A low-code data fusion platform and data fusion method based on AI

Abstract

The present invention proposes an AI-based low-code data fusion platform and data fusion method. The AI-based low-code data fusion method includes: identifying the data sources that need to be fused and collecting data information in real time to obtain target data information; preprocessing the target data information corresponding to each of the data sources, and assigning weight values ​​to each of the data sources; combining the weight values ​​corresponding to each of the data sources to fuse the target data information corresponding to multiple data sources, and obtaining the fused data information. The system includes modules corresponding to the steps of the method. By automatically identifying data sources and collecting data information in real time, the method significantly reduces the workload of manual configuration and monitoring, and speeds up the process of data fusion. In particular, for dynamically changing data environments, it can quickly adapt and continuously provide the latest data.

Description

Technical Field

[0001] The present invention relates to the field of data fusion technology, and in particular to an AI-based low-code data fusion platform and a data fusion method. Background Art

[0002] In the current information age, the value of data is becoming increasingly prominent, and data fusion technology, as a key technology for extracting, integrating and integrating useful information from multiple data sources, is essential for achieving data-driven intelligent decision-making and efficient management. However, traditional data fusion methods often rely on complex programming and manual configuration, which limits the popularization and application of data fusion technology. In addition, with the increase in the types of data sources and the surge in data volume, how to fuse data efficiently and in real time while ensuring data quality has become an urgent problem to be solved. Summary of the invention

[0003] The present invention provides an AI-based low-code data fusion platform and a data fusion method to solve the problems mentioned in the above background technology:

[0004] The present invention proposes an AI-based low-code data fusion method, which includes:

[0005] Identify the data sources that need to be integrated and collect data information in real time to obtain target data information;

[0006] Preprocessing the target data information corresponding to each of the data sources, and assigning a weight value to each of the data sources;

[0007] The target data information corresponding to the multiple data sources is fused in combination with the weight value corresponding to each of the data sources to obtain fused data information.

[0008] Furthermore, identify the data sources that need to be integrated and collect data information in real time, including:

[0009] Extract the list of data sources and the data types required to fuse the data;

[0010] Acquire a data source with data fusion requirements by identifying data information in the data source list;

[0011] According to the data type required by the fused data, data information is collected from the data source in real time to obtain data information that meets the data type required by the fused data;

[0012] The data information that meets the data type required by the fusion data is used as the target data information.

[0013] Further, the target data information corresponding to each of the data sources is preprocessed, and a weight value is assigned to each of the data sources, including:

[0014] Performing data preprocessing on the target data information corresponding to each of the data sources, wherein the data preprocessing includes noise reduction processing and invalid data filtering processing;

[0015] Real-time monitoring of each data source to obtain data collection operation information of its corresponding target data information; wherein the data collection information includes bandwidth utilization, data collection rate, data collection volume, data quality evaluation coefficient and data collection response delay;

[0016] Extracting bandwidth utilization, data acquisition rate and data acquisition amount contained in the data acquisition information, and obtaining a first weight adjustment coefficient using the bandwidth utilization, data acquisition rate and data acquisition amount;

[0017] Extracting the data collection response delay and the data quality evaluation coefficient contained in the data collection information, and using the data collection response delay and the data quality evaluation coefficient to obtain a second weight adjustment coefficient;

[0018] Obtaining a weight value corresponding to each of the data sources using the first weight adjustment coefficient and the second weight adjustment coefficient;

[0019] A weight value is allocated to each of the data sources according to the weight value corresponding to each of the data sources.

[0020] Furthermore, the method for obtaining the data quality evaluation coefficient includes:

[0021] Obtaining a data volume ratio between invalid data and valid data according to the data volumes of invalid data and valid data obtained during data preprocessing;

[0022] Extracting the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein the abnormal data includes data information corresponding to the requirements of the target data information and data information that meets the requirements of the target data information but is garbled;

[0023] The data quality evaluation coefficient corresponding to the target data information is obtained by using the data volume ratio between the invalid data and the valid data and the data volume corresponding to the duplicate data and the abnormal data contained in the invalid data;

[0024] Further, combining the weight value corresponding to each of the data sources to fuse the target data information corresponding to the multiple data sources to obtain the fused data information includes:

[0025] Extracting target data information corresponding to each of the data sources, and performing standardization processing on the target data information corresponding to each of the data sources to obtain the target data information after standardization processing;

[0026] Inputting the standardized target data information into a deep learning model for feature recognition to obtain feature values ​​corresponding to each standardized target data information that meet data fusion requirements;

[0027] The characteristic values ​​corresponding to each of the standardized target data information that meet the data fusion requirements are combined with the weight values ​​corresponding to each of the data sources to perform data fusion in a weighted average manner to obtain the fused data information.

[0028] The present invention proposes an AI-based low-code data fusion platform, which includes:

[0029] Target data acquisition module: identifies the data source that needs to be fused and collects data information in real time to obtain target data information;

[0030] A weight value allocation module: pre-processes the target data information corresponding to each of the data sources, and allocates a weight value to each of the data sources;

[0031] Data fusion module: combines the weight value corresponding to each of the data sources to fuse the target data information corresponding to multiple data sources to obtain fused data information.

[0032] Furthermore, the target data acquisition module includes:

[0033] Type extraction module: extracts the data source list and the data types required for fusion data;

[0034] Demand acquisition module: acquires data sources with data fusion requirements by identifying data information in the data source list;

[0035] Data collection module: collects data information from the data source in real time according to the data type required by the fusion data, and obtains data information that meets the data type required by the fusion data;

[0036] Target determination module: taking the data information that meets the data type required by the fusion data as target data information.

[0037] Furthermore, the weight value allocation module includes:

[0038] Preprocessing module: performing data preprocessing on the target data information corresponding to each of the data sources, wherein the data preprocessing includes noise reduction processing and invalid data filtering processing;

[0039] Real-time monitoring module: real-time monitoring of each data source to obtain data collection operation information of its corresponding target data information; wherein the data collection information includes bandwidth utilization, data collection rate, data collection volume, data quality evaluation coefficient and data collection response delay;

[0040] A first weight adjustment module: extracting bandwidth utilization, data acquisition rate and data acquisition amount contained in the data acquisition information, and obtaining a first weight adjustment coefficient using the bandwidth utilization, data acquisition rate and data acquisition amount;

[0041] A second weight adjustment coefficient module: extracting the data acquisition response delay and the data quality evaluation coefficient contained in the data acquisition information, and obtaining a second weight adjustment coefficient using the data acquisition response delay and the data quality evaluation coefficient;

[0042] A data source acquisition module: using the first weight adjustment coefficient and the second weight adjustment coefficient to acquire a weight value corresponding to each of the data sources;

[0043] Weight value allocation module: allocates a weight value to each of the data sources according to the weight value corresponding to each of the data sources.

[0044] Furthermore, the method for obtaining the data quality evaluation coefficient includes:

[0045] Data volume ratio acquisition module: acquires the data volume ratio between invalid data and valid data according to the data volume of invalid data and valid data acquired in the data preprocessing process;

[0046] Data volume extraction module: extracting the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein the abnormal data includes data information corresponding to the requirements of the target data information and data information that meets the requirements of the target data information but is garbled;

[0047] Data quality evaluation coefficient acquisition module: obtains the data quality evaluation coefficient corresponding to the target data information by using the data volume ratio between invalid data and valid data and the data volume corresponding to the duplicate data and abnormal data contained in the invalid data;

[0048] Furthermore, the data fusion module includes:

[0049] Target data extraction module: extracts target data information corresponding to each of the data sources, and performs standardization processing on the target data information corresponding to each of the data sources to obtain the target data information after standardization processing;

[0050] Model input module: input the standardized target data information into the deep learning model for feature recognition, and obtain the feature value corresponding to each standardized target data information that meets the data fusion requirements;

[0051] Weighted average module: Utilize the characteristic values ​​corresponding to each of the standardized target data information that meet the data fusion requirements and combine with the weight values ​​corresponding to each of the data sources to perform data fusion by weighted average, and obtain the fused data information.

[0052] Beneficial effects of the present invention: By automatically identifying data sources and collecting data information in real time, the method significantly reduces the workload of manual configuration and monitoring, and accelerates the process of data fusion. In particular, for dynamically changing data environments, it can quickly adapt to and continuously provide the latest data; noise reduction and invalid data filtering in the data preprocessing stage, as well as the application of data quality evaluation coefficients, ensure the accuracy and reliability of data before fusion. In particular, by considering the impact of duplicate data and abnormal data, the credibility of the final fusion result is improved; based on data acquisition operation information (such as bandwidth utilization, acquisition rate, etc.), the weight value of the data source is dynamically adjusted. This mechanism can automatically adjust its influence in the fusion process according to the real-time performance of the data source, ensuring high-quality and efficient data integration; using deep learning models for feature recognition, not only improves the accuracy of feature extraction, but also makes data fusion more intelligent. This method can capture the complex relationship between data and enhance the comprehensive analysis value of fused data; as a low-code solution, this method lowers the technical threshold, allowing non-professional developers to quickly build and adjust data fusion processes. At the same time, its design makes it easy to add new data sources or adjust the fusion logic to adapt to changes in future data integration needs; the final integrated data information, because it has undergone strict screening, preprocessing and intelligent fusion, provides a more comprehensive and accurate foundation for subsequent business analysis and decision support, helping enterprises or organizations make more accurate strategic decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a step diagram of the method of the present invention;

[0054] Figure 2 This is a system module diagram of the present invention. DETAILED DESCRIPTION

[0055] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0056] Embodiment 1,

[0057] In this embodiment, Figure 1As shown, a low-code data fusion method based on AI is characterized in that the low-code data fusion method based on AI includes:

[0058] Identify the data sources that need to be integrated and collect data information in real time to obtain target data information;

[0059] Preprocessing the target data information corresponding to each of the data sources, and assigning a weight value to each of the data sources;

[0060] The target data information corresponding to the multiple data sources is fused in combination with the weight value corresponding to each of the data sources to obtain fused data information.

[0061] The working principle of the above technical solution is as follows: First, the system will identify the data sources that need to be fused, and collect data information from these data sources in real time to obtain the target data information. This may involve extracting data from various data sources and monitoring them in real time to ensure the latest data information; for the target data information corresponding to each data source, the system will perform preprocessing operations, which may include data cleaning, denoising, standardization and other operations to ensure data quality. At the same time, for each data source, the system will assign weight values ​​to determine the importance and influence of different data sources in the fusion; after obtaining the weight values ​​corresponding to each data source, the system will combine these weight values ​​to perform data fusion on the target data information corresponding to multiple data sources. This may involve using weighted average or other algorithms to integrate information from different data sources to obtain fused data information.

[0062] The effects of the above technical solution are as follows: by utilizing AI technology, it is possible to automatically identify data sources that need to be fused, and collect data information in real time, making the data fusion process more efficient and timely; by preprocessing the target data information corresponding to each data source, including noise reduction and invalid data filtering, the quality and accuracy of the data can be improved, thereby improving the effect of data fusion; by assigning weight values ​​to each data source, data fusion can be performed according to the importance and influence of different data sources, so that the fused data is more in line with the actual situation; by combining the weight values ​​corresponding to each data source for data fusion, the information of each data source can be integrated, thereby enhancing the information integrity and comprehensiveness of the data; this method is a data fusion method based on low-code, which reduces the need for complex programming, allows more people to participate in the data fusion process, and improves ease of use and participation.

[0063] Embodiment 2,

[0064] In this embodiment, identifying the data source to be fused and collecting data information in real time includes:

[0065] Extract the list of data sources and the data types required to fuse the data;

[0066] Acquire a data source with data fusion requirements by identifying data information in the data source list; wherein the data source includes but is not limited to a database, an API, etc.;

[0067] According to the data type required by the fused data, data information is collected from the data source in real time to obtain data information that meets the data type required by the fused data;

[0068] The data information that meets the data type required by the fusion data is used as the target data information.

[0069] The working principle of the above technical solution is as follows: the system will extract a list of data sources that need to be fused and determine the data type required for the fused data. This includes extracting a list of data sources from various data storage media and analyzing and determining the data types that need to be fused, such as text data, digital data, time series data, etc. The system will identify the data information in the data source list to determine which data sources have data fusion requirements. This includes scanning and analyzing data sources such as databases and APIs to determine whether they contain data information that needs to be fused; once the data sources with data fusion requirements are determined, the system will collect data information from these data sources in real time. This includes obtaining data information that meets the data type required for fusion data in real time through API calls, database queries, etc.; finally, the system will use the data information obtained from the data source in real time that meets the data type required for fusion data as the target data information. This data information will be used in the subsequent data fusion process.

[0070] The effects of the above technical solution are: by identifying data sources with data fusion requirements and collecting data information from these data sources in real time, it is possible to ensure that the latest data is obtained, making the fused data real-time and accurate; collecting data information according to the data type required for the fused data can ensure that the acquired data information meets the requirements of the fused data, avoiding the complexity of subsequent data conversion and cleaning, and improving the quality and availability of the data; the system can automatically identify the data sources that need to be fused and the required data types, reducing the need for manual intervention, improving work efficiency, and reducing the possibility of errors; the data sources mentioned in the technical solution include but are not limited to databases, APIs, etc., which means that data information can be collected from different types of data sources, thereby obtaining more comprehensive and diversified data, improving the comprehensiveness and reliability of data fusion; through real-time data collection and automated data identification, the entire data fusion process becomes more efficient, saving time and labor costs, and improving the overall efficiency of data fusion.

[0071] Embodiment 3,

[0072] In this embodiment, the target data information corresponding to each data source is preprocessed, and a weight value is assigned to each data source, including:

[0073] Performing data preprocessing on the target data information corresponding to each of the data sources, wherein the data preprocessing includes noise reduction processing and invalid data filtering processing;

[0074] Real-time monitoring of each data source to obtain data collection operation information of its corresponding target data information; wherein the data collection information includes bandwidth utilization, data collection rate, data collection volume, data quality evaluation coefficient and data collection response delay;

[0075] Extract the bandwidth utilization, data acquisition rate and data acquisition amount contained in the data acquisition information, and use the bandwidth utilization, data acquisition rate and data acquisition amount to obtain a first weight adjustment coefficient; wherein the first weight adjustment coefficient is obtained by the following formula:

[0076]

[0077] Among them, X 01 represents the first weight adjustment coefficient; n represents the number of unit times contained in the data collection duration of the data source, and the unit time is 1s; P i represents the bandwidth utilization rate of the i-th unit time; λ Si and λ Ci They represent the data acquisition rate parameter and data acquisition amount parameter of the i-th unit time respectively; P 0 represents a preset bandwidth utilization threshold; wherein the data acquisition rate parameter and the data acquisition amount parameter are obtained by the following formula:

[0078]

[0079] Among them, S min and S max Respectively represent the minimum and maximum data collection rates that occur during the data collection duration of the data source; C min and C max They represent the minimum and maximum values ​​of the data collected during the data collection time of the data source; S cmin and S cmax Respectively represent the data collection rate corresponding to the minimum and maximum values ​​of the data collection amount; C i represents the amount of data collected per unit time; S i represents the data acquisition rate corresponding to the i-th unit time;

[0080] Extract the data acquisition response delay and the data quality evaluation coefficient contained in the data acquisition information, and use the data acquisition response delay and the data quality evaluation coefficient to obtain a second weight adjustment coefficient; wherein the second weight adjustment coefficient is obtained by the following formula:

[0081]

[0082] Among them, X 02 represents the second weight adjustment coefficient; K represents the data quality evaluation coefficient; T i represents the data acquisition response delay rate corresponding to the i-th unit time;

[0083] The weight value corresponding to each of the data sources is obtained by using the first weight adjustment coefficient and the second weight adjustment coefficient; wherein the weight value corresponding to each of the data sources is obtained by the following formula:

[0084]

[0085] Among them, w represents the weight value corresponding to the data source; w 0 Indicates the preset weight reference value; ε indicates a positive number close to 0, which is used to prevent the denominator from being 0;

[0086] A weight value is allocated to each of the data sources according to the weight value corresponding to each of the data sources.

[0087] The working principle of the above technical solution is as follows: for the target data information obtained by each data source, data preprocessing is performed, including noise reduction processing and invalid data filtering processing. These preprocessing operations can improve the quality and availability of data and provide a better data basis for subsequent data fusion and analysis; for each data source, the data acquisition operation information for obtaining the target data information is monitored in real time, including bandwidth utilization, data acquisition rate, data acquisition volume, data quality evaluation coefficient and data acquisition response delay. According to these data acquisition information, the bandwidth utilization, data acquisition rate and data acquisition volume are extracted, and the first weight adjustment coefficient is calculated; the data acquisition response delay and data quality evaluation coefficient are extracted, and the second weight adjustment coefficient is calculated. Then these adjustment coefficients are used to calculate the weight value of each data source; according to the calculated weight value, the weight value is assigned to each data source. The weight value reflects the importance of each data source in data fusion, and the data source can be prioritized according to actual needs or corresponding weighted processing can be performed during the data fusion process.

[0088] The effects of the above technical solutions are as follows: by preprocessing the target data information corresponding to each data source, including noise reduction processing and invalid data filtering processing, the data quality can be effectively improved, the impact of noise and invalid data can be reduced, and the accuracy and credibility of subsequent data processing and analysis can be improved; by real-time monitoring of the data acquisition operation information of each data source to obtain the target data information, including indicators such as bandwidth utilization, data acquisition rate, and data acquisition volume, problems and bottlenecks in the data acquisition process can be discovered in a timely manner, and adjustments and optimizations can be made to ensure the efficiency and stability of data acquisition; by extracting key indicators from the data acquisition information and using these indicators to calculate the weight adjustment coefficient, the importance and contribution of each data source in data fusion can be more objectively evaluated, and the allocation of weight values ​​can be optimized, so that important data sources and high-quality data sources can obtain higher weights and greater influence in the data fusion process; through the implementation of the above technical solutions, the efficiency and accuracy of data fusion can be improved, more reliable data support can be provided for decision-making, and enterprises and organizations can be helped to more accurately understand and grasp the business situation, thereby improving the efficiency and accuracy of decision-making and achieving better business results. The various formulas in the above technical solutions comprehensively consider multiple factors of data collection, can effectively monitor and adjust various indicators in the data collection process, and improve the accuracy, real-time and efficiency of data collection. The application of these formulas helps to optimize the performance of the data collection system and improve the effect and quality of data processing. At the same time, the first weight adjustment coefficient formula can accurately reflect the real-time situation of the data collection operation status by considering the changes in bandwidth utilization, data collection rate and data collection volume, combined with the preset bandwidth utilization threshold. According to the first weight adjustment coefficient, the weight value of each data source can be adjusted in time, so that resources can be allocated more effectively during the data collection process, and the efficiency and accuracy of data collection can be improved. The data collection rate parameter and data collection volume parameter calculation formula take into account the fluctuation range and trend of the data collection rate and data collection volume, which helps to identify abnormal situations and make adjustments. By calculating the rate parameter and the volume parameter, the changes in the data collection process can be more comprehensively evaluated, providing a reference basis for subsequent data processing and resource allocation. The second weight adjustment coefficient formula combines the data quality evaluation coefficient and the data collection response delay rate, which can comprehensively consider the data quality and data collection response speed, thereby more comprehensively evaluating the data collection effect. The weight value of the data source can be adjusted through the second weight adjustment coefficient, making the priority of data collection more reasonable and improving the real-time and stability of data collection. The weight value calculation formula combines the first weight adjustment coefficient and the second weight adjustment coefficient, and can dynamically calculate the weight value of each data source according to the real-time data collection situation and data quality evaluation situation. Through the weight value calculation formula, resources and weights can be reasonably allocated according to actual conditions, improving the overall efficiency and quality of data collection, and ensuring the smooth progress of the data collection process.

[0089] Embodiment 4,

[0090] In this embodiment, the method for obtaining the data quality evaluation coefficient includes:

[0091] Obtaining a data volume ratio between invalid data and valid data according to the data volumes of invalid data and valid data obtained during data preprocessing;

[0092] Extracting the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein the abnormal data includes data information corresponding to the requirements of the target data information and data information that meets the requirements of the target data information but is garbled;

[0093] The data quality evaluation coefficient corresponding to the target data information is obtained by using the data volume ratio between invalid data and valid data and the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein the data quality evaluation coefficient is obtained by the following formula:

[0094]

[0095] Wherein, m represents the total number of times data preprocessing is performed on the target data information corresponding to each of the data sources; C si represents the amount of abnormal data contained in the invalid data of the i-th data preprocessing; C fi represents the amount of duplicate data contained in the invalid data of the i-th data preprocessing; C wi represents the amount of invalid data in the i-th data preprocessing; C yi Indicates the amount of valid data in the i-th data preprocessing.

[0096] The working principle of the above technical solution is as follows: in the process of data preprocessing, the data volume of valid data and invalid data can be obtained by denoising and filtering invalid data. According to the data volume ratio between valid data and invalid data, the quality of the data is preliminarily evaluated; in the invalid data, there may be some duplicate data and abnormal data, which will affect the evaluation of data quality. By analyzing the invalid data, the duplicate data and abnormal data therein can be identified and their corresponding data volume can be calculated; according to the data volume ratio between valid data and invalid data and the data volume of duplicate data and abnormal data contained in the invalid data, the data quality evaluation coefficient corresponding to the target data information can be calculated. The smaller the coefficient, the higher the data quality.

[0097] The effect of the above technical solution is: by obtaining the data quality evaluation coefficient corresponding to the target data information based on the data volume ratio between valid data and invalid data and the data volume of duplicate data and abnormal data contained in the invalid data, the quality of the data can be evaluated more accurately. This helps organizations to objectively and scientifically evaluate data quality during data processing and analysis; the technical solution can help identify and quantify duplicate data and abnormal data in invalid data, including data information that meets the requirements of target data information but is garbled. This helps to timely discover and process abnormal data and improve the accuracy and reliability of data; the obtained data quality evaluation coefficient can be used as a basis for data quality improvement, which helps organizations to formulate targeted data quality improvement measures and improve the accuracy, completeness and consistency of data. The above data quality evaluation coefficient comprehensively considers multiple aspects of data quality, provides an effective and comprehensive evaluation method, and helps to more accurately evaluate the quality status of target data information. At the same time, the formula takes into account the ratio between invalid data and valid data, the number of abnormal data and the number of duplicate data, comprehensively reflects multiple aspects of data quality, and more comprehensively evaluates the reliability and accuracy of the data; taking the number of abnormal data as one of the evaluation indicators can effectively reflect the abnormal situation of the data and take it into consideration in the evaluation, thereby improving the processing ability of abnormal data and the stability of data quality; the formula takes into account the possible impact of duplicate data on data quality, and incorporates the number of duplicate data into the evaluation indicators, which helps to evaluate the effect of data deduplication processing and further improves the accuracy and credibility of the data; by normalizing the evaluation indicators, the evaluation results under different data preprocessing times are comparable, ensuring the fairness and accuracy of the evaluation, and making the evaluation results more operational and practical; the formula can be adjusted and expanded according to specific business needs, and is suitable for data of different types and sizes, with certain flexibility and applicability.

[0098] Embodiment 5,

[0099] In this embodiment, data fusion is performed on target data information corresponding to multiple data sources in combination with the weight value corresponding to each of the data sources to obtain fused data information, including:

[0100] Extracting target data information corresponding to each of the data sources, and performing standardization processing on the target data information corresponding to each of the data sources to obtain the target data information after standardization processing;

[0101] Inputting the standardized target data information into a deep learning model for feature recognition to obtain feature values ​​corresponding to each standardized target data information that meet data fusion requirements;

[0102] The characteristic values ​​corresponding to each of the standardized target data information that meet the data fusion requirements are combined with the weight values ​​corresponding to each of the data sources to perform data fusion in a weighted average manner to obtain the fused data information.

[0103] The working principle of the above technical solution is: extract the target data information from each data source and standardize the information to ensure the comparability and consistency of data from different data sources; input the standardized target data information into a deep learning model specifically used for feature recognition. The deep learning model can learn and extract the important features of the data, thereby providing more effective guidance information for subsequent data fusion; combine the weight value corresponding to each data source with the feature value obtained by the deep learning model that meets the data fusion requirements, and perform data fusion by weighted averaging. The weight value can be determined based on the reliability, importance and contribution of different data sources to the target data information, to ensure that the fused data information can fully reflect the impact of each data source.

[0104] The effects of the above technical solution are: by combining the weight values ​​corresponding to each data source and the target data information after standardization, as well as the feature values ​​extracted by the deep learning model that meet the data fusion requirements, data fusion is performed in a weighted average manner, which can more effectively utilize the information of each data source and improve the precision and accuracy of data fusion; using the deep learning model to perform feature recognition on the target data information after standardization, it is possible to better mine the hidden feature information in the data, improve the ability to understand and analyze the data, and thus provide stronger support for data fusion; combining the weight values ​​corresponding to each data source for data fusion can perform reasonable weighted processing according to the importance and contribution of different data sources, ensure that the fused data is more in line with the actual situation, and avoid certain data sources from having too much influence on the results; through data fusion, the limitations and deviations that may exist in a single data source can be eliminated, the comprehensiveness and reliability of the data can be improved, thereby improving the data quality and providing more reliable data support for subsequent decision-making.

[0105] Embodiment 6,

[0106] In this embodiment, Figure 2 As shown, an AI-based low-code data fusion platform, the AI-based low-code data fusion platform includes:

[0107] Target data acquisition module: identifies the data source that needs to be fused and collects data information in real time to obtain target data information;

[0108] A weight value allocation module: pre-processes the target data information corresponding to each of the data sources, and allocates a weight value to each of the data sources;

[0109] Data fusion module: combines the weight value corresponding to each of the data sources to fuse the target data information corresponding to multiple data sources to obtain fused data information.

[0110] The working principle of the above technical solution is as follows: First, the system will identify the data sources that need to be fused, and collect data information from these data sources in real time to obtain the target data information. This may involve extracting data from various data sources and monitoring them in real time to ensure the latest data information; for the target data information corresponding to each data source, the system will perform preprocessing operations, which may include data cleaning, denoising, standardization and other operations to ensure data quality. At the same time, for each data source, the system will assign weight values ​​to determine the importance and influence of different data sources in the fusion; after obtaining the weight values ​​corresponding to each data source, the system will combine these weight values ​​to perform data fusion on the target data information corresponding to multiple data sources. This may involve using weighted average or other algorithms to integrate information from different data sources to obtain fused data information.

[0111] The effects of the above technical solution are as follows: by utilizing AI technology, it is possible to automatically identify data sources that need to be fused, and collect data information in real time, making the data fusion process more efficient and timely; by preprocessing the target data information corresponding to each data source, including noise reduction and invalid data filtering, the quality and accuracy of the data can be improved, thereby improving the effect of data fusion; by assigning weight values ​​to each data source, data fusion can be performed according to the importance and influence of different data sources, so that the fused data is more in line with the actual situation; by combining the weight values ​​corresponding to each data source for data fusion, the information of each data source can be integrated, thereby enhancing the information integrity and comprehensiveness of the data; this method is a data fusion method based on low-code, which reduces the need for complex programming, allows more people to participate in the data fusion process, and improves ease of use and participation.

[0112] Embodiment 7,

[0113] In this embodiment, the target data acquisition module includes:

[0114] Type extraction module: extracts the data source list and the data types required for fusion data;

[0115] Demand acquisition module: acquires data sources with data fusion requirements by identifying data information in the data source list; wherein the data sources include but are not limited to databases, APIs, etc.;

[0116] Data collection module: collects data information from the data source in real time according to the data type required by the fusion data, and obtains data information that meets the data type required by the fusion data;

[0117] Target determination module: taking the data information that meets the data type required by the fusion data as target data information.

[0118] The working principle of the above technical solution is as follows: the system will extract a list of data sources that need to be fused and determine the data type required for the fused data. This includes extracting a list of data sources from various data storage media and analyzing and determining the data types that need to be fused, such as text data, digital data, time series data, etc. The system will identify the data information in the data source list to determine which data sources have data fusion requirements. This includes scanning and analyzing data sources such as databases and APIs to determine whether they contain data information that needs to be fused; once the data sources with data fusion requirements are determined, the system will collect data information from these data sources in real time. This includes obtaining data information that meets the data type required for fusion data in real time through API calls, database queries, etc.; finally, the system will use the data information obtained from the data source in real time that meets the data type required for fusion data as the target data information. This data information will be used in the subsequent data fusion process.

[0119] The effects of the above technical solution are: by identifying data sources with data fusion requirements and collecting data information from these data sources in real time, it is possible to ensure that the latest data is obtained, making the fused data real-time and accurate; collecting data information according to the data type required for the fused data can ensure that the acquired data information meets the requirements of the fused data, avoiding the complexity of subsequent data conversion and cleaning, and improving the quality and availability of the data; the system can automatically identify the data sources that need to be fused and the required data types, reducing the need for manual intervention, improving work efficiency, and reducing the possibility of errors; the data sources mentioned in the technical solution include but are not limited to databases, APIs, etc., which means that data information can be collected from different types of data sources, thereby obtaining more comprehensive and diversified data, improving the comprehensiveness and reliability of data fusion; through real-time data collection and automated data identification, the entire data fusion process becomes more efficient, saving time and labor costs, and improving the overall efficiency of data fusion.

[0120] Embodiment 8,

[0121] In this embodiment, the weight value allocation module includes:

[0122] Preprocessing module: performing data preprocessing on the target data information corresponding to each of the data sources, wherein the data preprocessing includes noise reduction processing and invalid data filtering processing;

[0123] Real-time monitoring module: real-time monitoring of each data source to obtain data collection operation information of its corresponding target data information; wherein the data collection information includes bandwidth utilization, data collection rate, data collection volume, data quality evaluation coefficient and data collection response delay;

[0124] The first weight adjustment module extracts the bandwidth utilization, data acquisition rate and data acquisition amount contained in the data acquisition information, and obtains the first weight adjustment coefficient by using the bandwidth utilization, data acquisition rate and data acquisition amount; wherein the first weight adjustment coefficient is obtained by the following formula:

[0125]

[0126] Among them, X 01 represents the first weight adjustment coefficient; n represents the number of unit times contained in the data collection duration of the data source, and the unit time is 1s; P i represents the bandwidth utilization rate of the i-th unit time; λ Si and λ Ci They represent the data acquisition rate parameter and data acquisition amount parameter of the i-th unit time respectively; P 0 represents a preset bandwidth utilization threshold; wherein the data acquisition rate parameter and the data acquisition amount parameter are obtained by the following formula:

[0127]

[0128] Among them, S min and S max Respectively represent the minimum and maximum data collection rates that occur during the data collection duration of the data source; C min and C max They represent the minimum and maximum values ​​of the data collected during the data collection time of the data source; S cmin and S cmax Respectively represent the data collection rate corresponding to the minimum and maximum values ​​of the data collection amount; C i represents the amount of data collected per unit time; S i represents the data acquisition rate corresponding to the i-th unit time;

[0129] The second weight adjustment coefficient module extracts the data acquisition response delay and the data quality evaluation coefficient contained in the data acquisition information, and uses the data acquisition response delay and the data quality evaluation coefficient to obtain the second weight adjustment coefficient; wherein the second weight adjustment coefficient is obtained by the following formula:

[0130]

[0131] Among them, X 02 represents the second weight adjustment coefficient; K represents the data quality evaluation coefficient; T i represents the data acquisition response delay rate corresponding to the i-th unit time;

[0132] Data source acquisition module: using the first weight adjustment coefficient and the second weight adjustment coefficient to obtain the weight value corresponding to each of the data sources; wherein the weight value corresponding to each of the data sources is obtained by the following formula:

[0133]

[0134] Among them, w represents the weight value corresponding to the data source; w 0 Indicates the preset weight reference value; ε indicates a positive number close to 0, which is used to prevent the denominator from being 0;

[0135] Weight value allocation module: allocates a weight value to each of the data sources according to the weight value corresponding to each of the data sources.

[0136] The working principle of the above technical solution is as follows: for the target data information obtained by each data source, data preprocessing is performed, including noise reduction processing and invalid data filtering processing. These preprocessing operations can improve the quality and availability of data and provide a better data basis for subsequent data fusion and analysis; for each data source, the data acquisition operation information for obtaining the target data information is monitored in real time, including bandwidth utilization, data acquisition rate, data acquisition volume, data quality evaluation coefficient and data acquisition response delay. According to these data acquisition information, the bandwidth utilization, data acquisition rate and data acquisition volume are extracted, and the first weight adjustment coefficient is calculated; the data acquisition response delay and data quality evaluation coefficient are extracted, and the second weight adjustment coefficient is calculated. Then these adjustment coefficients are used to calculate the weight value of each data source; according to the calculated weight value, the weight value is assigned to each data source. The weight value reflects the importance of each data source in data fusion, and the data source can be prioritized according to actual needs or corresponding weighted processing can be performed during the data fusion process.

[0137] The effects of the above technical solutions are as follows: by preprocessing the target data information corresponding to each data source, including noise reduction processing and invalid data filtering processing, the data quality can be effectively improved, the impact of noise and invalid data can be reduced, and the accuracy and credibility of subsequent data processing and analysis can be improved; by real-time monitoring of the data acquisition operation information of each data source to obtain the target data information, including indicators such as bandwidth utilization, data acquisition rate, and data acquisition volume, problems and bottlenecks in the data acquisition process can be discovered in a timely manner, and adjustments and optimizations can be made to ensure the efficiency and stability of data acquisition; by extracting key indicators from the data acquisition information and using these indicators to calculate the weight adjustment coefficient, the importance and contribution of each data source in data fusion can be more objectively evaluated, and the allocation of weight values ​​can be optimized, so that important data sources and high-quality data sources can obtain higher weights and greater influence in the data fusion process; through the implementation of the above technical solutions, the efficiency and accuracy of data fusion can be improved, more reliable data support can be provided for decision-making, and enterprises and organizations can be helped to more accurately understand and grasp the business situation, thereby improving the efficiency and accuracy of decision-making and achieving better business results. The various formulas in the above technical solutions comprehensively consider multiple factors of data collection, can effectively monitor and adjust various indicators in the data collection process, and improve the accuracy, real-time and efficiency of data collection. The application of these formulas helps to optimize the performance of the data collection system and improve the effect and quality of data processing. At the same time, the first weight adjustment coefficient formula can accurately reflect the real-time situation of the data collection operation status by considering the changes in bandwidth utilization, data collection rate and data collection volume, combined with the preset bandwidth utilization threshold. According to the first weight adjustment coefficient, the weight value of each data source can be adjusted in time, so that resources can be allocated more effectively during the data collection process, and the efficiency and accuracy of data collection can be improved. The data collection rate parameter and data collection volume parameter calculation formula take into account the fluctuation range and trend of the data collection rate and data collection volume, which helps to identify abnormal situations and make adjustments. By calculating the rate parameter and the volume parameter, the changes in the data collection process can be more comprehensively evaluated, providing a reference basis for subsequent data processing and resource allocation. The second weight adjustment coefficient formula combines the data quality evaluation coefficient and the data collection response delay rate, which can comprehensively consider the data quality and data collection response speed, thereby more comprehensively evaluating the data collection effect. The weight value of the data source can be adjusted through the second weight adjustment coefficient, making the priority of data collection more reasonable and improving the real-time and stability of data collection. The weight value calculation formula combines the first weight adjustment coefficient and the second weight adjustment coefficient, and can dynamically calculate the weight value of each data source according to the real-time data collection situation and data quality evaluation situation. Through the weight value calculation formula, resources and weights can be reasonably allocated according to actual conditions, improving the overall efficiency and quality of data collection, and ensuring the smooth progress of the data collection process.

[0138] Embodiment 9,

[0139] In this embodiment, the method for obtaining the data quality evaluation coefficient includes:

[0140] Data volume ratio acquisition module: acquires the data volume ratio between invalid data and valid data according to the data volume of invalid data and valid data acquired in the data preprocessing process;

[0141] Data volume extraction module: extracting the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein the abnormal data includes data information corresponding to the requirements of the target data information and data information that meets the requirements of the target data information but is garbled;

[0142] Data quality evaluation coefficient acquisition module: obtain the data quality evaluation coefficient corresponding to the target data information by using the data volume ratio between invalid data and valid data and the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein, the data quality evaluation coefficient is obtained by the following formula:

[0143]

[0144] Wherein, m represents the total number of times data preprocessing is performed on the target data information corresponding to each of the data sources; C si represents the amount of abnormal data contained in the invalid data of the i-th data preprocessing; C fi represents the amount of duplicate data contained in the invalid data of the i-th data preprocessing; C wi represents the amount of invalid data in the i-th data preprocessing; C yi Indicates the amount of valid data in the i-th data preprocessing.

[0145] The working principle of the above technical solution is as follows: in the process of data preprocessing, the data volume of valid data and invalid data can be obtained by denoising and filtering invalid data. According to the data volume ratio between valid data and invalid data, the quality of the data is preliminarily evaluated; in the invalid data, there may be some duplicate data and abnormal data, which will affect the evaluation of data quality. By analyzing the invalid data, the duplicate data and abnormal data therein can be identified and their corresponding data volume can be calculated; according to the data volume ratio between valid data and invalid data and the data volume of duplicate data and abnormal data contained in the invalid data, the data quality evaluation coefficient corresponding to the target data information can be calculated. The smaller the coefficient, the higher the data quality.

[0146] The effect of the above technical solution is: by obtaining the data quality evaluation coefficient corresponding to the target data information based on the data volume ratio between valid data and invalid data and the data volume of duplicate data and abnormal data contained in the invalid data, the quality of the data can be evaluated more accurately. This helps organizations to objectively and scientifically evaluate data quality during data processing and analysis; the technical solution can help identify and quantify duplicate data and abnormal data in invalid data, including data information that meets the requirements of target data information but is garbled. This helps to timely discover and process abnormal data and improve the accuracy and reliability of data; the obtained data quality evaluation coefficient can be used as a basis for data quality improvement, which helps organizations to formulate targeted data quality improvement measures and improve the accuracy, completeness and consistency of data. The above data quality evaluation coefficient comprehensively considers multiple aspects of data quality, provides an effective and comprehensive evaluation method, and helps to more accurately evaluate the quality status of target data information. At the same time, the formula takes into account the ratio between invalid data and valid data, the number of abnormal data and the number of duplicate data, comprehensively reflects multiple aspects of data quality, and more comprehensively evaluates the reliability and accuracy of the data; taking the number of abnormal data as one of the evaluation indicators can effectively reflect the abnormal situation of the data and take it into consideration in the evaluation, thereby improving the processing ability of abnormal data and the stability of data quality; the formula takes into account the possible impact of duplicate data on data quality, and incorporates the number of duplicate data into the evaluation indicators, which helps to evaluate the effect of data deduplication processing and further improves the accuracy and credibility of the data; by normalizing the evaluation indicators, the evaluation results under different data preprocessing times are comparable, ensuring the fairness and accuracy of the evaluation, and making the evaluation results more operational and practical; the formula can be adjusted and expanded according to specific business needs, and is suitable for data of different types and sizes, with certain flexibility and applicability.

[0147] Embodiment 10

[0148] In this embodiment, the data fusion module includes:

[0149] Target data extraction module: extracts target data information corresponding to each of the data sources, and performs standardization processing on the target data information corresponding to each of the data sources to obtain the target data information after standardization processing;

[0150] Model input module: input the standardized target data information into the deep learning model for feature recognition, and obtain the feature value corresponding to each standardized target data information that meets the data fusion requirements;

[0151] Weighted average module: Utilize the characteristic values ​​corresponding to each standardized target data information that meet the data fusion requirements and combine with the weight values ​​corresponding to each data source to perform data fusion by weighted average, and obtain the fused data information.

[0152] The working principle of the above technical solution is: extract the target data information from each data source and standardize the information to ensure the comparability and consistency of data from different data sources; input the standardized target data information into a deep learning model specifically used for feature recognition. The deep learning model can learn and extract the important features of the data, thereby providing more effective guidance information for subsequent data fusion; combine the weight value corresponding to each data source with the feature value obtained by the deep learning model that meets the data fusion requirements, and perform data fusion by weighted averaging. The weight value can be determined based on the reliability, importance and contribution of different data sources to the target data information, to ensure that the fused data information can fully reflect the impact of each data source.

[0153] The effects of the above technical solution are: by combining the weight values ​​corresponding to each data source and the target data information after standardization, as well as the feature values ​​extracted by the deep learning model that meet the data fusion requirements, data fusion is performed in a weighted average manner, which can more effectively utilize the information of each data source and improve the precision and accuracy of data fusion; using the deep learning model to perform feature recognition on the target data information after standardization, it is possible to better mine the hidden feature information in the data, improve the ability to understand and analyze the data, and thus provide stronger support for data fusion; combining the weight values ​​corresponding to each data source for data fusion can perform reasonable weighted processing according to the importance and contribution of different data sources, ensure that the fused data is more in line with the actual situation, and avoid certain data sources from having too much influence on the results; through data fusion, the limitations and deviations that may exist in a single data source can be eliminated, the comprehensiveness and reliability of the data can be improved, thereby improving the data quality and providing more reliable data support for subsequent decision-making.

[0154] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A low-code data fusion method based on AI, characterized in that: The AI-based low-code data fusion method includes: Identify the data sources that need to be integrated and collect data information in real time to obtain target data information; Preprocessing the target data information corresponding to each of the data sources, and assigning a weight value to each of the data sources; Combining the weight values ​​corresponding to each of the data sources, data fusion is performed on the target data information corresponding to the multiple data sources to obtain fused data information; The target data information corresponding to each data source is preprocessed, and a weight value is assigned to each data source, including: Performing data preprocessing on the target data information corresponding to each of the data sources, wherein the data preprocessing includes noise reduction processing and invalid data filtering processing; Real-time monitoring of each data source to obtain data collection operation information of its corresponding target data information; wherein the data collection information includes bandwidth utilization, data collection rate, data collection volume, data quality evaluation coefficient and data collection response delay; Extract the bandwidth utilization, data acquisition rate and data acquisition amount contained in the data acquisition information, and use the bandwidth utilization, data acquisition rate and data acquisition amount to obtain a first weight adjustment coefficient; wherein the first weight adjustment coefficient is obtained by the following formula: Among them, X 01 represents the first weight adjustment coefficient; n represents the number of unit times contained in the data collection duration of the data source, and the unit time is 1s; P i represents the bandwidth utilization rate of the i-th unit time; Si and λ Ci Respectively represent the data acquisition rate parameter and data acquisition amount parameter of the i-th unit time; P0 represents the preset bandwidth utilization threshold; wherein the data acquisition rate parameter and data acquisition amount parameter are obtained by the following formula: Among them, S min and S max Respectively represent the minimum and maximum data collection rates that occur during the data collection duration of the data source; C min and C max They represent the minimum and maximum values ​​of the data collected during the data collection time of the data source; S cmin and S cmax Respectively represent the data collection rate corresponding to the minimum and maximum values ​​of the data collection amount; C i represents the amount of data collected per unit time; S i represents the data acquisition rate corresponding to the i-th unit time; Extract the data acquisition response delay and the data quality evaluation coefficient contained in the data acquisition information, and use the data acquisition response delay and the data quality evaluation coefficient to obtain a second weight adjustment coefficient; wherein the second weight adjustment coefficient is obtained by the following formula: Among them, X 02 represents the second weight adjustment coefficient; K represents the data quality evaluation coefficient; T i represents the data acquisition response delay rate corresponding to the i-th unit time; The weight value corresponding to each of the data sources is obtained by using the first weight adjustment coefficient and the second weight adjustment coefficient; wherein the weight value corresponding to each of the data sources is obtained by the following formula: Among them, w represents the weight value corresponding to the data source; w0 represents the preset weight reference value; ε represents a positive number close to 0, which is used to prevent the denominator from being 0; A weight value is allocated to each of the data sources according to the weight value corresponding to each of the data sources.

2. According to the AI-based low-code data fusion method of claim 1, it is characterized in that: Identify data sources that need to be integrated and collect data information in real time, including: Extract the list of data sources and the data types required to fuse the data; Acquire a data source with data fusion requirements by identifying data information in the data source list; According to the data type required by the fused data, data information is collected from the data source in real time to obtain data information that meets the data type required by the fused data; The data information that meets the data type required by the fusion data is used as the target data information.

3. The low-code data fusion method based on AI according to claim 1 is characterized in that: The method for obtaining the data quality evaluation coefficient includes: Obtaining a data volume ratio between invalid data and valid data according to the data volumes of invalid data and valid data obtained during data preprocessing; Extracting the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein the abnormal data includes data information corresponding to the requirements of the target data information and data information that meets the requirements of the target data information but is garbled; The data quality evaluation coefficient corresponding to the target data information is obtained by using the data volume ratio between the invalid data and the valid data and the data volume corresponding to the duplicate data and abnormal data contained in the invalid data.

4. The low-code data fusion method based on AI according to claim 1 is characterized in that: Combining the weight values ​​corresponding to each of the data sources to fuse the target data information corresponding to the multiple data sources to obtain the fused data information includes: Extracting target data information corresponding to each of the data sources, and performing standardization processing on the target data information corresponding to each of the data sources to obtain the target data information after standardization processing; Inputting the standardized target data information into a deep learning model for feature recognition to obtain feature values ​​corresponding to each standardized target data information that meet data fusion requirements; The characteristic values ​​corresponding to each of the standardized target data information that meet the data fusion requirements are combined with the weight values ​​corresponding to each of the data sources to perform data fusion in a weighted average manner to obtain the fused data information.

5. A low-code data fusion platform based on AI, characterized in that: The AI-based low-code data fusion platform includes: Target data acquisition module: identifies the data source that needs to be fused and collects data information in real time to obtain target data information; A weight value allocation module: pre-processes the target data information corresponding to each of the data sources, and allocates a weight value to each of the data sources; Data fusion module: combining the weight value corresponding to each of the data sources to fuse the target data information corresponding to the multiple data sources, and obtaining the fused data information; Wherein, the weight value allocation module includes: Preprocessing module: performing data preprocessing on the target data information corresponding to each of the data sources, wherein the data preprocessing includes noise reduction processing and invalid data filtering processing; Real-time monitoring module: real-time monitoring of each data source to obtain data collection operation information of its corresponding target data information; wherein the data collection information includes bandwidth utilization, data collection rate, data collection volume, data quality evaluation coefficient and data collection response delay; The first weight adjustment module extracts the bandwidth utilization, data acquisition rate and data acquisition amount contained in the data acquisition information, and obtains the first weight adjustment coefficient by using the bandwidth utilization, data acquisition rate and data acquisition amount; wherein the first weight adjustment coefficient is obtained by the following formula: Among them, X 01 represents the first weight adjustment coefficient; n represents the number of unit times contained in the data collection duration of the data source, and the unit time is 1s; P i represents the bandwidth utilization rate of the i-th unit time; Si and λ Ci Respectively represent the data acquisition rate parameter and data acquisition amount parameter of the i-th unit time; P0 represents the preset bandwidth utilization threshold; wherein the data acquisition rate parameter and data acquisition amount parameter are obtained by the following formula: Among them, S min and S max Respectively represent the minimum and maximum data collection rates that occur during the data collection duration of the data source; C min and C max They represent the minimum and maximum values ​​of the data collected during the data collection time of the data source; S cmin and S cmax Respectively represent the data collection rate corresponding to the minimum and maximum values ​​of the data collection amount; C i represents the amount of data collected per unit time; S i represents the data acquisition rate corresponding to the i-th unit time; The second weight adjustment coefficient module extracts the data acquisition response delay and the data quality evaluation coefficient contained in the data acquisition information, and uses the data acquisition response delay and the data quality evaluation coefficient to obtain the second weight adjustment coefficient; wherein the second weight adjustment coefficient is obtained by the following formula: Among them, X 02 represents the second weight adjustment coefficient; K represents the data quality evaluation coefficient; T i represents the data acquisition response delay rate corresponding to the i-th unit time; Data source acquisition module: using the first weight adjustment coefficient and the second weight adjustment coefficient to obtain the weight value corresponding to each of the data sources; wherein the weight value corresponding to each of the data sources is obtained by the following formula: Among them, w represents the weight value corresponding to the data source; w0 represents the preset weight reference value; ε represents a positive number close to 0, which is used to prevent the denominator from being 0; Weight value allocation module: allocates a weight value to each of the data sources according to the weight value corresponding to each of the data sources.

6. The AI-based low-code data fusion platform according to claim 5 is characterized in that: The target data acquisition module comprises: Type extraction module: extracts the data source list and the data type required for fusion data; Demand acquisition module: acquires data sources with data fusion requirements by identifying data information in the data source list; Data collection module: collects data information from the data source in real time according to the data type required by the fusion data, and obtains data information that meets the data type required by the fusion data; Target determination module: taking the data information that meets the data type required by the fusion data as target data information.

7. The AI-based low-code data fusion platform according to claim 5 is characterized in that: The method for obtaining the data quality evaluation coefficient includes: Data volume ratio acquisition module: acquires the data volume ratio between invalid data and valid data according to the data volume of invalid data and valid data acquired in the data preprocessing process; Data volume extraction module: extracting the data volume corresponding to the duplicate data and abnormal data contained in the invalid data; wherein the abnormal data includes data information corresponding to the requirements of the target data information and data information that meets the requirements of the target data information but is garbled; Data quality evaluation coefficient acquisition module: obtains the data quality evaluation coefficient corresponding to the target data information by using the data volume ratio between invalid data and valid data and the data volume corresponding to the duplicate data and abnormal data contained in the invalid data.

8. The AI-based low-code data fusion platform according to claim 5 is characterized in that: The data fusion module comprises: Target data extraction module: extracts target data information corresponding to each of the data sources, and performs standardization processing on the target data information corresponding to each of the data sources to obtain the target data information after standardization processing; Model input module: input the standardized target data information into the deep learning model for feature recognition, and obtain the feature value corresponding to each standardized target data information that meets the data fusion requirements; Weighted average module: Utilize the characteristic values ​​corresponding to each standardized target data information that meet the data fusion requirements and combine with the weight values ​​corresponding to each data source to perform data fusion by weighted average, and obtain the fused data information.

Images