A method for quickly searching city infrastructure ownership based on big data analysis

Abstract

The application discloses a kind of based on the method for quickly finding city infrastructure ownership of big data analysis, specifically related to data analysis technical field;Collect GIS, property registration database, internet of things sensor, remote sensing image and historical archives data, use natural language processing and optical character recognition technology to parse unstructured text, compare the spatial information of GIS database and property registration database, mark geographic space inconsistency data, and combine property unit change trend analysis, identify property correlation instability data, train machine learning model, build infrastructure ownership prediction rule, improve the credibility evaluation ability of ownership informationThe application significantly improves the accuracy and response speed of infrastructure ownership query, especially in the scene of emergency repair and accident disposal, the property unit can be quickly identified, the repair delay caused by information lag or wrong matching is avoided, and the city management efficiency and public safety guarantee ability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to a method for quickly searching for the ownership of urban infrastructure based on big data analysis. BACKGROUND

[0002] The method for quickly searching for the ownership of urban infrastructure based on big data analysis refers to using big data technology to efficiently search and match the ownership information of infrastructure such as roads, bridges, pipe networks, and public facilities in cities. This technology usually relies on GIS (Geographic Information System), the Internet of Things, remote sensing images, property registration databases, and other multi-source data. Through methods such as machine learning, data mining, and natural language processing, it automatically identifies, analyzes, and correlates ownership information to quickly determine the ownership unit, management department, or property subject of the infrastructure, thereby improving urban management efficiency, reducing property disputes, and accelerating infrastructure maintenance and planning decisions.

[0003] The prior art has the following deficiencies:

[0004] During the process of emergency repair of urban infrastructure or major accident disposal, repair personnel often need to quickly determine the property unit of the facility in order to contact the relevant responsible party for repair or emergency treatment. However, due to the problem of outdated updates in existing GIS systems and property databases, the repair team may not be able to identify the ownership of the infrastructure within a short period of time. For example, in the case of a gas pipeline leakage accident, if the ownership information in the GIS database is not updated in time, the repair personnel may contact the wrong gas company, resulting in the valve not being closed in time and causing an explosion accident. In addition, the existing property registration system mostly uses keyword matching for retrieval, but in the face of unit name changes, address adjustments, and difficulties in recognizing handwritten archives, the query system may fail to match or produce incorrect results, thereby affecting the efficiency of emergency treatment. Due to the large amount of data and the dependence on traditional SQL retrieval for query, the existing system responds slowly and cannot meet the real-time query requirements in emergency situations. SUMMARY

[0005] The purpose of the present application is to provide a method for quickly searching for the ownership of urban infrastructure based on big data analysis to solve the problems in the background art.

[0006] To achieve the above purpose, the present application provides the following technical solution: a method for quickly searching for the ownership of urban infrastructure based on big data analysis, comprising:

[0007] Collecting GIS, property registration database, Internet of Things sensor, remote sensing image, and historical archive data, using natural language processing and optical character recognition technology for analysis, and unifying geographic information format through coordinate conversion and projection standardization;

[0008] Marking out the geographic spatial inconsistency data by comparing the infrastructure spatial information in the GIS database with the geographic coordinates of the property registration database, analyzing the change trend of the key fields of the property unit name, contact information and record time, and marking the property correlation instability data if there are multiple changes within a fixed time window;

[0009] Combining the geographic spatial inconsistency data and the property correlation instability data, training an infrastructure ownership prediction model by using a machine learning model, establishing an ownership matching rule, and judging whether the ownership information is reliable;

[0010] After the user inputs the query condition, the system calls the matching model in real time, provides the ownership information query result, and marks the matching reliability. If the matching reliability of the query result is lower than a predetermined threshold, an artificial review process is triggered.

[0011] Preferably, the natural language processing technology includes named entity recognition, syntax analysis and keyword extraction, which is used to extract the property unit, address, facility name and change time information from the property registration text, historical archives and handwritten documents, and analyze the unstructured text by combining the optical character recognition technology.

[0012] Preferably, the coordinate conversion and projection standardization includes coordinate format conversion, projection transformation and error correction, which unifies the different coordinate systems in the GIS database and the property registration database to the same geographic reference, and corrects the error by using the nearest neighbor interpolation or affine transformation.

[0013] Preferably, the spatial deviation anomaly index is generated after analyzing the spatial deviation data change of the same infrastructure in the two databases, and the method for obtaining the spatial deviation anomaly index is as follows:

[0014] The coordinate information of the infrastructure is obtained from the two databases, and the spatial deviation ΔD of each facility at different times is calculated, and the expression is as follows: Wherein, X GIS , Y GIS are the coordinates in the GIS database, and X 登记 , Y 登记 are the coordinates in the property registration database; the exponentially weighted moving average value of the spatial deviation data is calculated, and the expression is as follows: S t = αΔD + (1-α)S t-1 ; wherein: S t represents the exponentially weighted moving average value of the spatial deviation data at the t time, α is a smoothing coefficient, and the value range is 0<α<1, S t-1 is the exponentially weighted moving average value of the spatial deviation data at the t-1 time, and the spatial deviation anomaly index is calculated as follows: Wherein: S min , S maxThe minimum and maximum exponential moving average values known in the historical data are used for normalization processing, and the DC is a spatial deviation anomaly index.

[0015] Preferably, the data change rate fluctuation index is generated after analyzing the data change rate fluctuation in the property association instability data, and the method for obtaining the data change rate fluctuation index is as follows:

[0016] The historical change data of the field is collected, the change rate of the property field in a fixed time window is calculated, a feature vector is constructed, the data of each time window is converted into a format available for machine learning, including the number of changes, the change interval time and the change duration; a random forest model is trained, the past property change rate data is used as input, the random forest learns the normal change mode, the test data is used to evaluate the prediction error of the model, the current time window change data is input, the random forest predicts the normal change rate range, and the difference between the actual change rate and the predicted change rate is calculated as the data change rate fluctuation index.

[0017] Preferably, the spatial deviation anomaly index and the data change rate fluctuation index are converted into a comprehensive feature vector, the comprehensive feature vector is used as the input of the machine learning model, the machine learning model takes each set of comprehensive feature vector as the prediction target to predict the infrastructure ownership prediction confidence score label, minimizes the sum of prediction errors of all infrastructure ownership prediction confidence score labels as the training target, trains the machine learning model until the sum of prediction errors converges, and stops the model training, and determines the infrastructure ownership prediction confidence score according to the model output result, wherein the machine learning model is a polynomial regression model.

[0018] Preferably, when Y>0.8, it indicates that the ownership information of the infrastructure is highly reliable; when Y is between 0.5 and 0.8, it indicates that there is a certain uncertainty in the ownership information, which needs to be further checked; and when Y<0.5, it indicates that the ownership information is abnormal and needs to be manually reviewed.

[0019] Preferably, the matching confidence C of the infrastructure ownership information is set as: m C m =λ1Y+λ2S+λ3R; wherein: m C is the ownership matching confidence, indicating the confidence of the query result of the infrastructure, Y is the infrastructure ownership prediction confidence score, S is the query information completeness score, R is the historical matching success rate, indicating the frequency of successful matching of the infrastructure in the historical query record, λ1, λ2, and λ3 are weight coefficients for adjusting the influence of each index on the matching confidence, and satisfy λ1+λ2+λ3=1;

[0020] The user inputs a query condition of the infrastructure in the system, the system performs integrity check on the input information, and calculates a query information integrity score S;

[0021] A polynomial regression model is called to calculate an infrastructure ownership prediction confidence score Y;

[0022] The query integrity score S input by the user and the historical matching success rate R are combined to calculate a matching confidence C m ;

[0023] A matching confidence threshold C z is set: if C m ≥ C z , it is considered that the matching result is reliable, and the query result is directly returned; if C m < C z , it is considered that there is a problem with the matching result, and an artificial review process is triggered for further confirmation by the management personnel.

[0024] In the above technical solution, the technical effects and advantages provided by the present application are as follows:

[0025] 1. The present application solves the problems of lagging information update, low retrieval efficiency and inaccurate matching of existing GIS systems and property database by means of multi-source data fusion, intelligent analysis, machine learning modeling and real-time matching. The method uses natural language processing (NLP) and optical character recognition (OCR) to analyze property information, combines coordinate conversion and projection standardization to unify geographic information format, and ensures the consistency of data from different sources. By calculating the spatial deviation anomaly index and the data change rate fluctuation index, the system can automatically identify the geographical spatial anomalies and change instability in the property information, reducing the failure of property matching caused by information lag or errors. During the query process, the system predicts the infrastructure ownership confidence score based on the machine learning model, and calculates the matching confidence according to the query integrity and historical matching success rate, ensuring the accuracy and reliability of the query result.

[0026] 2. The present application can effectively improve the accuracy and response speed of urban infrastructure property information query, especially in emergency repair, emergency accident disposal, urban planning and property dispute resolution application scenarios, which can quickly match the property unit, reduce the delay of repair, and ensure public safety. In addition, the system provides property data services through GIS visualization display and API interface, supports government management departments, enterprises and municipal companies to efficiently retrieve and analyze infrastructure property information, and optimizes the property management process. The present application realizes the intelligentization of data-driven urban infrastructure ownership management, reduces manual intervention, improves query efficiency, and provides reliable technical support for smart city construction. BRIEF DESCRIPTION OF DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed in the embodiments will be briefly introduced as follows. Obviously, the accompanying drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art based on these drawings.

[0028] Figure 1 The method flowchart of the present application. DETAILED DESCRIPTION

[0029] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0030] Embodiment, please refer to Figure 1 As shown in the accompanying drawings, the method for quickly searching the ownership of urban infrastructure based on big data analysis comprises the following steps.

[0031] Collecting GIS, property registration database, Internet of Things sensor, remote sensing image and historical archive data, using natural language processing and optical character recognition technology for analysis, and unifying geographic information format through coordinate conversion and projection standardization;

[0032] Comparing the infrastructure spatial information in the GIS database with the geographic coordinates in the property registration database, marking out the geographic spatial inconsistency data, analyzing the change trend of the key fields of the property unit name, contact method and record time, and if multiple changes occur within a fixed time window, marking it as property association instability data;

[0033] Combining the geographic spatial inconsistency data and the property association instability data, using machine learning model to train infrastructure ownership prediction model, establishing ownership matching rules, and judging whether the ownership information is reliable;

[0034] After the user inputs the query condition, the system calls the matching model in real time, provides the ownership information query result, and marks the matching reliability. If the matching reliability of the query result is lower than the predetermined threshold, the artificial review process is triggered.

[0035] Acquire spatial location information of infrastructure (e.g., roads, bridges, underground pipelines) including point, line, and surface geometric data. Extract attribute data such as facility name, construction year, management unit, and maintenance responsibility party. Data formats are usually stored in Shapefile (.shp), GeoJSON, KML, or PostGIS database.

[0036] Collect property data from government property registration systems, including facility name, ownership unit, property certificate number, and historical change records. Since property registration data is usually stored as a structured database (e.g., MySQL, PostgreSQL) or text document (e.g., PDF, Word), format parsing is required.

[0037] Collect data from intelligent devices on infrastructure (e.g., intelligent manhole covers, water pressure sensors, pipe network leak detectors) to supplement the real-time status of ownership information. Obtain IoT data through MQTT protocol, HTTP API, or WebSocket, and the format is usually JSON, XML, or Protobuf.

[0038] Obtain image data of urban infrastructure through satellite remote sensing, high-resolution aerial photography, and unmanned aerial photography. Remote sensing data is usually stored in GeoTIFF, JPEG 2000, HDF5 format, and image registration is required to match GIS coordinates.

[0039] Collect historical property registration archives, including scanned paper files, electronic PDFs, and handwritten archives. These archives are usually stored in image formats (e.g., JPEG, TIFF) or non-editable PDFs, and text parsing is required.

[0040] Parse unstructured text in the property database (e.g., property description, historical change records). Use named entity recognition (NER) to extract key information such as property unit, address, facility name, and property change time. Perform syntactic analysis to identify property relationships such as "XX Company transferred to YY Company in 2020."

[0041] Recognize property information in scanned documents or handwritten text through OCR engines (e.g., Tesseract OCR, PaddleOCR). Use layout analysis techniques to detect tables, seals, and annotations to extract key property fields. Combine deep learning models (e.g., CRNN+CTC) to correct OCR recognition errors and improve text parsing accuracy.

[0042] Different data sources use different coordinate formats (e.g. latitude and longitude coordinates, projected coordinates, geographic grid numbers), which need to be converted. For example: WGS-84 (GPS standard) -> GCJ-02 (Chinese Mars coordinate) -> BD-09 (Baidu map coordinate); UTM (Universal Transverse Mercator) -> EPSG:4326 (WGS-84).

[0043] Different surveying and mapping systems use different projection methods (e.g. Mercator projection, Lambert projection, Gauss-Kruger projection), which need to be unified to the same coordinate reference system (CRS). Use Proj4 or GDAL for projection transformation to ensure that all data sources are aligned on the same geographic reference. Use nearest neighbor interpolation and affine transformation to correct coordinate errors to ensure accurate geographic position matching of multi-source data.

[0044] Extract spatial coordinates (latitude / longitude / projection coordinates), facility number, property unit, etc. from the GIS database. Extract geographic coordinates, property ownership information, registration time, etc. from the property registration database.

[0045] Coordinate conversion: ensure that the coordinate systems of the two databases are consistent (e.g. WGS-84, GCJ-02, EPSG:4326), and use projection conversion algorithms (GDAL / Proj4) for standardization. Spatial comparison: calculate the spatial deviation ΔD of the same infrastructure in the two databases, defined as:

[0046] where X GIS , Y GIS are the coordinates in the GIS database, and X 登记 , Y 登记 are the coordinates in the property registration database. Set a spatial error threshold TD, if ΔD> TD (e.g. greater than 5 meters), mark the data as geospatial inconsistency data.

[0047] If the spatial distribution of multiple facilities of the same property unit in the GIS database is seriously inconsistent with the registration database, Voronoi diagram division or K-means clustering analysis can be performed to detect abnormal areas. For linear infrastructure such as pipe network, cable, road, etc., calculate the topological relationship to ensure that the topological structure in the GIS database is consistent with the property registration.

[0048] After analyzing the spatial deviation data changes of the same infrastructure in the two databases, generate a spatial deviation anomaly index, and the method for obtaining the spatial deviation anomaly index is:

[0049] Obtain the coordinate information of the infrastructure from the two databases (GIS database & property registration database), calculate the spatial deviation ΔD of each facility at different times, and the expression is: wherein X GIS , Y GIS are the coordinates in the GIS database, X 登记 , Y 登记 are the coordinates in the property registration database. The exponentially weighted moving average of the spatial deviation data is calculated, expressed as: S t = aD + (1-a)S t-1 ; wherein: S t represents the exponentially weighted moving average of the spatial deviation data at time t, a is a smoothing coefficient, and the value range is 0 < a < 1 (usually 0.1-0.3), S t-1 is the exponentially weighted moving average of the spatial deviation data at time t-1 (the initial value S0 can be set as the initial spatial deviation D0), and the spatial deviation anomaly index is calculated: wherein S min , S max are the minimum and maximum exponentially weighted moving average values known in the historical data, used for normalization processing, and DC is the spatial deviation anomaly index. When DC ≈ 1, it indicates that the spatial deviation anomaly degree is high, and manual review may be needed; when DC ≈ 0, it indicates that the spatial deviation is normal, and there is no abnormal situation.

[0050] The property unit name, contact information, and record time fields are extracted from the property registration database, and the historical change records are recorded. A fixed time window (such as 6 months or 1 year) is set, and the number of property field changes in the window is counted.

[0051] The number of changes N W of the property field in the time window T C is calculated, expressed as: wherein C t is the property unit name or contact information at time t, and δ is the change detection function, which counts if the field changes. The change rate R C is calculated: If R C exceeds the set threshold T C (such as more than 1 time per 3 months), the property information is marked as property association instability data.

[0052] If the property unit frequently changes, but the contact information and record time remain stable, it may be a change of enterprise name, and further verification of legal person information is needed. If the contact information frequently changes, but the property unit name remains unchanged, it may be a transfer of management right, and confirmation from the actual management unit is needed. If the record time is abnormally updated (such as multiple updates in a short time), it may involve property disputes or illegal changes, and manual checking is needed.

[0053] Store the anomaly labeling data in a SQL database (e.g. PostgreSQL + PostGIS) or a NoSQL database (e.g. MongoDB). Create an index table for anomaly data, recording information such as anomaly type (spatial inconsistency / unstable property rights), facility number, change frequency, change rate, etc. Provide a data interface (API) for the management department or query system to call, for further manual review or model optimization.

[0054] After analyzing the fluctuation of data change rate in the property rights correlation instability data, generate the data change rate fluctuation index. The method for obtaining the data change rate fluctuation index is:

[0055] Collect property change records, including historical change data of property unit name, contact information, and record time fields, calculate the change rate R of the property field within a fixed time window C ;

[0056] Build a feature vector to convert the data of each time window into a format that can be used by machine learning, including: change frequency (such as the number of unit name changes, contact information changes), change interval time (the average time between two changes), and change duration (the length of time the change field remains stable). Train a random forest model, divide the training set and test set, generally divide 80% training data and 20% test data, initialize the random forest regression model, set the number of decision trees (such as 100-500) and the maximum depth to prevent overfitting, train the model, use past property change rate data as input, let the random forest learn the normal change pattern, use test data to evaluate the prediction error of the model, adjust the hyperparameters to improve accuracy.

[0057] Input the change data of the current time window, let the random forest predict the normal change rate range, calculate the difference between the actual change rate and the predicted change rate as the data change rate fluctuation index.

[0058] If the data change rate fluctuation index exceeds 0.7 (for example, frequent changes of property units in a short period of time), mark the data as high risk and recommend manual review.

[0059] The spatial deviation anomaly index and the data change rate fluctuation index are converted into a comprehensive feature vector, the comprehensive feature vector is taken as an input of a machine learning model, the machine learning model takes a predicted infrastructure ownership prediction confidence score value label as a prediction target for each group of comprehensive feature vectors, takes minimizing a sum of prediction errors of all infrastructure ownership prediction confidence score value labels as a training target, and is trained until the sum of prediction errors converges, and then the model training is stopped, and a final infrastructure ownership prediction confidence score value is determined according to an output result of the model, wherein the machine learning model is a polynomial regression model.

[0060] The method for obtaining the infrastructure ownership prediction confidence score value is: obtaining a corresponding function expression Y=F(DC,GH) from comprehensive feature vector training data of the trained machine learning model; in the function expression, F is an output function of the model, DC is the spatial deviation anomaly index, GH is the data change rate fluctuation index, and Y is the infrastructure ownership prediction confidence score value.

[0061] When Y>0.8, it is indicated that the ownership information of the infrastructure is highly reliable; when Y is between 0.5 and 0.8, it is indicated that the ownership information has certain uncertainty and needs to be further checked; and when Y<0.5, it is indicated that the ownership information is abnormal, and there may be a property dispute or data error, and manual review is needed.

[0062] The prediction confidence score value is displayed in a GIS system, and a color gradient is used to represent the confidence (for example, green represents high confidence, and red represents low confidence). The API is used to provide the government departments, enterprises and the like with query results of the infrastructure ownership stability analysis, so as to optimize the property management process.

[0063] After a user inputs a query condition (for example, a geographic location, a facility number or a property unit name), the system needs to call a matching model in real time, provide query results according to a calculated infrastructure ownership prediction confidence score value, and mark the confidence.

[0064] The matching confidence C of the infrastructure ownership information is set as follows: m C m =λ1Y+λ2S+λ3R; wherein: C mTo configure the confidence of the infrastructure ownership, indicating the credibility of the query result of the infrastructure, the value range is [0, 1], the higher the value, the more reliable the matching result, Y is the infrastructure ownership prediction credibility score, calculated by a polynomial regression model, indicating the reliability of the infrastructure ownership, the value range is [0, 1]. S is the query information completeness score, used to measure whether the query information provided by the user is complete, such as whether the facility number, property unit name, geographic location and other fields are complete, the value range is [0, 1]. R is the historical matching success rate, indicating the frequency of successful matching of the infrastructure in the historical query record, that is, whether the historical ownership information of the facility is stable, the value range is [0, 1]. λ1, λ2, λ3 are weight coefficients, used to adjust the influence of each index on the matching confidence, and satisfy λ1+λ2+λ3=1 (which can be adjusted through experiments, usually λ1 takes a larger value, such as 0.5-0.7).

[0065] The user inputs the query conditions of the infrastructure in the system, such as facility number, property unit, geographic coordinates, etc. The system checks the completeness of the input information and calculates the query information completeness score S.

[0066] Call the polynomial regression model to calculate the infrastructure ownership prediction credibility score Y;

[0067] Combine the user input query completeness score S and the historical matching success rate R to calculate the matching confidence C m .

[0068] Set the matching confidence threshold C z (e.g. 0.6): if C m ≥ C z , the matching result is considered reliable, and the query result is directly returned; if C m < C z , it is considered that the matching result may have problems, and the manual review process is triggered to further confirm by the management personnel.

[0069] The above formulas are dimensionless and the numerical values are calculated, the formula is obtained by software simulation of a large amount of data to obtain a formula of the latest real situation, and the preset parameters in the formula are set by the person skilled in the art according to the actual situation.

[0070] The above-described embodiments can be implemented in part or in whole through software, hardware, firmware or any combination thereof. When implemented in software, the above-described embodiments can be implemented using one or more computer programs written in any suitable programming language. The computer programs can be stored in one or more computer-readable storage media, such as a memory, a magnetic disk, an optical disk, a hard disk, a floppy disk, a magnetic tape, a memory card, a ROM, a DVD, a Blu-ray Disc, a CD, a semiconductor memory, a flash memory, or the like. The computer programs can be loaded into a computer, a server, a computer network, or the like, and executed. The computer programs can be distributed to computer systems connected to a network, and executed in parallel. The computer programs can be executed by a computer system that is capable of accessing a network, such as the Internet, and executed in parallel.

[0071] It should be understood that the term "and / or" in this document is merely used to describe associated objects, and can represent that three relationships are possible. For example, A and / or B can mean that the following three cases are possible: A exists alone, A and B exist together, and B exists alone, where A and B can be singular or plural. In addition, the character " / " in this document generally represents that the associated objects before and after the " / " have an "or" relationship, but can also represent an "and / or" relationship. The specific meaning can be understood according to the context before and after. Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed in this document can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0072] The above describes only the specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, and all such changes or replacements should be covered within the protection scope of the present application.

Claims

1. A method for rapidly identifying the ownership of urban infrastructure based on big data analytics, characterized in that: include: Collect GIS, property registration database, IoT sensor, remote sensing image and historical archive data, analyze them using natural language processing and optical character recognition technology, and standardize and unify geographic information format through coordinate transformation and projection. By comparing the spatial information of infrastructure in the GIS database with the geographic coordinates in the property registration database, geospatial inconsistencies are marked. The changing trends of key fields such as the name of the property owner, contact information, and filing time are analyzed. If multiple changes occur within a fixed time window, they are marked as data with unstable property association. Among them, a spatial deviation anomaly index is generated after analyzing the spatial deviation data changes of the same infrastructure in two databases. The method for obtaining the spatial deviation anomaly index is as follows: The coordinate information of the infrastructure is obtained from two databases, and the spatial deviation of each facility at different times is calculated. The expression is: ;in, , Coordinates from a GIS database, , The coordinates of the property registration database are used; the exponentially weighted moving average of the spatial deviation data is calculated using the following expression: ;in: This represents the exponentially weighted moving average of the spatial deviation data at time t, where α is the smoothing coefficient, with a value ranging from 0 to 1. for The spatial deviation anomaly index is calculated by using the exponentially weighted moving average of the spatial deviation data at different times. ;in: , The minimum and maximum exponentially weighted moving averages known from historical data are used for normalization, and DC is the spatial deviation anomaly index. Among them, the data change rate fluctuation index is generated after analyzing the data change rate fluctuation in the data on the instability of property rights association. The method for obtaining the data change rate fluctuation index is as follows: Collect historical change data for the property rights field, calculate the change rate of the property rights field within a fixed time window, construct a feature vector, and convert the data of each time window into a machine learning-compatible format, including the number of changes, change interval time, and change duration; train a random forest model, using past property rights change rate data as input, to let the random forest learn normal change patterns, use test data to evaluate the model's prediction error, input the change data of the current time window, let the random forest predict the normal change rate range, and calculate the difference between the actual change rate and the predicted change rate as the data change rate fluctuation index; By combining geospatial inconsistency data and property rights instability data, a machine learning model is used to train an infrastructure ownership prediction model, establish ownership matching rules, and determine the reliability of ownership information. Specifically, this includes: The spatial deviation anomaly index and the data change rate fluctuation index are converted into a comprehensive feature vector. The comprehensive feature vector is used as the input of the machine learning model. The machine learning model uses the prediction of the infrastructure ownership prediction credibility score label for each set of comprehensive feature vectors as the prediction objective and the minimization of the sum of prediction errors for all infrastructure ownership prediction credibility score labels as the training objective. The machine learning model is trained until the sum of prediction errors converges and the model training stops. The infrastructure ownership prediction credibility score is determined based on the model output. The machine learning model is a multinomial regression model. After the user enters the query conditions, the system calls the matching model in real time, provides the ownership information query results, and marks the matching confidence. If the matching confidence of the query results is lower than the predetermined threshold, the manual review process is triggered.

2. The method for rapidly finding the ownership of urban infrastructure based on big data analysis according to claim 1, characterized in that: Natural language processing technologies, including named entity recognition, syntactic analysis, and keyword extraction, are used to extract information such as property owners, addresses, facility names, and change times from property registration texts, historical archives, and handwritten documents. They are also combined with optical character recognition technology to parse unstructured text.

3. The method for rapidly finding the ownership of urban infrastructure based on big data analysis according to claim 1, characterized in that: The coordinate transformation and projection standardization includes coordinate format conversion, projection transformation and error correction, which unifies the different coordinate systems in the GIS database and the property registration database to the same geographic benchmark, and uses nearest neighbor interpolation or affine transformation to correct errors.

4. The method for rapidly finding the ownership of urban infrastructure based on big data analysis according to claim 1, characterized in that: When Y > 0.8, it indicates that the ownership information of the infrastructure is highly reliable; when Y is in a certain range... A value between 0.5 and 0.6 indicates some uncertainty regarding ownership information, requiring further verification; a value between 0.5 and 0.6 indicates abnormal ownership information, requiring manual review. This is the credibility score for infrastructure ownership prediction.

5. The method for rapidly finding the ownership of urban infrastructure based on big data analysis according to claim 4, characterized in that: Set the matching confidence level of infrastructure ownership information for: ;in: To match the confidence score, which represents the credibility of the query results for this infrastructure, Y is the infrastructure ownership prediction confidence score, S is the query information completeness score, and R is the historical matching success rate, representing the frequency with which the infrastructure successfully matches in historical query records. These are weighting coefficients used to adjust the impact of each indicator on the matching confidence level, and they satisfy the following conditions: =1; Users input infrastructure query criteria into the system, and the system performs a completeness check on the input information and calculates a query information completeness score S. The multinomial regression model is invoked to calculate the confidence score Y for infrastructure ownership prediction. The matching confidence score is calculated by combining the user's query completeness score S and the historical matching success rate R. ; Set matching confidence threshold :like ≥ If the match is successful, the result is considered reliable, and the query result is returned directly; otherwise... < If the matching result is not found to be correct, a manual review process will be initiated, and the manager will conduct further confirmation.

Images