tessl/pypi-scikit-learn

A comprehensive machine learning library providing supervised and unsupervised learning algorithms with consistent APIs and extensive tools for data preprocessing, model evaluation, and deployment.

0.98x

Overview

Eval results

Files

Anomaly Analysis Toolkit

Name: tessl/pypi-scikit-learn
Rating: 0.87 (1 reviews)
Author: tessl

Build a module that learns unsupervised detectors for global and local anomalies, estimates robust covariance distances, and discovers biclusters in a rectangular dataset.

Capabilities

Global anomaly scoring

Fitting on 2D numeric data that mixes a tight cluster with two faraway points produces boolean predictions that flag the distant points and keep the cluster points as normal @test

Local neighborhood detection

After training on points that mostly share similar nearest neighbors plus a single remote point, the local anomaly score for the remote point is strictly higher than every inlier score @test

Robust covariance distances

Computing robust Mahalanobis-style distances from a fitted covariance model yields a markedly larger distance for an outlier point than for a typical cluster point @test

Bicluster segmentation

Fitting biclusters on a matrix containing a high-valued block distinct from its surroundings returns row and column labels that separate that block from the remaining rows and columns @test

Implementation

@generates

API

import numpy as np
from typing import Tuple

class AnomalySuite:
    def fit_global(self, samples: np.ndarray, contamination: float = 0.1) -> "AnomalySuite":
        """Fit a global anomaly detector on 2D numeric samples."""

    def predict_global(self, samples: np.ndarray) -> np.ndarray:
        """Return a boolean array marking anomalies in the provided samples."""

    def fit_local(self, samples: np.ndarray, neighbors: int = 20) -> "AnomalySuite":
        """Fit a neighborhood-sensitive anomaly detector."""

    def score_local(self, samples: np.ndarray) -> np.ndarray:
        """Return anomaly scores where higher values indicate stronger outlierness."""

    def fit_covariance(self, samples: np.ndarray) -> "AnomalySuite":
        """Estimate a robust covariance model from the samples."""

    def mahalanobis_distance(self, samples: np.ndarray) -> np.ndarray:
        """Compute robust Mahalanobis-style distances for each sample."""

    def fit_biclusters(self, matrix: np.ndarray, n_clusters: int) -> "AnomalySuite":
        """Fit biclusters on a 2D matrix representing feature-by-sample signals."""

    def bicluster_assignments(self) -> Tuple[np.ndarray, np.ndarray]:
        """Return row and column cluster labels from the last biclustering fit."""

Dependencies { .dependencies }

scikit-learn { .dependency }

Provides unsupervised anomaly detection, covariance estimation, and biclustering utilities.

Install with Tessl CLI

npx tessl i tessl/pypi-scikit-learn