Undress AI Tool

Undress images for free

Clothoff.io

Clothoff.io

AI clothes remover

Video Face Swap

Video Face Swap

Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Tools

Notepad++7.3.1

Notepad++7.3.1

Easy-to-use and free code editor

SublimeText3 Chinese version

SublimeText3 Chinese version

Chinese version, very easy to use

Zend Studio 13.0.1

Zend Studio 13.0.1

Powerful PHP integrated development environment

Dreamweaver CS6

Dreamweaver CS6

Visual web development tools

SublimeText3 Mac version

SublimeText3 Mac version

God-level code editing software (SublimeText3)

Hot Topics

Java Tutorial
1653
14
PHP Tutorial
1251
29
C# Tutorial
1224
24
Python vs. C  : Applications and Use Cases Compared Python vs. C : Applications and Use Cases Compared Apr 12, 2025 am 12:01 AM

Python is suitable for data science, web development and automation tasks, while C is suitable for system programming, game development and embedded systems. Python is known for its simplicity and powerful ecosystem, while C is known for its high performance and underlying control capabilities.

How Much Python Can You Learn in 2 Hours? How Much Python Can You Learn in 2 Hours? Apr 09, 2025 pm 04:33 PM

You can learn the basics of Python within two hours. 1. Learn variables and data types, 2. Master control structures such as if statements and loops, 3. Understand the definition and use of functions. These will help you start writing simple Python programs.

Python: Games, GUIs, and More Python: Games, GUIs, and More Apr 13, 2025 am 12:14 AM

Python excels in gaming and GUI development. 1) Game development uses Pygame, providing drawing, audio and other functions, which are suitable for creating 2D games. 2) GUI development can choose Tkinter or PyQt. Tkinter is simple and easy to use, PyQt has rich functions and is suitable for professional development.

The 2-Hour Python Plan: A Realistic Approach The 2-Hour Python Plan: A Realistic Approach Apr 11, 2025 am 12:04 AM

You can learn basic programming concepts and skills of Python within 2 hours. 1. Learn variables and data types, 2. Master control flow (conditional statements and loops), 3. Understand the definition and use of functions, 4. Quickly get started with Python programming through simple examples and code snippets.

Python vs. C  : Learning Curves and Ease of Use Python vs. C : Learning Curves and Ease of Use Apr 19, 2025 am 12:20 AM

Python is easier to learn and use, while C is more powerful but complex. 1. Python syntax is concise and suitable for beginners. Dynamic typing and automatic memory management make it easy to use, but may cause runtime errors. 2.C provides low-level control and advanced features, suitable for high-performance applications, but has a high learning threshold and requires manual memory and type safety management.

Python: Exploring Its Primary Applications Python: Exploring Its Primary Applications Apr 10, 2025 am 09:41 AM

Python is widely used in the fields of web development, data science, machine learning, automation and scripting. 1) In web development, Django and Flask frameworks simplify the development process. 2) In the fields of data science and machine learning, NumPy, Pandas, Scikit-learn and TensorFlow libraries provide strong support. 3) In terms of automation and scripting, Python is suitable for tasks such as automated testing and system management.

Python and Time: Making the Most of Your Study Time Python and Time: Making the Most of Your Study Time Apr 14, 2025 am 12:02 AM

To maximize the efficiency of learning Python in a limited time, you can use Python's datetime, time, and schedule modules. 1. The datetime module is used to record and plan learning time. 2. The time module helps to set study and rest time. 3. The schedule module automatically arranges weekly learning tasks.

Python: Automation, Scripting, and Task Management Python: Automation, Scripting, and Task Management Apr 16, 2025 am 12:14 AM

Python excels in automation, scripting, and task management. 1) Automation: File backup is realized through standard libraries such as os and shutil. 2) Script writing: Use the psutil library to monitor system resources. 3) Task management: Use the schedule library to schedule tasks. Python's ease of use and rich library support makes it the preferred tool in these areas.

See all articles