rsclassifier 1.5.1

Package for training rule set classifiers for tabular data.

rsclassifier

Overview

This package consists of two modules, rsclassifier and discretization. The first one implements a rule-based machine learning algorithm, while the second one implements an entropy-based supervised discretization algorithm and a class for booleanizing data.

Installation

To install the package, you can simply use pip:

pip install rsclassifier

First module: rsclassifier

This module contains the class RuleSetClassifier, which is a non-parametric supervised learning method that can be used for classification and data mining. As the name suggests, RuleSetClassifier produces classifiers which consist of a set of rules which are learned from the given data. As a concrete example, the following classifier was produced from the well-known Iris data set.

IF
(petal_length_in_cm > 2.45 AND petal_width_in_cm > 1.75) {support: 33, confidence: 0.97}
THEN virginica
ELSE IF
(petal_length_in_cm <= 2.45 AND petal_width_in_cm <= 1.75) {support: 31, confidence: 1.00}
THEN setosa
ELSE versicolor

Notice that each rule is accompanied by:

• Support: The number of data points that satisfy the rule.
• Confidence: The probability that a data point satisfying the rule is correctly classified.

As another concrete example, the following classifier was produced from the Breast Cancer Wisconsin data set.

IF
(bare_nuclei > 2.50 AND clump_thickness > 4.50) {support: 134, confidence: 0.94}
OR (uniformity_of_cell_size > 3.50) {support: 150, confidence: 0.94}
OR (bare_nuclei > 5.50) {support: 119, confidence: 0.97}
THEN 4
ELSE 2

This classifier classifies all tumors which satisfy one of the four rules listed above as malign (4) and all other tumors as benign (2).

Advantages

• RuleSetClassifier produces extremely interpretable and transparent classifiers.
• It is very easy to use, as it has only one main hyperparameter num_prop.
• It can handle both categorical and numerical data.
• The learning process is very fast.

How to use RuleSetClassifier

Let rsc be an instance of RuleSetClassifier, X be a pandas dataframe (input features) and y a pandas series (target labels).

• Load the data: Use rsc.load_data(X, y, boolean, categorical, numerical) where boolean, categorical and numerical are (possibly empty) lists specifying which features in X are boolean, categorical or numerical, respectively. This function converts the data into a Boolean form for rule learning and store it in rsc.
• Fit the classifier: After loading the data, call rsc.fit(num_prop, fs_algorithm, growth_size). Note that unlike in scikit-learn, this function doesn't take X and y directly as arguments; they are loaded beforehand as part of load_data. The hyperparameters num_prop, fs_algorithm and growth_size work as follows.
  • num_prop is an upper bound on the number of proposition symbols allowed in the rules. The smaller num_prop is, the more interpretable the models are. The downside of having small num_prop is of course that the resulting model has low accuracy (i.e., it underfits), so an optimal value for num_prop is the one which strikes a balance between interpretability and accuracy.
  • fs_algorithm determines the algorithm used for selecting the Boolean features used by the classifier. It has two options: dt (which is the default) and brute. dt uses decision trees for feature selection, brute finds the set of features for which the error on training data is minimized. Note that running brute with a large num_prop can take a long time.
  • growth_size is a float in the range (0, 1], determining the proportion of X used for learning rules. The remaining portion is used for pruning. If growth_size is set to 1, which is the default value, no pruning is performed. Also 2/3 seems to work well in practice.
• Make predictions: Use rsc.predict(X) to generate predictions. This function returns a pandas Series.
• Visualize the classifier: Simply print the classifier to visualize the learned rules (together with their support and confidence).

Note: At present, RuleSetClassifier does not support datasets with missing values. You will need to preprocess your data (e.g., removing missing values) before using the classifier.

Background

The rule learning method implemented by RuleSetClassifier was inspired by and extends the approach taken in the paper, which we refer to here as the ideal DNF-method. The ideal DNF-method goes as follows. First, the input data is Booleanized. Then, a small number of promising features is selected. Finally, a DNF-formula is computed for those promising features for which the number of misclassified points is as small as possible.

The way RuleSetClassifier extends and modifies the ideal DNF-method is primarily as follows.

• We use an entropy-based Booleanization for numerical features with minimum description length principle working as a stopping rule.
• RuleSetClassifier is not restricted to binary classification tasks.
• We use the Quine-McCluskey algorithm for finding near-optimal size DNF-formulas.
• We also implement rule pruning as a postprocessing step. This is important because it makes the rules shorter and hence more interpretable.

Example

import pandas as pd
from sklearn import datasets
from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split
from rsclassifier import RuleSetClassifier

# Load the data set.
iris = datasets.load_iris()
df = pd.DataFrame(data= iris.data, columns= iris.feature_names)
df['target'] = iris.target

# Split it into train and test.
X = df.drop(columns = ['target'], axis = 1)
y = df['target']
X_train, X_test, y_train, y_test = train_test_split(X, y, train_size = 0.8)

# Initialize RuleSetClassifier.
rsc = RuleSetClassifier()
# All the features of iris.csv are numerical.
rsc.load_data(X = X_train, y = y_train, numerical = X.columns)
# Fit the classifier with a maximum of 2 proposition symbols.
rsc.fit(num_prop = 2)

# Measure the accuracy of the resulting classifier.
train_accuracy = accuracy_score(rsc.predict(X_train), y_train)
test_accuracy = accuracy_score(rsc.predict(X_test), y_test)

# Display the classifier and its accuracies.
print()
print(rsc)
print(f'Rule set classifier training accuracy: {train_accuracy}')
print(f'Rule set classifier test accuracy: {test_accuracy}')

Second module: discretization

This module contains the find_pivots function and the Booleanizer class.

find_pivots

find_pivots can be used for entropy-based supervised discretization of numeric features. This is the function that RuleSetClassifier and Booleanizer use for Booleanizing numerical data.

How does it work?

At a high level, the algorithm behind find_pivots works as follows:

1. Sorting: The feature column x is sorted to ensure that potential pivots represent transitions between distinct data values.
2. Candidate Pivot Calculation: Midpoints between consecutive unique values in the sorted list are calculated as candidate pivots.
3. Split Evaluation: Each candidate pivot is evaluated by splitting the dataset into two subsets:
   • One subset contains records with feature values ≤ the pivot.
   • The other subset contains records with feature values > the pivot.
4. Information Gain Calculation: Information gain is calculated to assess the quality of each split.
5. Recursion: If a split significantly increases information gain, the process is recursively applied to each subset until no further significant gains can be achieved.

For more details, see Section 7.2 of Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations by Ian H. Witten and Eibe Frank.

Example:

import pandas as pd
from sklearn import datasets
from discretization import find_pivots

# Load the dataset
iris = datasets.load_iris()
df = pd.DataFrame(data=iris.data, columns=iris.feature_names)
df['target'] = iris.target

# Calculate pivots for the feature "petal length (cm)"
pivots = find_pivots(df['petal length (cm)'], df['target'])
print(pivots)  # Output: [2.45, 4.75]

Booleanizer

The Booleanizer class is designed to transform a dataset into a booleanized format for classification purposes. The booleanized data is obtained by applying one-hot encoding to categorical features and splitting numerical features into boolean columns based on learned pivot values.

How to use:

Let booleanizer be an instance of Booleanizer, X a pandas dataframe (input features) and y a pandas series (target labels).

• Collect unique values for categorical features: For categorical features we need to store the unique values (or classes) in each categorical feature. This step is done by calling the store_classes_for_cat_features method.
• Learn pivots for numerical features: For the numerical features, Booleanizer uses entropy-based discretization to learn pivot points which it will then store. This step is done using the store_pivots_for_num_features method.
• Booleanize the data: After storing the categories for categorical features and the pivot points for numerical features, the data can be booleanized using the booleanize_dataframe method.

Note that a single instance of Booleanizer can be used to booleanize several different datasets. In particular, a Booleanizer trained on the training data can be used to booleanize the test data.

Example:

import pandas as pd
from sklearn import datasets
from sklearn.model_selection import train_test_split
from discretization import Booleanizer

# Load the data.
iris = datasets.load_iris()
X = pd.DataFrame(data=iris.data, columns=iris.feature_names)
y = pd.Series(iris.target)

# Split data into train and test sets.
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)

# Initialize the booleanizer.
booleanizer = Booleanizer()

# Learn pivots from the training data.
booleanizer.store_pivots_for_num_features(X_train, y_train, X.columns)

# Booleanize the training data and the test data using the same pivots.
X_train_bool = booleanizer.booleanize_dataframe(X_train)
X_test_bool = booleanizer.booleanize_dataframe(X_test)