**Author Introduction**

Hong Hailan specializes in leveraging large language models (LLMs) to automate the extraction of knowledge from unstructured medical records, construct knowledge graphs in traditional Chinese medicine, and uncover the experiences of renowned TCM practitioners.


# Background


Building knowledge graphs presents dual challenges: improving the efficiency of knowledge extraction and reducing human labor costs. To effectively address these issues, this paper builds on the schema layer and explores methods for automated knowledge extraction. The following sections detail the process of constructing a knowledge graph for the treatment of rhinitis by Master Gan Zuwang, a renowned traditional Chinese medicine practitioner.
![](https://nebula-website-cn.oss-cn-hangzhou.aliyuncs.com/nebula-website-5.0/share/%E9%BC%BB%E7%82%8E%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1/1.png)
The data for this study comes from clinical records of rhinitis patients, with a total of 1,081 detailed records selected for analysis. These valuable records contain various basic patient information, such as name, workplace, gender, age, diagnosis time, disease type, symptom details, diagnostic results, observations of tongue and pulse, treatment records, prescription content, and medication guidance.






# Knowledge Graph Construction Process Diagram





## Step 1: Prepare Sample Data

Utilize the 1,081 preprocessed medical records as sample data.


## Step 2: Define Extraction Targets

Based on the ontology-defined schema layer, establish extraction targets like diseases, related diseases, symptoms, examinations, tongue appearance, pulse appearance, and compatibility. Define extraction rules for each target type.




## Step 3: Automated Knowledge Extraction



Create a prompt template based on the defined extraction targets, using a "demonstration case + relationship list" format to guide large language models in automatically extracting medical record data and generating triples.



**The following is a specific Prompt example:**

Imagine you are a large language model designed to extract relationships in Traditional Chinese Medicine. Your task is to extract the relationships between symptoms and related diseases for rhinitis patients. Here's an example medical case:

"Rhinitis has been severe for three days, the condition naturally flares up, with excessive nasal discharge, difficulty blowing the nose. Headache. Nasal congestion. Nasal mucosal edema, with secretions. Thin white tongue coating, floating and tight pulse. Prescription: mulberry leaf 6g, peppermint 6g, platycodon root 6g, calamus 6g, spatholobus stem 10g, schizonepeta 6g, angelica dahurica 6g, xanthium fruit 10g, fagopyrum 10g, vitex 10g."

The output format is:[(“rhinitis”, “symptom”, “excessive nasal discharge”), (‘rhinitis’, ‘symptom’, ‘difficulty blowing the nose’), (‘rhinitis’, ‘symptom’, ‘headache’), (‘rhinitis’, ‘symptom’, ‘nasal congestion’), (‘rhinitis’, ‘examination’, ‘nasal mucosal edema’), (‘rhinitis’, ‘examination’, ‘secretions’), (‘rhinitis’, ‘tongue appearance’, ‘thin white tongue coating’), (‘rhinitis’, ‘pulse appearance’, ‘floating and tight pulse’), (‘rhinitis’, ‘combination’, ‘calamus 6g’), (‘rhinitis’, ‘combination’, ‘spatholobus stem 10g’), (‘rhinitis’, ‘combination’, ‘schizonepeta 6g’), (‘rhinitis’, ‘combination’, ‘angelica dahurica 6g’), (‘rhinitis’, ‘combination’, ‘xanthium fruit 10g’), (‘rhinitis’, ‘combination’, ‘fagopyrum 10g’), (‘rhinitis’, ‘combination’, ‘vitex 10g’)]. Follow this format to extract the following medical case in triplet format. The given sentence is: ‘***’. Given the relationship list: [‘symptom’, ‘related disease’, ‘combination’, ‘tongue appearance’, ‘examination’, ‘pulse appearance’], provide the relationships in the list. If they do not exist, output: none.


## Step 4: Clean the Extracted Results

Use regular expressions to clean the automatically generated triplet data by removing errors, redundancies, irrelevant, and incomplete triplets. This process ensures the accuracy and reliability of the data.


## Step 5: Store the Triplets

Store the final extracted triplet data in the NebulaGraph database.

This paper directly extracts diseases and Chinese herbs into corresponding triplets to ensure the reliability of the knowledge extraction process using large language models. During the writing process, these triplets are further decomposed to ensure the rationality of the knowledge graph. Additionally, the entities and attributes of Chinese herbs are jointly extracted during the knowledge extraction. For example, the dosage of Fagopyrum 10g in a rhinitis patient’s prescription is extracted as ('rhinitis', 'combination', 'fagopyrum 10g').

Before storing the triplets, this paper uses regular expressions to extract the entities and their corresponding attributes. For instance, ('rhinitis', 'combination', 'fagopyrum 10g') is decomposed into ('rhinitis', 'combination', 'Fagopyrum') and ('fagopyrum', 'dosage', '10g'), as demonstrated in the constructed knowledge graph.




# Using NebulaGraph to build a knowledge graph




**Why NebulaGraph**？NebulaGraph is a high-performance and user-friendly graph database that offers researchers a new way to explore and analyze data thanks to its outstanding features and broad applications. It comes with a graph query language called **nGQL**, which provides users with a rich set of graph operation functions. Whether for basic CRUD operations on nodes and edges or complex graph algorithms and path queries, nGQL handles them effortlessly. This **flexibility and robust functionality make NebulaGraph an excellent choice for handling complex network structures and maximizing the value of medical case data**.

**Specific steps:**

* Algorithm pseudocode

* Construct prompt templates, related cases, and relationship lists

* Write functions to connect to **NebulaGraph**

* for i := 0 to N do // traverse medical case data

* Generate questions by combining data and templates

* Pass the questions to the LLM to get answers

* Clean the answers and remove unnecessary triples

* Build the IDs for patients, examinations, symptoms, and prescriptions in the current case

* for j := 0 to M do // iterate through the triple list


```
if relation'=="combination”:       
      Determine whether the Chinese herb in the triplet contains a dosage unit attribute. If it does, extract the attribute and store the extracted result in the knowledge graph. 
Else: Directly write the extracted results into the knowledge graph.
end for
```





# Experiment



This study utilizes a dataset of 353 pre-prepared medical cases as the foundation for knowledge extraction experiments using large language models (LLMs). In the experiment, other LLMs are employed for data annotation, followed by manual verification and correction to ensure the accuracy and reliability of the annotated data. This process serves to evaluate the knowledge extraction performance of each model. The Bert-BiLSTM-CRF model is selected as the baseline for joint entity and relationship extraction, and the results of automatic extraction using the Bert-BiLSTM-CRF model are compared with those obtained from large language models, primarily GLM3, GLM4, and ChatGPT-4. The evaluation metrics include F1 score, precision, and recall, with the experimental results presented in the table.
![](https://nebula-website-cn.oss-cn-hangzhou.aliyuncs.com/nebula-website-5.0/share/%E9%BC%BB%E7%82%8E%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1/2.png)



# Knowledge Graph Display


This study leverages the constructed knowledge graph to analyze and mine medication patterns used by a master physician in treating rhinitis, such as identifying the top 10 drugs most frequently prescribed for severe sneezing symptoms. The following Cypher query is executed to achieve this:

```
MATCH (v:Patient)-[:HAS_SYMPTOM]->()-[:SPECIFIC_SYMPTOM]->(v1) 
WHERE id(v1) = 'severe sneezing' 
MATCH (v:Patient)-[:PRESCRIBED]->(v3)-[e2:COMPONENT_OF]->(v4) 
RETURN v4, count(v4) AS cnt 
ORDER BY cnt DESC 
LIMIT 10;
```



This query identifies patients exhibiting severe sneezing symptoms, retrieves their treatment prescriptions, and sorts the drugs by frequency of occurrence, returning the top 10 most frequently prescribed drugs. The top 10 drugs are: Cicada Skin, Windproof, Ink Plant, Licorice, Earthworm, Black Soybean Coat, Chebulic Myrobalan, White Atractylodes, Rubia Root, and Astragalus.
![](https://nebula-website-cn.oss-cn-hangzhou.aliyuncs.com/nebula-website-5.0/share/%E9%BC%BB%E7%82%8E%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1/3.png)


This paper introduces how to build a knowledge graph for the treatment of rhinitis with traditional Chinese medicine using NebulaGraph.

Rhinitis Knowledge Graph in Traditional Medicine Using NebulaGraph


Although vector and graph databases might seem similar at first glance, the differences between them are much bigger than you might initially think. Among other things, they differ in terms of data retrieval and analysis, data structure, queries, and performance. As you can gather from their names, graph databases use graphs for managing data, while vector databases rely on vectors.

When [choosing a database](https://www.nebula-graph.io/posts/how-to-choose-database) solution for your business, you should first consider your business needs, the characteristics of your data, and the company use cases. By finding the right system for your particular requirements, you can squeeze maximum value from every software dollar.

Before explaining the pros and cons of each system, as well as their ideal use cases, let’s first define these two terms.  


# What are graph databases?

The main characteristic of a graph database is that it uses a graphical structure to store company data. Everything within this database is categorized by using a combination of vertex and edges. While vertices represent large categories of data, edges are utilized to establish relationships between these bigger entities.

The best thing about graph databases is how they process relationships. With this software, you can easily handle larger sets of data and determine connections between nodes. These systems excel at finding paths, identifying various patterns, and traversing relationships, which makes them perfect for making complex business decisions.

While there are numerous use cases for graph databases, they’re especially popular for [recommendation systems](https://www.nebula-graph.io/posts/use-cases-of-graph-databases-in-real-time-recommendation), [social networks](https://www.nebula-graph.io/posts/social-networks-with-graph-database-1), business knowledge graphs, and [fraud detection](https://www.nebula-graph.io/posts/fraud-detection-using-knowledge-and-graph-database). One of the best examples of a [graph database is NebulaGraph](https://www.nebula-graph.io/), which is open-source software with unlimited scalability and minimal latency.


# What are vector databases?

In this particular case, all data is stored within vectors that have a specific number of dimensions. You can use them for just about any type of unstructured or structured data, including text, images, and audio, making them extremely versatile. Experts praise this software for its scalability and performance compared to traditional systems.

Vector databases are growing in popularity as of late, as they provide lots of benefits for AI-driven systems. These databases can quickly establish similarities between data and detect closest neighbors. As such, they are perfect for any type of browsing, recommendations, clustering, and anomaly detection.

Some of the best examples of vector databases are Milvus and Pinecone.

# Graph database pros and cons

Graph databases are renowned for their speed and flexibility. A company can utilize them in all sorts of ways, and as such, they're almost a must-have for any modern business.   

## Pros

* **Flexibility and quickness** – Graph databases provide a lot of flexibility for your day-to-day operations. You can integrate these systems with various sources, adapt them for new use cases, and ensure that all your data is in accordance with business goals
* **Improved training** – Many companies use graph databases to improve their training and onboarding. The software allows the creation of efficient knowledge bases that can serve as a company-wide information source
* **Enhanced protection** – In financial risk control use cases, graph databases are ideal for detecting fraud rings and other sophisticated scams as it can unlock the highly complex correlation in fraudulent transactions
* **Better decision-making** – Graph databases are ideal for making business decisions. This type of software can establish relationships between different data groups, which makes it easier to determine the right moves for your company
* **Better for complex queries** – One of the best things about graph databases is that they allow you to process intricate queries by establishing cycles and discovering shortest paths


## Cons

* **Issues with scalability** – Except for NebulaGraph, most graph databases have problems with scalability. Luckily, many software companies are slowly working on resolving this issue and providing users with increased value
* **Additional overhead** – Certain datasets won't benefit from this technology. Basically, in certain cases, you don't need to establish relationships between different data categories, making your graph database solution redundant
* **Steep learning curve** – This type of software requires knowledge of complex query languages such as Neo4j and Cypher, which require more time to master


# Vector database pros and cons

In many ways, vector and graph databases are polar opposites. Vector solutions are ideal for processing search queries and can easily be used for machine learning models.


## Pros

* **Content flexibility** – These databases can process any type of data, including text, images, and audio
* **Machine learning benefits** – The software can easily be integrated with ML models, making it ideal for modern business and [marketing challenges](https://www.algocentric.ca/blog/saas-marketing-challenges/) 
* **Improved similarity searches** – Vector databases can easily find data points that are close to each other within multi-dimensional space. As such, it is a perfect tool for determining differences and similarities between various data points
* **Improved automation** – Given that machine learning is vital for automation, vector databases can, by proxy,  automate your business processes
* **Enhanced scalability** – Vector systems allow you to scale the processes as much as you want without ever encountering issues with availability or speed


## Cons

* **Lower accuracy** – Due to the speed of certain types of data retrieval, there are cases where vector databases provide subpar accuracy
* **Issues with dimensionality** – As the dimensionality increases, you’ll notice a drop-off in search efficiency and data availability. Although the creators use certain techniques that would mitigate the problem, it’s still a notable issue for this type of software
* **High requirements** – Vector databases are notorious for having high storage and memory requirements. This is especially true if you’re handling large datasets or as you start scaling the business


# What are the key similarities between graph and vector databases?

While these two database types are extremely different, and we implement them in different use cases, they also have a handful of similarities. For example, both of them can manage complex data and graph structure with ease. They are based on proven mathematical principles, allowing them to establish relationships between numerous data points.

The two database types excel at tackling complex queries, ensuring that businesses have the best, most relevant insights at their disposal. Modern graph databases, specifically, are fantastic for interconnecting data and establishing their connections. Both can be used to store data as well as manage and process complex structures.

The great thing about these databases is that you can use them for all sorts of tasks. Through their complex data handling and implementation of large language models, companies can build recommendation algorithms, perform social media analysis, detect fraud, and index search engines.

Whether you use them to draw insights and understand data or you utilize them for managing relationships between data points, graphs and vectors will allow you to create a niche for your business. No matter the approach you're using, they can tackle various complex tasks, thus giving users a competitive edge.


# What are the key differences between graph and vector databases?


Despite certain similarities, the two databases are, ultimately, made to tackle unique procedures and provide specific results. The main difference between them is rooted in their data models. While knowledge graphs rely on nodes and edges to visually showcase relationships between entities, vector databases rely on a vector space model and a multi-dimensional space.

Basically, vectors are ideal when analyzing the proximity and relationship of one data set compared to others. By relying on algorithms and indexing techniques, vector databases can easily track and analyze all vectors with similar traits to a given query vector.

On the other hand, graph databases are much more focused on the relationships themselves, not necessarily comparing entities to each other. They utilize previously mentioned nodes and edges to create a net of information. Because of this characteristic, graph databases excel for applications where relationship analysis takes center stage.

There are also some major differences in terms of querying. Specifically, vector databases primarily use mathematical computations for their queries, for example, cosine similarity and dot product computations, to establish relationships. By comparison, graphs are reliant on patterns and traversal algorithms to go from one node to another.

Some of these differences are rather stark and are the main reason why you might prefer one solution to another, depending on a specific case. A vector database is a perfect solution when you need to perform calculations, while graph databases excel at establishing complex relationships.


# Combining vector database and graph database

Although combining the two types of software might be tricky, and some companies might even consider it gimmicky, there are lots of benefits to it. Besides giving you more versatility, the use of vector and graph databases can ensure better query options, richer data representation, improved recommendations, and a unified data system.

The use of combined solutions has become much more common in the last few years, with the development of natural language processing, generative AI, and real-time data. Certain databases are even introducing vector similarity search as a way of tackling LLM hallucinations and retrieving insights (we refer to this as retrieval augmented generation).

Whatever the case might be, here are a few benefits you can experience by combining a knowledge graph database with a vector database:
* By combining the two, you get access to enhanced query options. Using a vector database concurrently with a graph database will allow you to discover similarities leading to better insights and, thus, better decision-making
* Richer data representation is another major benefit. By using vectors, you get a better understanding of data points. On the other hand, graph databases can visually showcase data in space, pointing out their relationships
* By adding vectors to knowledge graphs, you can achieve much higher scaling. You can analyze various relationships between data while simultaneously gaining fantastic insights from them
* Due to all these previously mentioned characteristics, the combination of a graph and vector database is a perfect solution for building recommendation systems
* Allows you to manage structured and unstructured data without a hitch. Users can streamline their daily operations and utilize a simple infrastructure to achieve their goals

Whatever the case, you need to study all available solutions before committing to one of them or using them in conjunction. That way, you can maximize your software expenditures while leaving competitors behind.
 

# The main challenges of integrating the two databases

Based on everything mentioned so far, combining the two technologies would make a lot of sense, especially for specific, complex tasks. Alas, using them in unison can turn out to be much more calling than you initially envisioned.

The first thing we need to consider is the financial side of things. Most databases charge companies per month per user. So, implementing both systems concurrently will double your expenses. We also need to consider all the manpower and maintenance required to effectively use them in unison.

Graph databases, in particular, can be costly. As with any other modern software, they require constant updates, which can be slow at times. They require more capacity and memory compared to vector databases, which are much simpler. You will also have to create proper streams that will update these relationships as a well of increasing input.

Although these problems might sound a bit too much, there are a few tricks you can implement to overcome them. The best way to start is by using specialized tools that will allivate some of the problem. We also suggest that you analyze the cost efficiency of combining the two databases before you start a project.


Although vector and graph databases might seem similar at first glance, the differences between them are much bigger than you might initially think. Among other things, they differ in terms of data retrieval and analysis, data structure, queries, and performance.

Vector Database vs. Graph Database: What Is Better for Your Project?

Exploring NebulaGraph RAG Pipeline with the Philadelphia Phillies

7 Query Strategies for Navigating Knowledge Graphs With NebulaGraph and LlamaIndex

As AI applications, like ChatGPT, gain popularity globally, large language models (LLMs) have garnered significant attention from both enterprises and individual users across various industries. Graph databases, which structure and link information graphically, can enhance LLMs' understanding of the relationships between entities, thereby improving their expressive and reasoning capabilities. So, what kind of innovations might arise when LLMs and graph technology intersect in specific application areas?

Currently, NebulaGraph is collaborating with several globally-renowned LLM teams to integrate LLMs with graph databases. We are excited to reveal that NebulaGraph now supports Text2Cypher, a combined Graph + LLM technology that translates natural language into graph queries. Through a dialogue interface, users can effortlessly construct and query knowledge graphs using natural language. Additionally, we provide a turnkey enterprise-level service. For a firsthand experience with our next-generation Knowledge Graph application powered by NebulaGraph, please follow the link provided at the end of this article.

# How can LLMs make knowledge graphs smarter?

AS one of the most common applications in graph technology, knowledge graphs organize and represent knowledge in the form of a graph structure consisting of nodes and edges. [Knowledge graphs](https://www.nebula-graph.io/posts/fraud-detection-using-knowledge-and-graph-database) not only represent the relationships between entities (e.g., people, places, events, and concepts) but also contain the semantic meanings of these relationships. They vividly describe knowledge in the real world in a structured manner, enabling computers to understand and reason about this knowledge. Knowledge graphs are widely used in risk control, production and manufacturing, drug development, scientific research, and other fields.

Text2Cypher, as its name implies, converts natural language text into Cypher queries. One of the challenges that previously impeded the broader adoption of [graph databases](https://www.nebula-graph.io/) and knowledge graphs was the technical barrier associated with querying these databases. So, how did we manage this prior to the advent of large language models?

## Traditional Text2Cypher

Before the emergence of large language models, the most common application of knowledge graphs was to convert natural language text into Cypher queries. For example, the internal system of a KBQA (Knowledge Base Question Answering) system is essentially text2cypher.

When a question is submitted to the system, it initially conducts intent and entity recognition. Subsequently, it employs an NLP model or code to formulate a query statement based on the identified intent and entity. The system then queries the graph database and crafts a response based on the retrieved structure.

So, what are the challenges associated with this approach?

- Understanding intent from natural language: Which category of questions does the question correspond to?
- Identifying entities: What are the main entities involved in the question?
- Constructing a query statement based on intent and entities

This is not an easy development task. A truly implementable model or rule code must consider many boundary conditions.

## LLM-based Text2Cypher

In the era of large language models, the above-mentioned "smart" application scenarios that previously required specialized training or rule writing can now be completed using a general model and prompt engineering.

> Note: Prompt engineering refers to the use of natural language descriptions to enable generative models and language models to complete smart tasks.

NebulaGraph has undertaken extensive R&D validation to successfully integrate large language models with graph query languages. Now, when users submit a task, they can utilize natural language for schema acquisition, Cypher statement generation, prompts, and various other large language model (LLM) calls and result processing. **This offers a truly code-free, ready-to-use solution for knowledge graph construction and querying.**


# Easy Knowledge Graph Construction and Querying with NebulaGraph

With the integration of NebulaGraph and large language models, enterprise users can easily construct and query knowledge graphs. Not only can they complete high-accuracy searches and intuitive visualizations at a lower cost, but they also no longer need to invest a long time in learning how to obtain important insights from graphs.

Compared with the previous generation of knowledge graphs, the combination of NebulaGraph and large models enables enterprises to construct knowledge graphs based on massive data at a lower cost and with greater convenience. Additionally, users can directly use natural language for interactive questioning and querying, thereby reducing the threshold for enterprise use and lowering the learning curve for users.

# LLM + Knowledge Graph Demo

Click the [demo](https://www.nebula-graph.io/posts/build-knowledgegraph-via-natural-language-demo) for a firsthand experience of constructing and querying knowledge graphs using natural language. 
For more about NebulaGraph and to connect to the digital future with NebulaGraph, visit our [official website](https://www.nebula-graph.io/). 





NebulaGraph now supports Text2Cypher, a combined Graph + LLM technology that translates natural language into graph queries. 

Knowledge Graphs & LLMs: Integrating Large Language Models with NebulaGraph

# Financial Fraud Detection: One of the Best Practices of Knowledge Graph

Hello, I'm Qingxin Kong from Bangsun Technology. I'm glad that I have this opportunity to share the topic **Financial Fraud Detection: Best Practices of Knowledge Graph **with NebulaGraph community members.

### Background: Problems with traditional fraud detection solutions

Financial frauds remain a hot topic in recent years. There's a new trend emerging that fraudsters are organized and ganged up on. It is difficult to verify whether the customer's information provided is fake or not. Such inconsistent information makes it difficult to identify the real frauds and leads to a lot of problems. The bank may control the passing rate by raising the requirements to handle loan applications due to the uncertainty of the customer's information, which is effective but brings losses.

The 2022 government work report said that financial inclusion will be further expanded and the customer base will include rural customers, which means there will be "white accounts" having no credit reports. Traditional fraudulent behaviors include fake credit reports, commission agencies, and collusion. 

Some agencies focus on a few banks for credit card applications or loans. Those agencies know the bottom line and red line of the banks. As long as the bottom line is not touched, they can successfully apply for loans. Those agencies may even know how to make the loan amount higher.

Traditional fraud detection solutions rely heavily on expert rules, that is, risk control experts set a batch of rules through experience and then adjust the trigger conditions to find a business balance between the pass rate and rejection rate. It costs little to forge things like ID numbers and bank card flows. There are no good means of control for mass and high-volume frauds.

### Empowering graphs: Paths to intelligent fraud detection solutions
![](http://www-cdn.nebula-graph.io/nebula-website-5.0/images/blogs/financial-fraud-detection-one-of-the-best-practices-of-knowledge-graph/figure1.png)

Let's take a look at the development path of intelligent fraud detection solutions. 

The first step is to build expert rules, but it is difficult to cover all risks through expert experience. The second step is to build a [machine learning](https://www.nebula-graph.io/posts/fraud-detection-with-graph-machine-learning) model to prevent and control the overall risk and then use expert rules to help the model. Expert rules and machine learning models well complement each other. However, based on the current experience, we may face lots of problems when building the machine learning model. The first reason is insufficient data and the second is the limited growth of the sample data. Because there are probably be 6~7 ways to apply for loans or credit cards, manual inspection is required in the end, so the data growth is very low. There will be a case below to explain the problem.

Therefore, we need to use relational graphs to crack the problems of information inconsistency. Both group and agency frauds have a batch of application behaviors. We can use the visualized graphs to recognize fraud samples. In this way, we may capture dozens to hundreds of fraudulent applications. It is also a very effective way for sample replenishment. Graphs can also be used for consistency checks. The multiple sources of information prove the information inconsistency. Graphs are displayed from the dimension of relations which helps us to do secondary verification. For example, multiple applications sharing the same information like the same phone number and IP address are considered group fraud.

To detect these group frauds, we use graphs that help us find rules of relations. These rules are complementary to the original expert rules. On the other hand, we can extract graph information to optimize machine learning patterns which increases the pattern's ability to detect group frauds and label more suspicious objects. Therefore, the expert rules, machine learning models, and graphs construct an effective ecological closed loop and complement each other to detect frauds. 

In terms of the specific applications of relational graphs, I've listed the following applications:

**_Application Graph_**: An example is using graphs to detect fraudulent applicants for loans and credit cards. One characteristic of a group is that information is shared by the group members. From the visual queries of graph data, a lot of correlations will be found. The more common information includes identity card, cell phone number, device fingerprint, e-mail address, etc. We use this information to explore relationships among these data, and then visually analyze the specific fraudulent techniques.**_‍_**

**_Transaction Graph_**: For multi-level transfer relationships, we mainly focus on the final flow of funds and who are the beneficiaries. The recent gambling fraud issue is the main focus of concern for financial supervisors, which is also one of the main directions we focus on. We have many customer cases. Through the list of suspects provided by the supervision of our internal rule models, suspicious transactions can be detected. In this process, we use historical flow data to explore more layers of data starting with the collected accounts. We may also use the graph exploration feature or a graph platform to analyze existing data for incremental prevention and control. 

**_Enterprise Graph/Internal Control Graph_**: Nowadays, there are interconnections between two or more enterprise businesses. The impact of external risks on the enterprise is increasing. Using graphs can outline the whole picture of risks for an enterprise. Combined with the introduction of external risks, the impact of external risks on the enterprise can be predicted as early as possible. Internal control graphs are more often used for detecting ethical or operational risks, such as misuses or failure to do certain operations in accordance with rules and regulations. Moral risks occur more often in unusual financial transactions between employees and enterprises, such as illegal fundraising or misappropriation of funds. The prevention core is to use graphs to check the actual control accounts of employees. Because bank employees are well acquainted with the bank's business. Few employees take their own accounts for some crimes. The identification of these control accounts can be explored by the means of graphs. For example, the accounts of relatives. We have a case of a joint-stock bank. In this case, the target was located in the account of the employee's girlfriend, because from visual graphs we found that many funds actually flowed to the girlfriend's account. It is very difficult to detect a case like this by traditional means.

**_Money Laundering Graph_**: The money laundering problem has existed for a long time. There are many derivative money laundering ways in different periods. Popular ways include the underground money bank and the "order sneaking" platform. Criminals hide their money on the platform where order-sneaking customers' daily consumption transactions are normal, however, there is actually a fixed pattern on visual graphs to detect cases like this. So how creating business relations with graphs and making graph rules to perform real-time monitoring and detections have become an important task for Bangsun, taken as a graph technology platform, to empower business developers. We will talk about it later. 

**_Vehicle Insurance/Ops Graph_**: For the vehicle insurance/ops field, we combine the initialization network with the design of vertices and edges, then query for a subgraph where the data is relatively appropriate and related, and then combine the experience of business experts to perform some exploratory and analytical work.

### **Use case 1: Flow of funds**
![](http://www-cdn.nebula-graph.io/nebula-website-5.0/images/blogs/financial-fraud-detection-one-of-the-best-practices-of-knowledge-graph/figure2.png)

Here, some specific application scenarios are also briefly listed. The first one is the flow of funds after the loan is approved.

We can focus on business and consumer loans through graphs, and track whether the approved loans flow into the housing market or the stock market. In this process, we will use the penetration ability of graphs to calculate the amount and proportion of transferred funds. Thereafter, we can easily find risks and violations and provide a basis for investigation. The regulators currently follow the principle of "three approaches and one guideline" for internet credits, so we establish real-time risk control through graphs and focus on the fund flows.

Next, we can show the specific methods. For example, loans flow directly into the blacklisted accounts or their own accounts are transferred to an account not under their name, but the account may be a controlled account. The account is used for investment and financial management. These can be analyzed with graphs and the specific pattern of the funds because the fund flows will form a very obvious distribution or pattern on visual graphs.

This pattern is difficult to portray with the previous expert rules, because the expert rules can only find a one-dimensional relationship, up to two-dimensional relationships, while the graphs are more adept at multi-dimensional relationship detecting. In massive transaction structures, we can also find some patterned structures. For example, the above figure shows 4-5 vertices, which involve some centralized transfer in, decentralized transfer out, and chain transaction structures like tree branches and ant nests, which are all abnormal funding patterns. Through visual graphs and combination with business knowledge, in many cases, we can see that it is problematic at a glance. As for where the specific problem is, we can check it through some functions provided by the graph, such as the k-degree query or the discovery algorithm of important nodes.

### **Use case 2: Insurance frauds** 
![](http://www-cdn.nebula-graph.io/nebula-website-5.0/images/blogs/financial-fraud-detection-one-of-the-best-practices-of-knowledge-graph/figure3.png)

The second use case is frauds in insurance, such as vehicle insurance or health insurance. The primary vehicle insurance issues are collusive claim frauds. By monitoring and associating shared identity information, it is easier to find group members and excludes irrelevant data. As for health insurance, we generally focus more on the improper relationship between doctors and patients or providers. Once there is a large number of discrete values on a drug or disease, it represents fraud. This is the common way to detect individual fraud in the past. Now the group fraud problem is severe. Group fraudsters have a shift in thinking that dozens of accounts buy the same drug through individual payment accounts. There are also aggregation accounts and many other commonalities. Combined with graph analysis and algorithms, those group frauds can be detected.

### **Summary: Advantages of graph fraud detection**
![](http://www-cdn.nebula-graph.io/nebula-website-5.0/images/blogs/financial-fraud-detection-one-of-the-best-practices-of-knowledge-graph/figure4.png)

The advantages of graph fraud detection mainly include the following four aspects.

Association analysis: This means that people take the initiative to explore and discover suspicious features, and then use graphs to make some visual associations. Business developers take the initiative to use graph tools for visual exploration and then combine the features of different business areas, such as money laundering, credit card applications, and loan applications to find some suspicious points. The suspicious points are certainly different in different areas, but the business developers who often use graph tools have a certain sensitivity to such suspicious points.

Graph rules: This aspect is more for incremental monitoring. Unlike the previous expert rules, graph rules use the means of graph exploration associated with knowledge comparison to draw some conclusions. We conclude graph rules from the dimension of relationships, and then the system uses the rules to detect risks.

Pattern analysis: We recently called the same model, because many fraud patterns have a solidified pattern, such as cash-out from credit cards. After analyzing these solidified patterns, we convert the pattern in a graph query language into a graph library for a traversal query. Compared to the graph rules introduced above, it is a method for detecting existing data.

Community analysis: As mentioned above, insufficient data is lacking for modeling graph data, the community analysis is actually a very good feature to discover fraudulent groups by algorithms and combination with data labels.

The above four graph fraud detection advantages can help us detect frauds in real-time, on time, and afterward. They are both the main graph functions for anti-frauds and are considered our strengths. These advantages are currently applied by about dozens of our customers.

This is all we have for today's article! Please look forward to the next topic Introduction of Bangsun Knowledge Graph Platform and Practical Cases of State-owned Banks.

‍

A knowledge Graph specialist from Bungsun shares his best Practices of Knowledge graph in the Financial Fraud detection field.

Financial Fraud Detection: One of the Best Practices of Knowledge Graph

Fraud detection is important to any organization, be it a financial service or simply a web service. With the increase in technology and the ease of information being shared with other parties, traditional methods of fraud detection have been rendered obsolete. In this post, we will be looking into fraud detection using knowledge graph and how you could use graph databases like **NebulaGraph** to build a quality knowledge graph for fraud detection.


The main goal of fraud detection is to minimize losses and prevent further damage. In order to achieve this goal, we need to understand the nature of fraud and how it affects our business processes.

For instance, in case of credit card fraud, there are many cases where the criminals use stolen data to buy high-end products such as luxury cars, jewelry, and so on. As a result of this, banks have to pay heavy fines for these transactions and sometimes even lose their license due to their negligence in maintaining customers’ information securely.

In order to deal with this problem, banks use a variety of techniques such as machine learning, neural networks and artificial intelligence techniques to detect credit card frauds. However, these approaches have some limitations because these techniques rely on statistical models which can be easily fooled by hackers using synthetic data sets or through other methods known as adversarial attacks.

# Graph database for fraud detection: How to detect and visualize fraudulent activities using knowledge graph

Knowledge graph is a state of the art of fraud detection. The reason is that it contains a massive amount of data, and even if one piece of information is incorrect or missing, the system will still be able to identify the user as fraudulent.

Knowledge graphs are used by financial institutions to detect fake identities (e.g., people trying to open accounts with fake ID cards), credit card fraud (e.g., someone applying for credit with a stolen credit card), and money laundering (e.g., someone trying to move money from one account into another).

# What is a knowledge graph?

A knowledge graph is a database of facts and relations between different entities.

A knowledge graph can be used to represent the world, with objects being concepts or physical things, and their attributes, relationships and metadata. For example, a financial institution could have a knowledge graph containing information about its customers, loans, and employees. The institute might also have a separate knowledge graph containing information about its offices and locations.

Graph databases like NebulaGraph are a natural way of building a knowledge graph because they provide an efficient way of storing relations between entities. A fact can be represented as an entity and its relationship with another entity can be represented as an edge between them. This representation enables us to use graph algorithms on our knowledge graph to find answers to various questions such as if a user is a spam user or if a deal is likely fraudulent.

**Related:**  [What is a Graph Database and What are the Benefits of Graph Databases
](https://www.nebula-graph.io/posts/why-use-graph-databases)

NebulaGraph is an open-source [graph database](https://www.nebula-graph.io) and it is best for dealing with large volumes of graph data.

NebulaGraph is a distributed, fault-tolerant and transactional graph database that provides ACID transactions, high availability, horizontal scalability and sub-second latency. It's designed to be easy to deploy and operate, both as a standalone instance or as part of a larger cluster.

The steps below outline a typical graph approach to fraud detection.

**Related:** [How I cracked Chinese Wordle using knowledge graph](https://www.nebula-graph.io/posts/crack-chinese-wordle-using-knowledge-graph)

# Data modeling

Fraud detection is a difficult problem for many reasons. For example, it is necessary to distinguish between legitimate requests and fraudulent requests, but many requests are legitimate on the surface but actually fraudulent in nature.

To solve this problem, you need to create a graph of relationships between the various pieces of information available about individual users. The most important piece of information is an account ID. This ID must be associated with all available information: account numbers, user names, IP addresses, social media accounts, email addresses, identification numbers and so on.

Once you have created this graph of relationships, it's possible to use graph algorithms such as the Louvain algorithm and Depth First Search to identify patterns that indicate fraud. This process can be automated so that once a pattern has been identified in one case (e.g., an account number), it can be applied to other cases with similar characteristics.

Knowledge Graphs are very effective in detecting fraud. The relationships between different entities provide context to the data, which helps in detecting fraudulent activity. For example, if you have a knowledge graph of relationships between people and their social media accounts, then you can find out if two accounts belong to the same person or not.

In addition to using graph algorithms, it's also possible to use rule-based systems or human experts who manually inspect suspicious cases and determine whether they're fraudulent or not. This can be done using NebulaGraph's data visualization tool[ Nebula Explorer](https://www.nebula-graph.io/products/explorer), which allows you to manually explore your graph data on a canvas.

# Define suspicious patterns

The goal of fraud detection is to identify suspicious activity. One of the most important steps is to define what suspicious activity to look for.

By using a knowledge graph, you can build powerful rules that detect known fraudulent behavior. Here are some examples of what you might want to look for:

*   Common attributes (same email addresses, tax identification number, or phone number, for instance)
*   Multiple parties using the same account
*   Short paths between transactions (a rapid return of purchase with no support call or reason given, for example)
*   Transactions outside normal patterns (a large payment made at 3 AM on a Saturday morning)

# Run graph algorithms

Once you have defined suspicious patterns in your fraud detection system, you can run graph queries using various graph algorithms in the knowledge graph to help you identifies these behaviors.

Here are some of the graph algorithms that are best for fraud detection in knowledge graph. You can use Nebula Algorithm To use these algorithms in your knowledge graph, which is also open source.

## Louvain

The Louvain method is a simple, efficient and easy-to-implement method for identifying communities in large networks.

The method unveils hierarchies of communities and allows to zoom within communities to discover sub-communities, sub-sub-communities, etc. It is today one of the most widely used methods for detecting communities in large networks.

![](http://www-cdn.nebula-graph.io/nebula-website-5.0/images/blogs/fraud-detection-using-knowledge-and-graph-database/figure1.png)

The Louvain method has been shown capable of detecting communities even when they are not well separated from each other, which makes it useful when analyzing complex networks such as fraud networks where communities may overlap with each other.

## PageRank

The PageRank algorithm was introduced by Google founder Lary Page. It was first used to rank web pages in the Google search engine. Nowadays, it is more and more used in many different fields, for example, in uncovering influential individuals and identifying high-frequency paths in fraud detection.

Influential individuals and high-frequency paths are both important characteristics of fraudsters as they tend to be more influential than others in their network and also tend to be connected through multiple schemes.

## Cycle detection

Cycle detection is especially a good fit for anti-money laundering applications. One of the money laundering patterns is a circular money flow. Money is being transferred from one account to another and eventually lands back in the originator’s account.

Using the cycle detection algorithm in your knowledge graph you can easily find such loops. For example, if a transaction involves multiple users and accounts, the system can detect if there is a circular flow between them.

![](http://www-cdn.nebula-graph.io/nebula-website-5.0/images/blogs/fraud-detection-using-knowledge-and-graph-database/figure2.png)

Cycle detection does not only work for money laundering but also for other forms of fraud like credit card fraud or identity theft.

# Visualize your data

Data visualization is the process of presenting data in a visual form that helps to identify patterns, trends and relationships. The goal of data visualization is to turn raw numbers into meaningful information that can be understood by anyone.

[Graph visualization](https://www.nebula-graph.io/posts/graph-database-visualization) helps you quickly analyze large amounts of data and spot patterns that indicate fraudulent activity. It allows you to investigate specific transaction patterns, find anomalies in real-time, and visualize complex relationships between entities. With this visual approach, you can also take advantage of graph algorithms to identify fraudulent transactions more efficiently than humans can.

Graph visualization cuts the manual analyst review time in half, allowing them to stop fraudulent transactions sooner and reduce wait times for non-fraudulent customers.

[Nebula Explorer](https://nebula-graph.io/products/explorer) is a graph data exploration app built upon NebulaGraph database. Taking advantage of the latest technologies, it is easy and fast to load large-scale graph data, query the data with visualized graphs, and identify fraudulent activities in a visualized way.

![](http://www-cdn.nebula-graph.io/nebula-website-5.0/images/blogs/fraud-detection-using-knowledge-and-graph-database/figure3.png)

# Conclusion

The global economy is facing increasing risks due to fraud. The technological challenge of quickly recognizing and analyzing massive volumes of transaction data makes fraud detection a very complex task. To meet mounting expectations, financial institutions and banks must use innovative technologies for efficient, comprehensive and reliable fraud detection such as graph technology.

It is clear that graph technology offers many advantages when it comes to fraud detection. This is an area in which "the more the better" also applies, as the number of available variables increases and their interconnections get richer and deeper. Graph databases can handle this without loading any additional costs – that is a key advantage. To conclude: fraud detection has been not just made more efficient, but more manageable in general by the use of graph technology.

## About the author

Nick Russell is a freelancing technical writer with a special focus on graph technology, machine learning and cloud computing. 

## References

*   Internet Financial Fraud Detection Based on a Distributed Big Data Approach With Node2vec - [IEEE](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9363921)
*   Fraud Detection Using Graph Analytics -[ Md. Ekramul Hoque Shajib](https://medium.com/pathaoengineering/fraud-detection-using-graph-analytics-4b7386fb648d)
*   Financial Crime &amp; Fraud Detection Using Graph Computing: Application Considerations &amp; Outlook - [arvix](https://arxiv.org/pdf/2103.01854.pdf)

# How NebulaGraph Works
1. [NebulaGraph Architecture — A Bird’s Eye View](https://www.nebula-graph.io/posts/nebula-graph-architecture-overview)
2. [Benchmark: NebulaGraph vs Dgraph vs JanusGraph](https://www.nebula-graph.io/posts/benchmarking-mainstraim-graph-databases-dgraph-nebula-graph-janusgraph)
3. [Comparison: Neo4j vs NebulaGraph vs JanusGraph](https://www.nebula-graph.io/posts/performance-comparison-neo4j-janusgraph-nebula-graph)
4. [Deploy the Graph Database on Kubernetes](https://www.nebula-graph.io/posts/how-to-deploy-nebula-graph-in-kubernetes)



‍

Knowledge graph is a state of the art of fraud detection. The reason is that graph database contains a massive amount of data, and even if one piece of information is incorrect or missing, the system will still be able to identify the user as fraudulent.

Tag