Various uses and implementation examples of the Knowledge Graph

Machine Learning Artificial Intelligence Natural Language Processing Semantic Web Python Collecting AI Conference Papers Deep Learning Ontology Technology Digital Transformation Knowledge Information Processing Economy and Business Physics & Mathematics Navigate This blog

Knowledge Graph

A knowledge graph is a graph structure that represents information as a set of related nodes (vertices) and edges (connections), and is a data structure used to connect information on different subjects or domains and visualize their relationships.

Features and advantages of the knowledge graph include the following

Visualization of relevance: Knowledge graphs help visualize and intuitively understand the relevance of information, as described in “Relational Data Learning. The visual representation of the graph also makes it easier to grasp complex relationships and patterns.
Support for causal and other inferences: Knowledge graphs can be used to support causal and other inferences, such as those described in “Statistical Causal Inference and Causal Search” enabling new knowledge and insights by exploring routes and paths on the graph, applying inference rules, etc.
Question answering and decision support: Knowledge graphs can be applied to question answering systems, as described in “Applications of Knowledge Graphs to Question Answering Systems” or to decision support systems. This would involve using the information in the graph to find answers to specific questions or to provide decision support.
Integration of Domain Knowledge: Knowledge graphs can help integrate domain knowledge by centralizing different types of knowledge information as described in “Knowledge Information Processing Techniques” and “Ontology Techniques. By linking information from different data sources and domains in this way, the graph allows for easy reference to relevant information.
Integration with Machine Learning: Knowledge graphs can also be used in conjunction with various machine learning models, such as those described in “Overview of Relational Data Learning and Examples of Applications and Implementations” to use information on the graph as features or to gain insights using graph analysis techniques.

Knowledge graphs are powerful tools for connecting information and visualizing relationships, and are an important means for integrating domain knowledge, structuring information, and gaining insight.

Knowledge Graph Application Examples

Knowledge graphs are widely used in a variety of areas. Some specific examples of knowledge graph applications are described below.

Natural Language Processing (NLP): The knowledge graph is very useful in natural language processing tasks, for example, a question answering system can extract relevant information from the knowledge graph and use it to appropriately answer user questions. The knowledge graph can be used to effectively handle contextual and relevant information.
Recommendation systems: The knowledge graph is utilized in recommendation systems to express relevance among items. Specifically, individual recommendations can be generated based on user preferences and past behavior, taking into account the connections and paths between nodes in the graph.
Social Network Analysis: The Knowledge Graph is useful for analyzing relationships and connections in social networks. With nodes on the graph representing people and organizations, and edges representing relationships and connections, it can be applied to social network analysis, including information diffusion, measuring influence, and community detection.
Life Science: Knowledge graphs are also being used in the life science and medical domains. By representing biological entities (genes, proteins, diseases, etc.) and their relationships in graphs and analyzing complex networks, they can contribute to the discovery of new treatments, prediction of side effects, and understanding of disease mechanisms.
Knowledge Management: Knowledge graphs are also used to manage knowledge and information within an organization. Specifically, they can be used to represent data, documents, people’s skills, and relationships within a company as a graph to support information retrieval, sharing, and collaboration within an organization.
Machine Learning Applications: Machine learning is used in a variety of ways to leverage knowledge graphs. Specifically, it can be used to provide deep insights for inference and classification tasks.

By representing the relationships and connections between information, knowledge graphs can be a powerful tool to assist in the analysis and decision-making of small and complex data.

Their details and specific implementations are described below.

Question and answer system using knowledge graph

<Overview>.

Question-and-answer systems using the knowledge graph generate answers to users’ questions using information in the knowledge graph. The following describes the general procedure for constructing a question-and-answer system and specific applications.

Construction of the Knowledge Graph: To construct a question-and-answer system, it is first necessary to build a knowledge graph. This is done by collecting information about the domain and representing it in the nodes and edges of the graph, with the nodes representing entities and concepts and the edges showing their relationships.
Question Analysis: Analyze questions from users and extract information about the intent of the questions. This is done using Natural Language Processing (NLP) technology, which analyzes the question to determine which parts of the question are important keywords and the type of question (Yes/No question, information request, definition, etc.).
Graph Search: Based on the analyzed questions, appropriate information is searched for in the knowledge graph. The goal is to find the desired information by following the edges in the graph, considering keywords and relevance of the question. Graph search algorithms (e.g., path search, shortest path search) are used for the search.
Answer Generation: Answers are generated based on the information obtained from graph search. Answers are provided in the form of text, links, etc. The goal is to generate appropriate answers to questions by extracting information in the knowledge graph or combining related information.

A specific application of question answering is a question answering system based on medical information. A medical knowledge graph stores information about diseases, symptoms, treatments, etc. When a user asks a medical question, the system explores the knowledge graph to retrieve relevant information and provide accurate medical information and advice.

Question-and-answer systems based on knowledge graphs are expected to provide more sophisticated responses because they can answer questions by considering the relevance and context of the information, but they require attention to the accuracy and consistency of data during the process of graph construction and question analysis.

<Example Implementation in Python>

The following is a simple example implementation of a question-and-answer system using a knowledge graph in Python. In this example, a simple knowledge graph is created, and a system is created to return the best answer to a user’s question. First, consider the following knowledge graph.

knowledge_graph = {
    "Question1": "Answer1",
    "Question2": "Answer2",
    "Question3": "Answer3",
    # Add other questions and answers
}

Next, create a function that accepts questions from users and retrieves answers from the knowledge graph.

def search_answer(question, knowledge_graph):
    if question in knowledge_graph:
        return knowledge_graph[question]
    else:
        return "No answer found."

# Ask a question
user_question = input("Enter your question:.")

# Search and display answers
answer = search_answer(user_question, knowledge_graph)
print(answer)

In this example implementation, a question from a user is received as user_question, and the search_answer function searches the knowledge graph for an answer and returns it. If the question does not exist in the knowledge graph, “No answer found. is displayed.

Thus, the implementation of a question-and-answer system using the knowledge graph uses a simple dictionary structure to express the correspondence between questions and answers, and builds a mechanism to return appropriate answers in response to user questions.

Decision support using knowledge graphs

<Overview>

Knowledge graphs are useful tools in decision support systems. The following methods and procedures are commonly used in decision support using knowledge graphs.

Construction of Knowledge Graph: To construct a decision support system, it is necessary to construct a knowledge graph of the target domain. The knowledge graph represents domain concepts, relationships, rules, constraints, and other information in terms of nodes and edges.
Problem modeling: model the specific problem that is the subject of the decision. Modeling involves identifying appropriate nodes and edges in the knowledge graph to represent the elements and relationships of the problem. Modeling methods and representations may vary depending on the characteristics and objectives of the problem.
Knowledge Graph Exploration: Explore information in the knowledge graph to support decision making about the problem. Graph traversal algorithms and query languages are used for traversal. Here, relevant information in the knowledge graph is identified and the information needed for decision making is retrieved.
Analysis and Evaluation: The explored information is analyzed and evaluated for decision making. Utilizing the data and rules in the knowledge graph, factors such as benefits, risks, and constraints are evaluated, and indicators and evaluation values for decision-making are calculated.
Simulation of scenarios: The decision support system allows simulations based on different scenarios and alternatives. By modeling scenarios on the knowledge graph and predicting the outcomes and impacts in each scenario, the system supports the evaluation of decision options and risks.

Decision support using knowledge graphs can contribute to speeding up and improving the quality of the decision-making process by facilitating visualization of information and understanding of complex relationships. Examples of concrete applications include business strategy formulation, risk management, medical diagnosis support, and product development decision-making.

<Example of python implementation>

The following is an example of a Python implementation of a decision support system using a knowledge graph. In this example, a simple decision support system is created to support the decision-making process using a knowledge graph. First, consider the following knowledge graph.

knowledge_graph = {
    "Question1": {
        "Choice1": "Answer of choice1",
        "Choice2": "Answer of choice2",
        # Add other choices and answers
    },
    "Question2": {
        "Choice1": "Answer of Choice1",
        "Choice2": "Answer of Choice2",
        # Add other choices and answers
    },
    "Question3": {
        "Choice1": "Answer of Choice1",
        "Choice2": "Answer of Choice2",
        # Add other choices and answers
    },
    # Add other choices and answers
}

Next, a function is created that presents a question to the user and supports decision-making based on the answer.

def decision_support(knowledge_graph):
    current_node = knowledge_graph
    while isinstance(current_node, dict):
        question = input(current_node["Question"] + ": ")
        if question in current_node:
            current_node = current_node[question]
        else:
            print("Not a valid choice. Please select again.")
    print("Final result: " + current_node)

# Initiate decision support
decision_support(knowledge_graph)

In this implementation example, the decision support system is initiated by calling the decision_support function, and the process of presenting a question to the user, followed by the next question based on the user’s answer, is repeated until the final result is obtained.

In this way, a decision support system using a knowledge graph can express the relationship between hierarchical questions and answers, and can build a mechanism to derive the final result based on the user’s answers.

Knowledge graph-based recommendation system

<Overview>

Knowledge graph-based recommendation systems generate individual recommendations based on a user’s preferences and behavioral history, using relevant information in the knowledge graph. The following describes general knowledge graph-based recommendation system methods and specific applications.

Graph-based filtering: Recommends the best items to users based on the relevance of nodes and edges in the knowledge graph. Graph-based filtering explores connections and paths in the graph and selects relevant items based on user behavior and interests.
User Personalization: The knowledge graph can incorporate user profiles and attribute information. Here, nodes and edges in the graph that are relevant to the user’s attributes and preferences are identified and personalized recommendations are generated based on them.
Graph-based Ranking: Provides ranked recommendations based on the relevance and rating information of items in the knowledge graph. Here, items relevant to the user are explored in the knowledge graph and ranked based on the ratings and relevance of those items.

As specific examples of applications, the recommendation system using the knowledge graph is utilized in the following situations

Online stores: Data on a user’s purchase history, preferences, and other similar users is incorporated into the knowledge graph to recommend highly relevant products. The relationship between products is analyzed on the graph, and products of interest to the user are suggested.
Content delivery platform: Recommend relevant content based on users’ viewing history and preferences. Analyze the relationship between contents on the knowledge graph and provide contents that may be of interest to the user.
Social Network: Recommend content and events shared by friends and connections based on the user’s friendships and interests. Provide relevant information based on the user’s connections and common interests on the Knowledge Graph.

By utilizing the knowledge graph, individual user preferences and relevance can be captured with greater precision, enabling more personalized recommendations.

<Example of implementation in python>

This section describes a simple example of a Python implementation of a recommendation system using the knowledge graph. In this example, items are recommended based on user preferences and choices. First, consider the following knowledge graph.

knowledge_graph = {
    "User1": {
        "preference": ["itemA", "itemB", "itemC"],
        "recommend": ["itemD", "itemE"]
    },
    "User2": {
        "preference": ["itemB", "itemC", "itemD"],
        "recommend": ["itemA", "itemE"]
    },
    # Add other users and their preferences/recommendations
}

Next, create a function that recommends items based on user preferences.

def recommend_items(user, knowledge_graph):
    if user in knowledge_graph:
        user_preferences = knowledge_graph[user]["preference"]
        recommendations = knowledge_graph[user]["recommendation"]
        new_recommendations = [item for item in recommendations if item not in user_preferences]
        if new_recommendations:
            return new_recommendations
        else:
            return recommendations
    else:
        return []

# Recommending items to users
user = input("Please enter your user name:. ")
recommended_items = recommend_items(user, knowledge_graph)
if recommended_items:
    print("Recommended items: ", recommended_items)
else:
    print("No items recommended.")

In this implementation example, item recommendations are made to the user by calling the recommend_items function. Here, the user’s preferences and recommended items are retrieved from the knowledge graph, and based on the preferences, items that have not yet been tried are recommended, if any. If the user does not exist or if there are no recommended items, an empty list is returned.

In this way, the recommendation system using the knowledge graph can recommend items based on the user’s preferences and selections, making it possible to express user profiles and item associations on the knowledge graph and utilize them to build a system that makes recommendations suitable for individual users. The system can also express user profiles and item associations on the knowledge graph, and use them to build a system that makes recommendations appropriate for individual users.

Social Network Analysis using Knowledge Graph

<Overview>

In social network analysis using knowledge graphs, the objective is to represent the relationships and connections among people and organizations in the form of knowledge graphs and to analyze them to gain insights. The following describes the methodology and specific applications of social network analysis using knowledge graphs.

Graph Construction: A knowledge graph is constructed to represent the relationships in a social network. Entities such as users, organizations, and communities are considered as nodes, and their relationships are represented by edges. Edges include friendships, follower relationships, mentions, etc.
Network Analysis: Analyze the constructed knowledge graph to reveal the characteristics and patterns of the social network. Network analysis methods (e.g., graph theory, social network analysis) will be used to analyze centrality, clustering, measures of influence, and diffusion of information.
Community Detection: Detect groups and communities in the knowledge graph. Identify and understand the structure of communities using techniques such as clustering of nodes with strong relationships and modularity optimization.
Influence Analysis: measure and identify influences within the knowledge graph. Identify influential nodes and organizations using methods such as centrality indices (e.g., order centrality, mediational centrality, proximity centrality) and page rank described in “Overview and Implementation of the Page Rank Algorithm“.
Modeling information diffusion: Predict the diffusion and influence of information on the knowledge graph. Here, information diffusion models (e.g., epidemic model, information cascade model) are applied to the knowledge graph to predict information diffusion patterns and influence.

Specific applications include the following

Social Media Analysis: The knowledge graph is used to analyze the connections among users and the diffusion patterns of information. This is used to identify influential users and analyze trends in topics.
Collaboration Analysis: Analyze patterns of collaboration and information flow within teams and organizations. These are used to visualize decision-making processes and promote effective information sharing.
Event Analysis: Analyzes the connections among event participants and information diffusion patterns to identify the success factors and influential participants of the event.
Marketing Analytics: Analyze connections and influences among customers of a brand or product to optimize marketing strategies and customer segmentation.

Social network analysis using knowledge graphs can be a powerful method for gaining a deep understanding and insight into the relationships between people and organizations.

<Example implementation in python>

This section describes a simple example implementation of social network analysis using knowledge graphs in Python. In this example, a social network analysis is performed by utilizing a knowledge graph that represents the relationships and connections among users. First, consider the following knowledge graph.

knowledge_graph = {
    "UserA": ["UserB", "UserC"],
    "UserB": ["UserA", "UserD"],
    "UserC": ["UserA"],
    "UserD": ["UserB", "UserE"],
    "UserE": ["UserD"],
    # Add other users and their relationships
}

Next, create functions to analyze user connections and relationships.

def analyze_social_network(user, knowledge_graph):
    if user in knowledge_graph:
        connections = knowledge_graph[user]
        print(f"{user}connection: {', '.join(connections)}")
        print(f"{user}number of connection: {len(connections)}")
    else:
        print(f"{user} does not exist on the knowledge graph.")

# Analyze users' social networks
user = input("Please enter your user name: ")
analyze_social_network(user, knowledge_graph)

In this example implementation, the social network of the specified user is analyzed by calling the analyze_social_network function. It retrieves the related users of the specified user from the knowledge graph, displays them, and also displays the number of related users. If the user does not exist, a message is displayed.

Thus, social network analysis using the knowledge graph represents the relationships and connections among users and utilizes them to build a mechanism for analyzing the characteristics of users’ social networks and the number of connections.

Life Science with Knowledge Graph

<Overview>

The field of life science using knowledge graphs integrates knowledge and data from a variety of sources and utilizes it for research and applications in the life science and medical fields. Below are some specific examples of life science applications using the knowledge graph.

Bioinformatics: Construct a knowledge graph to integrate biological data and related information such as genes, proteins, metabolites, etc. Data exploration and association analysis on the knowledge graph can be used to identify new biomarkers and disease-related genes, and to predict drug targets.
Drug development: Integrate information on drugs, targets, diseases, side effects, and clinical trial data on the knowledge graph to predict and evaluate drug efficacy and safety. Drug-target-disease relationships are analyzed to support the search for effective therapeutic strategies and the prediction of side effects.
Clinical Decision Support: Integrate medical data, treatment guidelines, and literature information on the Knowledge Graph to support clinicians and healthcare professionals in decision making. This will support accurate diagnosis and treatment planning by leveraging relevant information and evidence in the Knowledge Graph.
Personalized Medicine: Integrate patients’ genetic information, clinical data, lifestyle habits, etc. on the Knowledge Graph to provide the best individualized treatment and prognosis prediction. By analyzing the knowledge graph, the association between a patient’s genetic mutation and treatment response and the risk of side effects can be evaluated to develop a personalized medical strategy.
Data integration and knowledge discovery: Integrate diverse data sources related to life science (e.g., gene expression data, protein interaction data, literature data) to support the discovery of new knowledge and hypotheses. The construction of knowledge graphs and the analysis of associations will enable a deeper understanding of new biological networks and disease mechanisms.

The knowledge graph will enable the effective integration of vast amounts of life science knowledge and data to facilitate insights and innovative discoveries in the fields of research and medicine.

<About the example implementation in python>

While it is difficult to provide specific examples of knowledge graph implementations in the life science field, we describe below some common methods and tools for utilizing knowledge graphs in the life science field.

In the life science field, organization, analysis, and visualization of biological information and related data is important. Knowledge graphs can be a useful method for representing the relationships among molecules, genes, proteins, diseases, drugs, etc.

Python has a variety of packages and tools for utilizing knowledge graphs in the life sciences.

NetworkX: NetworkX is a powerful library for network analysis in Python. In the life science field, it is used to represent molecular interaction networks and gene-related networks.
BioPython: BioPython is a Python package that supports biological data analysis. It is used to process and analyze bioinformatics data such as DNA, RNA, and proteins.
PyBEL: PyBEL will be a Python library for expressing biomolecule associations. These will use the Biological Expression Language (BEL) format to express gene and protein relationships.

Using these tools and packages, it will be possible to build knowledge graphs in the life science field to analyze and visualize relationships and patterns of biological information.

Knowledge management using knowledge graphs

<Overview>

Knowledge management using knowledge graphs is an approach that utilizes knowledge graph concepts and techniques to integrate and facilitate access to knowledge within an organization. Some specific examples of knowledge management using knowledge graphs are described below.

Knowledge Integration and Sharing: Integrate knowledge from various data sources and information sources within an organization into a knowledge graph. This allows members of the organization to easily share information and learn from each other by expressing the relevance of knowledge on the knowledge graph.
Expert Identification and Collaboration: Use the knowledge graph to identify experts and expertise within the organization and facilitate collaboration with them. These are used to visualize the skills and relevance of experts on the knowledge graph to support problem solving and decision making.
Problem Solving and Innovation: Information and relevance on the knowledge graph will be leveraged to drive problem solving and innovation. This enables users to explore knowledge and present relevant information on the knowledge graph to support the discovery of new ideas and solutions.
Knowledge Map Creation: Create a knowledge map of the organization using the knowledge graph. This enables visualization of the organization’s knowledge domains, areas of expertise, and the location of experts to assist in the understanding and management of knowledge.
Knowledge storage and reuse: Store the knowledge accumulated on the knowledge graph and make it accessible in a reusable format. Combining information in the knowledge graph with search and recommendation systems makes it easier for members of the organization to access the knowledge they need.

Knowledge management using the Knowledge Graph is an approach that supports the promotion of innovation and decision-making by streamlining the use and sharing of knowledge within an organization. These are important methods that contribute to improving the competitiveness and efficiency of organizations.

<About the example implementation in python>

Below is a simple example of knowledge management implementation using knowledge graphs in Python.

class KnowledgeGraph:
    def __init__(self):
        self.graph = {}

    def add_knowledge(self, topic, information):
        if topic not in self.graph:
            self.graph[topic] = []
        self.graph[topic].append(information)

    def get_knowledge(self, topic):
        if topic in self.graph:
            return self.graph[topic]
        else:
            return []

# Create a knowledge graph
knowledge_graph = KnowledgeGraph()

# Add Knowledge
knowledge_graph.add_knowledge("Programming", "Python is a popular programming language.")
knowledge_graph.add_knowledge("Programming", "Java is an object-oriented language.")
knowledge_graph.add_knowledge("Data Science", "Machine learning is part of data science.")

# Get knowledge on a specific topic
topic = "programming"
knowledge = knowledge_graph.get_knowledge(topic)
if knowledge:
    print(f"{topic}knowledge:")
    for info in knowledge:
        print(f"- {info}")
else:
    print(f"There is no knowledge of {topic}.")

In this example implementation, the KnowledgeGraph class is defined: the add_knowledge method adds knowledge (information) related to the specified topic to the knowledge graph, and the get_knowledge method retrieves knowledge for the specified topic. In the execution example, an instance of the KnowledgeGraph class is created and some knowledge is added. It then retrieves and displays the knowledge of the specified topic.

This example is a very simple implementation of a knowledge management system; more complex functions and database integration can be added. It is also common for actual knowledge management systems to implement knowledge classification, search, and sharing functions.

Using the Knowledge Graph in Machine Learning

<Overview>

Machine learning can be useful in many ways in the utilization of the knowledge graph. Below we discuss some examples of how the knowledge graph can be used in machine learning.

Inference and inference-based tasks: The knowledge graph contains related facts and concepts connected at the edges. Machine learning models can leverage the structure and data in the knowledge graph to perform inference and question answering tasks. For example, it can learn the relationships among entities in the graph and make inferences based on new queries.
Node classification and attribute prediction: Nodes in the knowledge graph have various attributes and features associated with them. Machine learning models can learn the attributes and features of nodes and classify new nodes or predict their attributes. This could include, for example, the use of the knowledge graph for medical data to classify diseases or to predict new patient characteristics.
Extending the Knowledge Graph: The knowledge graph may also be extended by machine learning models. Models can learn new data and relationships and add new nodes and edges to the knowledge graph, thereby making the knowledge graph more informative and enabling more inferences and predictions.
Improved graph querying and visualization: Machine learning can also help improve query processing and graph visualization on the knowledge graph. This could include, for example, using machine learning to find the best path based on user queries, or learning the structure and features of a graph to generate more insightful visualizations.

Machine learning will allow users to more effectively utilize information and gain insights from the knowledge graph. However, since specific applications depend on data and tasks, it is necessary to select the most appropriate methods and algorithms for specific situations.

<Example implementation in python>

As a concrete implementation example for utilizing the knowledge graph for machine learning, we describe a method for extracting features from the knowledge graph and using them to construct a machine learning model. Below is an example implementation of extracting features from the knowledge graph and using them for a classification task.

import networkx as nx
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score

# Create a graph showing the relevance of users and interest categories as an example of a knowledge graph
knowledge_graph = nx.Graph()
knowledge_graph.add_edge("userA", "category1")
knowledge_graph.add_edge("userA", "category2")
knowledge_graph.add_edge("userB", "category2")
knowledge_graph.add_edge("userB", "category3")
knowledge_graph.add_edge("userC", "category1")

# Extract features from the knowledge graph
users = list(knowledge_graph.nodes)
categories = list(set(knowledge_graph.nodes) - set(users))
features = []
labels = []
for user in users:
    user_features = [int(knowledge_graph.has_edge(user, category)) for category in categories]
    features.append(user_features)
    labels.append(user)

# Split data into training and test sets
X_train, X_test, y_train, y_test = train_test_split(features, labels, test_size=0.2, random_state=42)

# Vectorize features
vectorizer = TfidfVectorizer()
X_train_vectorized = vectorizer.fit_transform(X_train)
X_test_vectorized = vectorizer.transform(X_test)

# Training logistic regression models
model = LogisticRegression()
model.fit(X_train_vectorized, y_train)

# Forecasts are made on test data and evaluated for accuracy
y_pred = model.predict(X_test_vectorized)
accuracy = accuracy_score(y_test, y_pred)
print(f"Accuracy: {accuracy}")

In this example implementation, the networkx package is used to create a knowledge graph to represent relationships. Next, features are extracted from the knowledge graph. Here, we used 0 and 1 features that represent whether each user is related to each category. Using the extracted features and labels, the dataset is divided into a training set and a test set, and then the features are vectorized using TfidfVectorizer. Finally, the LogisticRegression model is trained and predictions are made on the test data. The accuracy of the prediction results is evaluated.

This example implementation shows the general flow of feature extraction from the knowledge graph and training and evaluation of the machine learning model. Depending on the specific data and task, the feature extraction method and machine learning model should be adjusted appropriately.

Reference Information and Reference Books

See “Artificial Intelligence Technology as a DX Case Study” for examples of Knowledge Graph applications. In addition, “Knowledge Information Processing Technology” “Ontology Technology” and “Semantic Web Technology” also describe various application examples and application technologies. Please refer to this page as well.

Reference book is “Building Knowledge Graphs“

“Knowledge Graphs and Big Data Processing“

“The Knowledge Graph Cookbook“

“Domain-Specific Knowledge Graph Construction“