Skip to content

Graph Filtering

GraphFiltering

GraphFiltering Class

This class implements graph-based filtering techniques for recommendation systems. Graph-based filtering leverages the structure of a graph to make recommendations by analyzing the relationships between users and items.

Attributes:

Name Type Description
graph Graph

The graph structure representing users and items.
similarity_metric str

The metric used to calculate similarity between nodes.
damping_factor float

The damping factor used in algorithms like PageRank.
max_iterations int

Maximum number of iterations for convergence in iterative algorithms.

Methods:

Name Description
build_graph

Constructs the graph from the given data, where nodes represent users and items, and edges represent interactions or similarities.
compute_similarity

Computes similarity scores between nodes using the specified similarity metric.
recommend

Generates top-N recommendations for a given user by analyzing the graph structure.
update_graph

Updates the graph with new interactions or changes in the data, allowing for dynamic recommendations.
Source code in engines/contentFilterEngine/graph_based_algorithms/graph_filtering.py 
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
class GraphFiltering:
    """
    GraphFiltering Class

    This class implements graph-based filtering techniques for recommendation systems.
    Graph-based filtering leverages the structure of a graph to make recommendations by
    analyzing the relationships between users and items.

    Attributes:
        graph (Graph): The graph structure representing users and items.
        similarity_metric (str): The metric used to calculate similarity between nodes.
        damping_factor (float): The damping factor used in algorithms like PageRank.
        max_iterations (int): Maximum number of iterations for convergence in iterative algorithms.

    Methods:
        build_graph(data):
            Constructs the graph from the given data, where nodes represent users and items,
            and edges represent interactions or similarities.

        compute_similarity():
            Computes similarity scores between nodes using the specified similarity metric.

        recommend(user_id, top_n=10):
            Generates top-N recommendations for a given user by analyzing the graph structure.

        update_graph(new_data):
            Updates the graph with new interactions or changes in the data, allowing for
            dynamic recommendations.
    """
    def __init__(self):
        self.graph = None

    def generate_graph(self, num_people, file_path="graph_dataset.csv", seed=None):
        """
        Generate a random graph and save it to a file.
        """
        self.graph = generate_random_graph(num_people, file_path, seed)
        return self.graph

    def generate_large_graph(self, num_people, file_path="large_random_graph.csv", seed=None):
        """
        Generate a large random graph using multiprocessing.
        """
        self.graph = generate_large_random_graph(num_people, file_path, seed)
        return self.graph

    def scale_and_save(self, input_file, output_dir, num_matrices):
        """
        Scale and save matrices to the specified directory.
        """
        scale_and_save_matrices(input_file, output_dir, num_matrices)

generate_graph(num_people, file_path='graph_dataset.csv', seed=None)

Generate a random graph and save it to a file.

Source code in engines/contentFilterEngine/graph_based_algorithms/graph_filtering.py 
40
41
42
43
44
45
def generate_graph(self, num_people, file_path="graph_dataset.csv", seed=None):
    """
    Generate a random graph and save it to a file.
    """
    self.graph = generate_random_graph(num_people, file_path, seed)
    return self.graph

generate_large_graph(num_people, file_path='large_random_graph.csv', seed=None)

Generate a large random graph using multiprocessing.

Source code in engines/contentFilterEngine/graph_based_algorithms/graph_filtering.py 
47
48
49
50
51
52
def generate_large_graph(self, num_people, file_path="large_random_graph.csv", seed=None):
    """
    Generate a large random graph using multiprocessing.
    """
    self.graph = generate_large_random_graph(num_people, file_path, seed)
    return self.graph

scale_and_save(input_file, output_dir, num_matrices)

Scale and save matrices to the specified directory.

Source code in engines/contentFilterEngine/graph_based_algorithms/graph_filtering.py 
54
55
56
57
58
def scale_and_save(self, input_file, output_dir, num_matrices):
    """
    Scale and save matrices to the specified directory.
    """
    scale_and_save_matrices(input_file, output_dir, num_matrices)