Karate Club

https://karateclub.readthedocs.io/en/latest/
Admin
Karate Club is an unsupervised machine learning extension library for NetworkX that provides graph...
Tokens:76,644
Snippets:579
Trust Score:9.6
Update:4 months ago
Show doc for...
Tokens:
Raw
### Walklets Embedding Example

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

Shows how to create a Walklets embedding with minimal changes from the DeepWalk example, highlighting the API-driven design.

```python
import networkx as nx
from karateclub.node_embedding.neighbourhood import Walklets

g = nx.newman_watts_strogatz_graph(100, 20, 0.05)

model = Walklets()
model.fit(g)
embedding = model.get_embedding()
```

--------------------------------

### Install Karate Club

Source: https://karateclub.readthedocs.io/en/latest/notes/installation.html

Install the package by using pip:

```bash
pip install karateclub

```

--------------------------------

### DeepWalk Embedding Example

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

Demonstrates how to create a DeepWalk embedding for a Watts-Strogatz graph using the Karate Club library.

```python
import networkx as nx
from karateclub import DeepWalk

g = nx.newman_watts_strogatz_graph(100, 20, 0.05)

model = DeepWalk()
model.fit(g)
embedding = model.get_embedding()
```

--------------------------------

### Upgrade Karate Club

Source: https://karateclub.readthedocs.io/en/latest/notes/installation.html

Upgrade your outdated Karate Club version by using:

```bash
pip install karateclub --upgrade

```

--------------------------------

### MNMF Example

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/community_detection/overlapping/mnmf.html

This example demonstrates how to use the MNMF algorithm to detect overlapping communities in a graph.

```python
import networkx as nx
from karateclub import MNMF

# Load a graph (example: Zachary's Karate Club)
G = nx.karate_club_graph()

# Initialize and train the MNMF model
model = MNMF(dimensions=16, iteration=20, lambd=1.0, gamma=0.1)
model.fit(G)

# Get the community embeddings
embeddings = model.get_embedding()

# Get the detected communities (overlapping)
communities = model.get_communities()

print("Community Embeddings:", embeddings)
print("Detected Communities:", communities)
```

--------------------------------

### Check Karate Club version

Source: https://karateclub.readthedocs.io/en/latest/notes/installation.html

To check your current package version just simply run:

```bash
pip freeze | grep karateclub

```

--------------------------------

### Setup Q Method

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/community_detection/overlapping/danmf.html

Method to set up the Q matrices, which are used in the update steps.

```python
def _setup_Q(self):
        """
        Setting up Q matrices.
        """
        self._Q_s = [None for _ in range(self._p + 1)]
        self._Q_s[self._p] = np.eye(self.layers[self._p - 1])
        for i in range(self._p - 1, -1, -1):
            self._Q_s[i] = np.dot(self._U_s[i], self._Q_s[i + 1])
```

--------------------------------

### Setup Target Matrices Method

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/community_detection/overlapping/danmf.html

Method to set up the target matrices (adjacency, Laplacian, and degree) for the pre-training process.

```python
def _setup_target_matrices(self, graph):
        """
        Setup target matrix for pre-training process.

        Arg types:
            * **graph** *(NetworkX graph)* - The graph being clustered.
        """
        self._graph = graph
        self._A = nx.adjacency_matrix(
            self._graph, nodelist=range(self._graph.number_of_nodes())
        )
        self._L = nx.laplacian_matrix(
            self._graph, nodelist=range(self._graph.number_of_nodes())
        )
        self._D = self._L + self._A
```

--------------------------------

### GEMSEC Helper Methods

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/community_detection/non_overlapping/gemsec.html

Internal helper methods for GEMSEC, including setup for sampling weights, initializing embeddings and cluster centers, sampling negative examples, and calculating noise and cluster vectors.

```python
    def _setup_sampling_weights(self, graph):
        """
        Creating a negative sampling table.

        Arg types:
            * **graph** *(NetworkX graph)* - The graph for negative sampling.
        """
        self._sampler = {}
        index = 0
        for node in graph.nodes():
            for _ in range(graph.degree(node)):
                self._sampler[index] = node
                index = index + 1
        self._global_index = index - 1

    def _initialize_node_embeddings(self, graph):
        """
        Creating a node embedding array.

        Arg types:
            * **graph** *(NetworkX graph)* - The graph for negative sampling.
        """
        shape = (graph.number_of_nodes(), self.dimensions)
        self._base_embedding = np.random.normal(0, 1.0 / self.dimensions, shape)

    def _initialize_cluster_centers(self, graph):
        """
        Creating a cluster center array.

        Arg types:
            * **graph** *(NetworkX graph)* - The graph for negative sampling.
        """
        shape = (self.dimensions, self.clusters)
        self._cluster_centers = np.random.normal(0, 1.0 / self.dimensions, shape)

    def _sample_negative_samples(self):
        """
        Sampling a batch of nodes as negative samples.

        Return types:
            * **negative_samples** *(list)*: List of negative sampled nodes.
        """
        negative_samples = [
            self._sampler[random.randint(0, self._global_index)]
            for _ in range(self.negative_samples)
        ]
        return negative_samples

    def _calculcate_noise_vector(self, negative_samples, source_node):
        """
        Getting the noise vector for the weight update.

        Arg types:
            * **negative_samples** *(list)*: List of negative sampled nodes.
            * **source_node** *(int)* - Source node in the walk.

        Return types:
            * **noise_vector** *(NumPy array) - Noise update vector.
        """
        noise_vectors = self._base_embedding[negative_samples, :]
        source_vector = self._base_embedding[int(source_node), :]
        raw_scores = noise_vectors.dot(source_vector.T)
        raw_scores = np.exp(np.clip(raw_scores, -15, 15))
        scores = raw_scores / np.sum(raw_scores)
        scores = scores.reshape(-1, 1)
        noise_vector = np.sum(scores * noise_vectors, axis=0)
        return noise_vector

    def _calculate_cluster_vector(self, source_node):
        """
        Getting the cluster vector for the weight update.

        Arg types:
            * **source_node** *(int)* - Source node in the walk.

        Return types:
            * **cluster_vector** *(NumPy array) - Cluster update vector.
            * **cluster_index** *(int)*: Node cluster membership index.
        """
        distances = (
            self._base_embedding[int(source_node), :].reshape(-1, 1)
            - self._cluster_centers
        )

```

--------------------------------

### Loading the Reddit 10K dataset

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet shows how to load the Reddit 10K dataset using GraphSetReader, obtaining both the graphs and their corresponding target vectors.

```python
from karateclub.dataset import GraphSetReader

reader = GraphSetReader("reddit10k")

graphs = reader.get_graphs()
y = reader.get_target()

```

--------------------------------

### Loading Facebook Dataset for Community Detection

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet shows how to load the Facebook page-page network dataset using GraphReader and retrieve the graph and target data.

```python
from karateclub import GraphReader

reader = GraphReader("facebook")

graph = reader.get_graph()
target = reader.get_target()

```

--------------------------------

### Node Embedding with Diff2Vec

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet demonstrates applying the Diff2Vec model for node embedding with custom parameters and retrieving the embedding features.

```python
from karateclub import Diff2Vec

model = Diff2Vec(diffusion_number=2, diffusion_cover=20, dimensions=16)
model.fit(graph)
X = model.get_embedding()

```

--------------------------------

### TENET Model Example

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/node_embedding/attributed/tene.html

This example demonstrates how to use the TENET model for node embedding. It includes initializing the model, fitting it to the data, and obtaining node embeddings.

```python
from karateclub.node_embedding.attributed import TENET
from karateclub.utils.data import load_data

# Load a sample dataset
adj, features, labels = load_data("cora")

# Initialize the TENET model
model = TENET(dimensions=128, 
              epochs=10, 
              learning_rate=0.01, 
              decay=0.001, 
              lambd=1.0, 
              nu=1.0, 
              seed=42)

# Fit the model to the data
model.fit(adj, features)

# Get node embeddings
embeddings = model.get_embedding()

print(f"Shape of embeddings: {embeddings.shape}")
```

--------------------------------

### Fitting a FEATHER graph level embedding

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet demonstrates fitting a FEATHER graph embedding model to the loaded graphs and retrieving the resulting graph embeddings.

```python
from karateclub import FeatherGraph

model = FeatherGraph()
model.fit(graphs)
X = model.get_embedding()

```

--------------------------------

### Loading Twitch UK Dataset for Node Embedding

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet shows how to load the Twitch UK user friendship dataset using GraphReader and retrieve the graph and target data.

```python
from karateclub.dataset import GraphReader

reader = GraphReader("twitch")

graph = reader.get_graph()
y = reader.get_target()

```

--------------------------------

### get_embedding

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/graph_embedding/waveletcharacteristic.html

Getting the embedding of graphs.

```python
def get_embedding(self) -> np.array:
        r"""Getting the embedding of graphs.

        Return types:
            * **embedding** *(Numpy array)* - The embedding of graphs.
        """
        return np.array(self._embedding)
```

--------------------------------

### NetLSD get_embedding method

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/graph_embedding/netlsd.html

Getting the embedding of graphs.

```python
def get_embedding(self) -> np.array:
    r"""Getting the embedding of graphs.

    Return types:
        * **embedding** *(Numpy array)* - The embedding of graphs.
    """
    return np.array(self._embedding)
```

--------------------------------

### Training a logistic regression model and evaluating performance

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet demonstrates training a logistic regression model on the training data, predicting probabilities on the test set, and evaluating performance using the ROC AUC score.

```python
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression

downstream_model = LogisticRegression(random_state=0).fit(X_train, y_train)
y_hat = downstream_model.predict_proba(X_test)[:, 1]
auc = roc_auc_score(y_test, y_hat)
print('AUC: {:.4f}'.format(auc))
>>> AUC: 0.7127

```

--------------------------------

### Setup Base Model

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/node_embedding/neighbourhood/boostne.html

Initializes the base model by setting up the shape, indices, and performing the initial NMF fit and score.

```python
def _setup_base_model(self):
        """
        Fitting NMF on the starting matrix.
        """
        self._shape = self._residuals.shape
        indices = self._residuals.nonzero()
        self._index_1 = indices[0]
        self._index_2 = indices[1]
        base_score, embedding = self._fit_and_score_NMF(self._residuals)
        self._embeddings = [embedding]
```

--------------------------------

### NMF-ADMM get_embedding method

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/node_embedding/neighbourhood/nmfadmm.html

Getting the node embedding.

```python
def get_embedding(self) -> np.array:
        r"""Getting the node embedding.

        Return types:
            * **embedding** *(Numpy array)* - The embedding of nodes.
        """
        embedding = np.concatenate([self._W_plus, self._H_plus.T], axis=1)
        return embedding
```

--------------------------------

### Get Embedding Method

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/graph_embedding/gl2vec.html

Retrieves the computed graph embeddings.

```python
    def get_embedding(self) -> np.array:
        r"""Getting the embedding of graphs.

        Return types:
            * **embedding** *(Numpy array)* - The embedding of graphs.
        """
        return np.array(self._embedding)

```

--------------------------------

### Clustering a Random Watts-Strogatz Graph

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet illustrates clustering an arbitrary NetworkX graph, such as a randomly generated Watts-Strogatz graph, using Label Propagation.

```python
import networkx as nx
from karateclub import LabelPropagation

graph = nx.newman_watts_strogatz_graph(100, 20, 0.05)

model = LabelPropagation()
model.fit(graph)
cluster_membership = model.get_memberships()

```

--------------------------------

### GraphWave get_embedding method

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/node_embedding/structural/graphwave.html

Getting the node embedding.

```python
    def get_embedding(self) -> np.array:
        r"""Getting the node embedding.

        Return types:
            * **embedding** *(Numpy array)* - The embedding of nodes.
        """
        embedding = [self._real_and_imaginary.real, self._real_and_imaginary.imag]
        embedding = np.concatenate(embedding, axis=1)
        return embedding

```

--------------------------------

### Evaluating Clustering with Normalized Mutual Information

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet shows how to evaluate the quality of the detected clusters using normalized mutual information (NMI) against the ground truth.

```python
from sklearn.metrics.cluster import normalized_mutual_info_score

cluster_membership = [cluster_membership[node] for node in range(len(cluster_membership))]

nmi = normalized_mutual_info_score(target, cluster_membership)
print('NMI: {:.4f}'.format(nmi))

```

--------------------------------

### Getting the node embedding

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/node_embedding/attributed/fscnmf.html

This method returns the learned node embedding.

```python
embedding = np.concatenate([self._B_1, self._U], axis=1)
return embedding
```

--------------------------------

### Role2Vec Embedding

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/node_embedding/structural/role2vec.html

This snippet shows how to get the node embedding after the model has been trained.

```python
alpha=self.learning_rate,
            seed=self.seed,
        )

        self._embedding = [model.dv[str(i)] for i, _ in enumerate(documents)]
```

--------------------------------

### Logistic Regression for Abusive User Prediction

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet demonstrates training a logistic regression model on the node embeddings and evaluating its performance using AUC.

```python
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression

downstream_model = LogisticRegression(random_state=0).fit(X_train, y_train)
y_hat = downstream_model.predict_proba(X_test)[:, 1]
auc = roc_auc_score(y_test, y_hat)
print('AUC: {:.4f}'.format(auc))

```

--------------------------------

### Get Embedding

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/community_detection/overlapping/mnmf.html

This method returns the node embedding, which is stored in the U matrix.

```python
def get_embedding(self) -> np.array:
        r"""Getting the node embedding.

        Return types:
            * **embedding** *(Numpy array)* - The embedding of nodes.
        """
        embedding = self._U
        return embedding
```

--------------------------------

### Community Detection with Label Propagation

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet demonstrates applying the Label Propagation algorithm for community detection on a graph and obtaining cluster memberships.

```python
from karateclub import LabelPropagation

model = LabelPropagation()
model.fit(graph)
cluster_membership = model.get_memberships()

```

--------------------------------

### GEMSEC Get Memberships Method

Source: https://karateclub.readthedocs.io/en/latest/modules/root.html

Retrieves the cluster memberships for each node from the GEMSEC model.

```python
get_memberships() -> Dict[int, int]
```

--------------------------------

### Get Embedding

Source: https://karateclub.readthedocs.io/en/latest/_modules/karateclub/node_embedding/neighbourhood/boostne.html

Retrieves the final node embedding by concatenating all learned embeddings.

```python
def get_embedding(self) -> np.array:
        r"""Getting the node embedding.

        Return types:
            * **embedding** *(Numpy array)* - The embedding of nodes.
        """
        embedding = np.concatenate(self._embeddings, axis=1)
        return embedding
```

--------------------------------

### LabelPropagation Class Initialization

Source: https://karateclub.readthedocs.io/en/latest/modules/root.html

Initializes the Label Propagation model for community detection.

```python
LabelPropagation(seed: int = 42, iterations: int = 100)
```

--------------------------------

### Train-Test Split for Node Embedding Features

Source: https://karateclub.readthedocs.io/en/latest/notes/introduction.html

This snippet shows how to split the node embedding features and target variable into training and testing sets for downstream classification.

```python
from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

```
Tokens:
Raw