Construction method based on knowledge graph blood index analysis

Abstract

The invention provides a knowledge graph construction method based on multi-source data analysis, and belongs to the field of data processing. According to the system, input and structured processing of multi-dimensional data are innovatively combined, data cleaning and standardization technologies are adopted, and the data quality is ensured. Nodes and relations are defined through a Neo4j graph database, and a complex knowledge graph capable of revealing individual data and state changes of the individual data is constructed. The system supports automatic data input and verification, carries out accurate prediction of a state trend in combination with a machine learning algorithm, and predicts a change result of the sixth year by using data of the first five years. Through introduction of node relation management of a graph database, the system can not only predict data trend, but also automatically generate a detailed analysis report, and help a user to carry out individualized analysis and decision support. The efficient data processing and prediction process of the system simplifies user operation through a graphical interface, is suitable for multi-source data analysis and prediction, provides comprehensive analysis and management support, and improves the depth and precision of data processing.

Description

Technical Field

[0001] The present invention relates to the field of data processing and analysis, and specifically to a knowledge graph construction method suitable for multi-source data analysis. Background Art

[0002] With the continuous development of data analysis technology, some indicators have received widespread attention as important factors in measuring individual status. The changes in these indicators not only reflect the individual's status, but are also closely related to a variety of potential risk factors. Therefore, studying the relationship between these indicators and status has important research and application value. At present, this type of research is mainly carried out through two methods: statistical analysis and specific data detection.

[0003] Through statistical analysis methods, the correlation between specific states and results can be identified, thus providing a basis for further data management and decision-making. This method relies on traditional statistical techniques, such as linear regression algorithms. The analysis results depend on the choice of model and the integrity of data. Insufficient data quality may affect the reliability of the analysis results. In addition, statistical analysis methods have limitations in capturing the relationship between complex factors and the interaction of multiple factors, and the accuracy of individualized trend prediction is low.

[0004] Data detection methods help develop personalized management strategies by providing specific status information. However, this method relies on expensive equipment and complex experimental operations, which are costly and difficult to operate, especially in a wide range of applications. In addition, the interaction between different data indicators and the influence of multiple factors also limit the prediction accuracy of this method.

[0005] Knowledge graph technology combines big data with artificial intelligence, providing a way to integrate multi-source heterogeneous data. It can reveal complex relationships and underlying mechanisms, providing more comprehensive analysis and accurate prediction results. Knowledge graph-based methods can utilize existing machine learning techniques, such as linear regression, to address the limitations of traditional methods, and improve efficiency and prediction accuracy by semi-automating the processing of large amounts of data.

[0006] The present invention constructs a comprehensive knowledge graph, combines machine learning algorithms and multi-source data, including indicators of multiple parameters such as white blood cell count, red blood cell count, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin content, mean corpuscular hemoglobin concentration, corpuscular volume distribution width, corpuscular volume distribution width, platelet count, mean platelet volume, hematocrit, platelet distribution width, large platelet ratio, neutrophil percentage, lymphocyte percentage, monocyte percentage, eosinophil percentage, basophil percentage, neutrophil count, lymphocyte count, monocyte count, eosinophil count, basophil count, C-reactive protein and health characteristics, and analyzes their impact on individual status. After adopting these technical means, the system can realize the prediction of future status trends and provide scientific decision support for individual management. Summary of the invention

[0007] In order to overcome the problems of complex data processing and insufficient prediction accuracy in state prediction research in the prior art, the present invention proposes a multi-source data analysis and future trend prediction method based on knowledge graph.

[0008] This approach has the following advantages:

[0009] 1. Data input module: Through the graphical interface, users can select files to input individual parameters and status data, and the system will perform data verification and prompt success or failure. The verification function of the input module is implemented through programming to ensure that the input data conforms to the expected format and range.

[0010] 2. Data processing module: The system will use the Neo4j graph database to build test result graphs, state nodes, state change nodes, and prediction nodes based on the data entered by the user. This module uses existing Neo4j technology and programming to complete the structured processing of data and graph construction.

[0011] 3. Data output module: The system uses existing machine learning algorithms to predict parameters and store and display the prediction results. Specifically, a linear regression model is used to generate trend prediction results and displayed in the user interface.

[0012] The overall process of the system includes three modules: data input, data processing, and data output. The data input module completes the input of individual data and status data through a graphical user interface. After the system verifies the validity of the input, it proceeds to the next step of processing. The data processing module uses the Neo4j database to create relevant nodes and process the input data, and finally generates the prediction results. The data output module uses a linear regression algorithm to fit the parameter data of the first five years, predict the results of the sixth year, and store and display them, such as Figure 1 shown.

[0013] During data processing, the system first checks the validity of the input data, such as whether there are incomplete or erroneous values. If an input error is detected, the system prompts the user to re-enter the data. The system uses a linear regression model to predict the results of the sixth year based on the parameter data of the previous five years. This prediction function is implemented through programming, and the linear regression algorithm is used to generate the prediction results. The system fits the data of the previous five years, calculates the model parameters, and predicts future trends based on this.

[0014] In the data output stage, the system uses the Neo4j graph database to match the corresponding treatment plan based on the input individual data and prediction results. The matching process is implemented through existing database query technology and programming, and finally the recommended treatment plan is presented to the user, such as Figure 2 shown.

[0015] The detailed data processing steps include data validation, data cleaning, filling missing values ​​with interpolation, and excluding outliers. Then, the system uses machine learning models for training and evaluation. In order to improve the prediction accuracy, the system adjusts the model through model evaluation and configuration optimization. After the system performs preliminary verification and cleaning of the data, it uses interpolation to handle missing values, excludes outliers, and trains through machine learning models. Finally, the system predicts the parameter results for the sixth year and stores and displays them, such as Figure 3 shown.

[0016] The maintenance of the knowledge graph includes the definition of nodes and relationships, regular updates, data storage, and generation of analysis reports. All node definitions and data updates are implemented through Neo4j graph database technology. Data privacy and security are guaranteed by existing encryption methods to ensure the reliability of the system during data processing, such as Figure 4 shown.

[0017] The system's graphical user interface is simple and intuitive. After the user enters the status data, the system prompts the next step. When invalid input is made, the system pops up a prompt box through programming to help the user correct the input data, thereby improving the convenience of user operation. Figure 5 shown.

[0018] The parameter data and prediction results displayed by the system are generated through a linear regression model to ensure that the data forecast for the sixth year is interpretable and transparent. Figure 6 shown.

[0019] Figure 7The data relationship diagram constructed by the system through the Neo4j graph database is shown. In this diagram, the system shows the relationship between various nodes such as input nodes, test result nodes, status nodes, prediction nodes, etc. Through the graphical display, users can intuitively view the relationship between various types of data generated by the system and the visualization effect of the processing process. In this way, the system can not only provide prediction results, but also show the potential relationship between multi-dimensional data, thereby providing more comprehensive data support for research and analysis, such as Figure 7 shown.

[0020] Graph Database Construction Explained

[0021] The present invention realizes the management and association analysis of multi-source data through the Neo4j graph database and constructs a complete knowledge graph. The system creates nodes and relationships related to the data based on multiple input data items, thus providing users with a comprehensive and intuitive data visualization display method.

[0022] Specifically, the Neo4j graph database uses the concepts of nodes and relationships to create corresponding nodes for each input data item and connect these nodes to each other through relationships. These nodes include input nodes, test result nodes, state nodes, state change nodes, and prediction nodes. In the system, different types of data, such as state changes and historical data, are represented by different types of nodes, and the relationship between them is shown through relationships. For example, the test result node is connected to the relevant state node through a specific relationship, thus building a complex association network.

[0023] In the actual application of the system, after the user enters the data, the system will generate corresponding nodes for each data item and organize these nodes using relationships. In this way, the system can not only show the relationship between different data items, but also show their changing trends over time. Each node stores specific data attributes, and the relationship between nodes shows the influence and association between data. Through this knowledge graph structure, users can easily query and analyze the potential relationships between data.

[0024] This construction method based on graph database has great flexibility and scalability, and is particularly suitable for processing and analyzing multi-source heterogeneous data. Through the Neo4j database, the system can easily process large amounts of data and display the complex relationships between data on a visual interface, helping users better understand the structure and changes of data. In actual applications, users can not only see the current data status, but also predict future trends and generate analysis reports based on the correlation of historical data.

[0025] By building this knowledge graph, the system provides researchers with a powerful data management and analysis tool. Through graphical display and intuitive query functions, researchers can more easily explore potential connections between data, provide personalized analysis suggestions, and optimize data management solutions, such as Figure 7 shown.

[0026] Graph database structure design

[0027] In order to construct complex data associations and display the input, processing and output relationships in the system, the present invention adopts Neo4j graph database technology. Through the graph database, the system can establish the input data as different nodes and connect these nodes through relationships to form a structured knowledge graph.

[0028] Specifically, the Neo4j graph database can generate corresponding nodes for each set of input data, including test result nodes, status nodes, state change nodes, and prediction nodes. These nodes are connected through relationships, such as extending from input nodes to status nodes, and then to test result and prediction nodes. The system shows the complete process from data input to result prediction. After the user enters data through the graphical interface, the system automatically generates these nodes and connects them to form a knowledge graph, showing the relevance and evolution of the data.

[0029] In actual applications, the system generates various nodes based on the data input by the user, and visualizes the relationship between the nodes through the graph database. Users can clearly see the relationship between different data nodes through the interface and analyze the data change process. This graphical presentation is not only intuitive, but also can display complex data association networks, providing strong support for in-depth data analysis.

[0030] This construction method based on graph database greatly improves the data processing efficiency and visualization effect of the system, and users can easily view and understand the structure and changes of data. Through the Neo4j database, the system realizes the integration and analysis of multi-source data, providing intuitive graphical support for subsequent research and analysis, such as Figure 7 shown.

[0031] Data analysis algorithms and model formulas

[0032] In the data analysis process of the present invention, the system uses a linear regression model to generate future state trends. The specific formula is shown in Formula 2:

[0033] y=β0+β1·t (Formula 1)

[0034] in:

[0035] y is the result value of the analysis

[0036] β0 is the intercept

[0037] β1 is the regression coefficient (slope)

[0038] t is a time parameter (such as year)

[0039] Beneficial Effects

[0040] The main innovation and beneficial effects of the present invention are reflected in the integration and association analysis of complex data through knowledge graph technology. The system constructs a knowledge graph through a custom graph database, and structures data from multiple sources, such as individual parameters, status information, and potential influencing factors. The knowledge graph enables the system to reveal the complex relationships between these data, thereby enhancing the depth and accuracy of the prediction results.

[0041] Although the linear regression model is used in this invention to predict future trends, the introduction of the knowledge graph is a deep mining of data relationships, which can help the system identify the interactions between multiple factors, which is of great significance for accurate prediction and decision analysis. The linear regression model is used to process the basic prediction of historical data, while the knowledge graph provides a more comprehensive background and information association.

[0042] The advantage of this method is the automated data processing and prediction process, which is implemented through programming, reducing human intervention and improving work efficiency. The transparency and interpretability of the prediction model ensure the scientific nature of the results and enhance the user's trust in the results. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The present invention is implemented in an overall flow chart.

[0044] Figure 2 This is a data input flow chart for the present invention.

[0045] Figure 3 It is a data processing flow chart of the present invention.

[0046] Figure 4 It is a data output flow chart of the present invention.

[0047] Figure 5 It is a detailed flow chart of data processing of the present invention.

[0048] Figure 6 A flow chart for maintaining the knowledge graph of the present invention.

[0049] Figure 7 This is an example diagram of the graph database of the present invention. DETAILED DESCRIPTION

[0050] The present invention is described in detail below with reference to the specific embodiments shown in the accompanying drawings.

[0051] Figure 1 The overall flow chart of a multi-source data analysis and future trend prediction method based on knowledge graph proposed by the present invention is shown. The system includes a data input module, a data processing module and a data output module. The data input module inputs individual parameter data by selecting a data file and gives a success or error prompt. The data processing module creates a test result node, a state node, a state change node and a prediction node to process the input data. The data output module predicts future trends and stores and displays the results, such as Figure 1 shown.

[0052] Figure 2 It is a data input flow chart, which describes in detail how the user selects and enters individual parameter data through the graphical interface. The system will verify the validity of the data. If the input is successful, the system will generate a test result node and prompt that the operation is successful; if the input fails, the system will prompt the user to re-enter the data, such as Figure 2 shown.

[0053] Figure 3 It is a data processing flow chart, which includes the following specific steps: First, the user inputs the individual's status data and parameter data of the previous five years, such as test results, indicator values, etc. The system will check whether these parameters have changed significantly within the numerical range determined based on historical data analysis or statistical standards. For example, the system will detect whether certain key parameters, such as the values ​​of specific indicators, are outside the normal range to determine whether these parameters exceed the upper and lower limits calculated through historical data or relevant standards. If these parameters do not exceed the set range, the system will use a regression model based on the data of the previous five years to predict the results of the sixth year. After the prediction is completed, the system will store and display the results, such as Figure 3 shown.

[0054] Figure 4 It is a data output flow chart, which describes how the system obtains individual status data and prediction results, and then matches the corresponding processing plan through the knowledge graph database. After the matching is completed, the system will determine the final analysis plan and display it to the user, such as Figure 4 shown.

[0055] Figure 5 The detailed steps of data processing are shown. First, the system verifies and cleans the input data to ensure the integrity and accuracy of the data. Then, interpolation is used to fill in missing values ​​and exclude outliers to ensure the continuity and consistency of the data. Then, the system trains and evaluates the machine learning model through model evaluation and configuration optimization to improve the prediction accuracy of the model. Finally, the system predicts the results of the sixth year based on the model, and outputs and stores these prediction results, such as Figure 5 shown.

[0056] When predicting future data, the system uses a linear regression model to generate future trend forecasts. The specific formula is as follows, as shown in Formula 1. Among them, y is the predicted result value, β0 is the intercept, β1 is the regression coefficient, and t is the time parameter. Through historical data {(t1,y1), (t2,y2),…, (t n ,y n )}, the system calculates the regression coefficient and intercept to predict future data trends.

[0057] The system calculates the regression coefficient β1 and intercept β0 through historical data, and the calculation formula is as shown in Formula 2. These formulas are used to generate regression prediction results of future states based on historical data.

[0058] Figure 6 The maintenance process of the knowledge graph is described. First, the nodes and relationships in the knowledge graph are defined, including individual parameter data, status data, and processing solutions. Then, the system regularly updates and maintains the knowledge graph to ensure the timeliness and accuracy of the data. The system also stores the prediction results and generates analysis reports, presenting the prediction data and suggestions through a graphical interface. At the same time, the system also implements data privacy and security measures to ensure the security of the data. The system can be horizontally expanded to handle more data and requests, such as Figure 6 shown.

[0059] Figure 7 The knowledge graph constructed by the system through the Neo4j graph database is shown. The graph shows the complex associations between different data nodes, including input nodes, test result nodes, status nodes, prediction nodes, etc. This graphical display enables users to clearly view the data relationships and changes in the system, so as to better understand the interaction and impact of data, such as Figure 7 shown.

[0060] The present invention realizes structured management and display of input data based on Neo4j through a custom developed graph database structure. The system creates different types of nodes and relationships through the graph database to form a complex knowledge graph network, which can effectively integrate multi-source data and provide users with powerful visualization and analysis functions.

[0061] In the application of the present invention, the Neo4j graph database is used to construct and maintain data nodes, such as test result nodes, state nodes, state change nodes, and prediction nodes. The system generates corresponding nodes for each data according to multiple data items input by the user, and connects these nodes through relationships to form a knowledge graph. The graph can show the association between data and visually show the change process of various types of data. In this way, the system can help users identify complex relationships between data, thereby providing support for data analysis and decision-making.

[0062] Specifically, the system creates corresponding nodes through the Neo4j graph database based on the previous five years of data input by the user, including input nodes, test result nodes, and status nodes. In the graph database, the relationship between nodes reflects the evolution of the data. For example, the state change node is connected to the prediction node through a relationship, showing the connection from historical data to future predictions. The graph database can not only efficiently manage large amounts of data, but also help users conduct in-depth analysis through its powerful query function.

[0063] The system provides an intuitive graphical user interface through which users can easily input data, view prediction results and perform interactive operations. The GUI port is implemented through Java Swing, providing a simple and feature-rich user experience. Users can input years of parameter data through the interface, and the system automatically processes this data and generates nodes, while displaying the data processing results and analysis reports on the interface.

[0064] The multi-step operation process in the GUI simplifies the interaction between users and the system. After the user enters the data, the system displays the validity check of the data through the GUI and generates a knowledge graph based on the data entered by the user. The knowledge graph is visualized through the GUI interface, showing the relationship between data nodes. Users can view the attributes, relationships and prediction results of various nodes on the interface to help them better understand the structure and change trends of the data.

[0065] This combination of graph database and GUI not only improves the processing efficiency of the system, but also enhances the visualization and analysis capabilities of data. Through graphical presentation and interaction, the system provides users with more intuitive and easy-to-understand analysis tools, helping researchers and analysts to develop individualized analysis plans and optimize data monitoring and management processes.

[0066] It should be understood that although this specification is described based on implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each implementation method may also be appropriately combined and implemented according to the understanding of those skilled in the art.

[0067] The series of detailed descriptions listed above are only specific descriptions of feasible implementation modes of the present invention. They are not used to limit the protection scope of the present invention. All equivalent implementation modes or changes that do not deviate from the spirit of the invention should be included in the protection scope of the present invention.

Claims

1. A knowledge graph construction method based on data indicator analysis, Features The following steps are involved: Data entry phase: Select and input multiple individual data through the graphical interface; including multiple parameters such as white blood cell count, red blood cell count, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin content, mean corpuscular hemoglobin concentration, and red blood cell volume Distribution width, red blood cell volume distribution width, platelet count, mean platelet volume, platelet hematocrit, platelet distribution width, large platelet ratio, neutrophil percentage, lymphocyte Baocyte percentage, monocyte percentage, eosinophil percentage, basophil percentage, neutrophil count, lymphocyte count, monocyte count, eosinophil count, basophil count; Verify the validity of data; If the input is successful, the system will generate a test result graph and prompt success; if the input fails, the system will prompt an error and require you to re-enter the data. Data processing stage: Create test result graphs, state nodes, state change nodes, and prediction nodes based on multiple input data; Check whether the individual's status has undergone critical changes, including: whether the value of a specific indicator exceeds the preset upper and lower limits, or whether the parameter has fluctuated significantly, that is, exceeds a specific multiple of the standard deviation or mean; If there is no change in status, the system uses the data from the previous five years to build a forecast model and calculate the forecast results for the sixth year; Interpolation was used to fill missing values ​​and exclude outliers to ensure data continuity and consistency; Train and evaluate machine learning models, and improve the prediction accuracy of models through model evaluation and configuration optimization; Data output stage: Obtain individual status data and prediction results; Match the corresponding analysis scheme through the knowledge graph database; Determine the final analysis plan and present it to the user; The prediction results are stored and displayed on the interface.

2. The method according to claim 1, It is characterized in that The data processing phase includes the following steps: Verify and clean input data to ensure data integrity and accuracy; Use interpolation to fill missing values ​​and exclude outliers; A linear regression model was built using the data from the first five years to calculate the forecast results for the sixth year.

3. The method according to claim 1, It is characterized in that The data output phase includes the following steps: Obtain individual status data and prediction results; Match the corresponding analysis scheme through the knowledge graph database; Determine the final analysis plan and present it to the user; The prediction results are stored and displayed on the interface.

4. The method according to claim 1, It is characterized in that The knowledge graph maintenance process includes the following steps: Define nodes and relationships in the knowledge graph, including multiple data of individuals, status data, and analysis schemes; Regularly update and maintain the knowledge graph. ; Store forecast results and generate analysis reports, presenting forecast data and suggestions through a graphical interface; Implement data privacy and security measures.

5. The method according to claim 1, It is characterized in that The system builds a knowledge graph through the Neo4j graph database, showing the complex relationships between different data nodes, including input nodes, test result nodes, status nodes, and prediction nodes; the graph shows the evolution and correlation of data relationships, and provides users with interactive analysis in a graphical way.

6. The method according to claim 1, It is characterized in that By building a graph database based on the knowledge graph, integrating multiple individual data, and defining and creating multiple nodes and relationships, a knowledge graph network including input nodes, test result nodes, status nodes, status change nodes, and prediction nodes is formed.

7. The method according to claim 1, It is characterized in that By applying knowledge graphs to data trend analysis, detailed analysis reports can be generated and individualized analysis strategies can be formulated; through structured processing of data through knowledge graphs, data management and subsequent intervention measures can be optimized.

8. The method according to claim 1, It is characterized in that The knowledge graph realizes the complex association display between nodes through the construction of a graph database; the nodes of the graph database include input data nodes, test result nodes, and status nodes.

Images