Concept Clustering package

The concept clustering method separates data instances into groups based on their similarity in multiple description spaces, i.e. sets of features.

Installation

• install via pip: pip install .

Usage

Checkout the example scripts in ./examples/.

Basic usage:

import pandas as pd
from concept_clustering.concept_clustering import ConceptClustering

# define description spaces
features_per_space = [
    ["Investment costs"],
    ["Yearly total costs", "posResilience"],
]
list_of_features = [item for sublist in features_per_space for item in sublist]

# set number of clusters
num_clusters = 3

# get data
data = pd.read_csv("./energy.csv", usecols=list_of_features)

# initialize and fit ConceptClustering
ConClus = ConceptClustering(
    description_spaces=features_per_space,
    n_clusters=num_clusters,
    max_iter=100,
).fit(X=data, centers="k")

print(ConClus.concepts_)