quantics 0.2.0

No project description provided

QUANTICS Class Documentation

Overview

The QUANTICS class provides methods for data normalization, dimensionality reduction, clustering, and extracting representative samples from clustered data.

Methods

__init__(self, dataset)

Initializes the QUANTICS object with a dataset.

Parameters:

• dataset: The input dataset, which must be either a numpy array or a pandas dataframe.

Raises:

• ValueError: If the dataset is neither a numpy array nor a pandas dataframe.

normalize(self, normalize_method='z-score', **args)

Normalizes the dataset.

Parameters:

• normalize_method: The normalization method to use. It can be either 'z-score' or 'minmax'. (default is 'z-score')
• **args: Additional arguments to pass to the scaler.

Raises:

• ValueError: If an invalid normalization method is provided.

reduce_dim(self, reduction_method='pca', dim_size=2, **args)

Performs dimensionality reduction on the normalized dataset.

Parameters:

• reduction_method: The reduction method to use. It can be either 'pca', 'tsne', or 'umap'. (default is 'pca')
• dim_size: The number of dimensions to reduce to. (default is 2)
• **args: Additional arguments to pass to the reduction algorithm.

Raises:

• ValueError: If an invalid reduction method is provided.

cluster(self, min_k=2, max_k=10, **args)

Clusters the reduced data using KMeans and determines the best number of clusters based on the maximum silhouette score.

Parameters:

• min_k: The minimum number of clusters to consider. (default is 2)
• max_k: The maximum number of clusters to consider. (default is 10)
• **args: Additional arguments to pass to the KMeans algorithm.

Raises:

• ValueError: If the reduced data is not available.

get_representative_samples(self, no_samples=5)

Extracts representative samples based on the clustered data.

Parameters:

• no_samples: The number of representative samples to extract. (default is 5)

Returns:

• A subset of the original dataset containing the representative samples.

Raises:

• ValueError: If no clustering model has been fit.

Example Usage

import numpy as np
import pandas as pd
from quantics import QUANTICS

dataset = np.random.rand(100, 5)
processor = QUANTICS(dataset)
processor.normalize(normalize_method='Z-Score', with_mean=False)
processor.reduce_dim(reduction_method='PCA', dim_size=2, svd_solver='full')
processor.cluster(min_k=2, max_k=10, random_state=42)
samples = processor.get_representative_samples(no_samples=5)
print(samples)