Intelligent diagnosis method and device

Abstract

The invention relates to the technical field of standard address intelligent diagnosis, and particularly provides an intelligent diagnosis method and device, and the method comprises the following steps: S1, building and managing a standard address library, carrying out the management of professional resource addresses, and building a 11-level standard address library and 11-level field attributes; s2, a rule is preset, and atomic operation on a specific logic unit or a logic unit is completed; s3, preprocessing before intelligent diagnosis: establishing a feature data model according to standard address element features, performing model training and model evaluation according to the features and address tags, continuously adjusting and optimizing the model according to training and evaluation results, and performing intelligent positioning and classification on the problems of information loss and non-standard input of Chinese addresses according to the optimized model; s4, performing intelligent diagnosis on the standard address; and S5, storing the standard address after intelligent diagnosis. Compared with the prior art, the method has the advantages that the problems of non-standardization, repetition, errors and the like in the address data are effectively solved, and the quality and availability of the address data are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of standard address intelligent diagnosis, and specifically provides an intelligent diagnosis method and device. Background Art

[0002] In order to meet the people's growing demand for network access, standard addresses are used as the first entry point for business acceptance, connecting the front-end and back-end processes. In the long-term business flow, due to the uncertainty of manual input and the complexity of Chinese addresses, standard addresses have accumulated a lot of duplicate, irregular, and inaccurate problematic data, which has brought obstacles to a large number of application scenarios based on address information.

[0003] In order to improve the quality and accuracy of addresses and efficiently support business development, we have built intelligent diagnosis capabilities for standard addresses based on AI model learning and training, and created an 11-level standard address intelligent conversion function for text addresses. This supports intelligent cutting and intelligent error correction of standard addresses, improves front-line operation and maintenance efficiency, and supports efficient business launch.

[0004] The current shortcomings of address data:

[0005] (1) Inaccurate address data: The 11-level standard addresses in the current system have missing address information, redundant address information, untimely address information updates, and incorrect address names. As a result, front-line business ordering personnel often have problems selecting addresses when activating business orders, or incorrect address selections lead to order errors, which affects business orders.

[0006] (2) Similar address names: Since address names generally contain about 28 Chinese characters, the prefix of the address is very long, which leads to incorrect address selection.

[0007] (3) Address conversion error: The address obtained by the front desk is a text address. When placing an order for a business, it needs to be converted into level 11 address data in the system, resulting in an address error.

[0008] (4) Complex address maintenance: Manual hierarchical maintenance of 11 levels of address data, with many attributes and many maintenance times, leads to inaccurate address data. Summary of the invention

[0009] The present invention aims at solving the above-mentioned deficiencies of the prior art and provides an intelligent diagnosis method with strong practicability.

[0010] A further technical task of the present invention is to provide an intelligent diagnostic device that is rationally designed, safe and applicable.

[0011] The technical solution adopted by the present invention to solve its technical problem is:

[0012] An intelligent diagnosis method comprises the following steps:

[0013] S1. Standard address database construction and management, management of professional resource addresses, establishment of 11-level standard address database and 11-level field attributes;

[0014] S2. Preset rules to split various operation instructions of standard address intelligent diagnosis into operation units of minimum granularity for management. The split unit operations are used as a set of instructions to complete atomic operations on specific logical units or logical units.

[0015] S3: Pre-processing before intelligent diagnosis: Establish a feature data model based on the features of standard address elements, extract features through deep learning algorithms, train and evaluate models based on features and address labels, continuously adjust and optimize the model based on training and evaluation results, and intelligently locate and classify missing information and irregular input problems of Chinese addresses according to the optimized model;

[0016] S4, standard address intelligent diagnosis;

[0017] S5. Standard address is put into storage after intelligent diagnosis.

[0018] Furthermore, in step S1, the 11-level field attributes are as follows: 1: province; 2: prefecture-level city; 3: district and county; 4: township and street; 5: road / lane / street; 6: house number; 7: community / administrative village / unit / building; 8: building; 9: unit; 10: floor; 11: room number.

[0019] Furthermore, in step S2, various operation instructions of the standard address intelligent diagnosis are split into operation units of minimum granularity for management, and the split unit operations are used as a set of instructions to complete atomic operations on specific logical units or logical units.

[0020] Furthermore, in step S3, it includes:

[0021] S3-1. Establish a standard address knowledge graph;

[0022] S3-2, collect data and establish address database;

[0023] S3-3. Training on accurate understanding and recognition of addresses.

[0024] Further, in step S3-1, a standard address knowledge graph is constructed according to the 11-level standard address structure;

[0025] In step S3-2, a unified and standardized address library is established, and the existing addresses are included in the standardized address library. Newly added and non-standardized addresses are classified and identified.

[0026] Furthermore, in step S3-3, the preprocessed data is input into the intelligent diagnosis program, and the system will automatically perform address element recognition, standardization, error correction, matching, inference and analysis operations to realize the conversion of addresses in various formats and expressions into a unified standardized format. At the same time, it has the address matching capability, that is, matching the address with the data in the address database to find the matching postal code and administrative division function.

[0027] Furthermore, in step S4, it includes:

[0028] S4-1, 11-level standard address intelligent error correction;

[0029] Based on the 11-level standard address knowledge graph and the collected 11-level standard address data, the deep learning algorithm is used to intelligently identify, classify and correct massive address data. The NLP technology is combined to improve the understanding and analysis capabilities of address data, automatically discover and correct these errors, and display the processing results to users in the form of GIS maps, reports, etc., and accept user feedback and command input;

[0030] S4-2, 11-level standard address intelligent cutting;

[0031] When entering or using an address, the address is entered or selected in sequence multiple times according to the level; the address names are highly similar, making it difficult to enter or select the address. Based on the intelligent diagnosis capability of standard addresses, the hierarchical presentation of standard addresses is achieved.

[0032] Furthermore, in step S5, the standard address database is automatically updated based on the newly added and existing standard address data that have been intelligently diagnosed.

[0033] An intelligent diagnostic device, comprising: at least one memory and at least one processor;

[0034] The at least one memory is used to store a machine-readable program;

[0035] The at least one processor is used to call the machine-readable program to execute an intelligent diagnosis method.

[0036] Compared with the prior art, the intelligent diagnosis method and device of the present invention have the following outstanding beneficial effects:

[0037] (1) Automation and intelligence: Through text analysis of addresses, address information can be automatically identified, standardized, corrected, matched, inferred, and analyzed, avoiding the tediousness and errors of manual operations.

[0038] (2) Reduce operation and maintenance costs and improve operation and maintenance efficiency: Through intelligent diagnosis of standard addresses, problems such as inaccurate addresses, missing address information, redundant address information, and untimely address information updates caused by manual operations can be reduced, operation and maintenance investment can be reduced, and operation and maintenance efficiency can be improved, thereby achieving cost reduction and efficiency improvement.

[0039] (3) Providing decision-making support capabilities based on big data analysis: Standard address intelligent diagnosis capabilities can conduct statistics and analysis on large amounts of address data, obtain valuable information and trends, and provide data support for decision-making, thereby better meeting various business needs.

[0040] (4) Optimize resource allocation: Based on accurate address data, enterprises can plan resource allocation more reasonably, avoid waste and redundancy of resources, and further reduce costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Attached Figure 1 It is a schematic diagram of the architecture of an intelligent diagnosis method;

[0043] Attached Figure 2 It is a schematic diagram of intelligent diagnosis analysis in an intelligent diagnosis method;

[0044] Attached Figure 3 It is a flow chart of an intelligent diagnosis method;

[0045] Attached Figure 4 It is a schematic diagram of intelligent cutting of 11 standard addresses in an intelligent diagnosis method;

[0046] Attached Figure 5 It is a flow chart for training accurate understanding and recognition of addresses in intelligent diagnosis methods. DETAILED DESCRIPTION

[0047] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] A best embodiment is given below:

[0049] An intelligent diagnosis method in this embodiment is implemented under the following conditions:

[0050] (1) Equipment manufacturers should adopt a more open approach to enable equipment to support network management.

[0051] (2) It should be possible to centrally manage all address data involved in the network under its jurisdiction through a network operation control platform.

[0052] like Figure 1 As shown, the steps are as follows:

[0053] S1. Standard address database construction and management, management of professional resource addresses, establishment of 11-level standard address database and 11-level field attributes;

[0054] First, we implement the management of professional resource addresses such as wireless, home customers, and customer collection, and establish an 11-level standard address library. The 11-level field attributes are as follows: 1: Province; 2: City; 3: District and County; 4: Township and Street; 5: Road / Lane / Street; 6: House Number; 7: Community / Administrative Village / Unit / Building; 8: Building; 9: Unit; 10: Floor; 11: Room Number. The complete standard address information is as follows: 1405, Unit 1, 14th Floor, No. 66, Jingsan Road, Wenhua Road, Jinshui District, Zhengzhou, Henan Province, China.

[0055] S2. Preset rules to split various operation instructions of standard address intelligent diagnosis into operation units of minimum granularity for management. The split unit operations are used as a set of instructions to complete atomic operations on specific logical units or logical units.

[0056] S3: Pre-processing before intelligent diagnosis: Establish a feature data model based on the features of standard address elements, extract features through deep learning algorithms, train and evaluate models based on features and address labels, continuously adjust and optimize the model based on training and evaluation results, and intelligently locate and classify missing information and irregular input problems of Chinese addresses according to the optimized model;

[0057] Further including:

[0058] S3-1. Establish a standard address knowledge graph: Construct a standard address knowledge graph according to the 11-level standard address structure;

[0059] S3-2. Collect data and establish an address database: Establish a unified and standardized address database, incorporate existing addresses into the standardized address database, and classify and identify new and non-standardized addresses;

[0060] S3-3, accurate understanding and recognition training of addresses;

[0061] like Figure 5As shown, data preprocessing: standardize Chinese addresses, such as using pinyin or Chinese characters uniformly, and perform word segmentation operations, so that the model can better process data.

[0062] Feature extraction: Use deep learning models (such as recurrent neural networks (RNNs)) to extract features from addresses, which may include features such as word frequency and geographic location.

[0063] Training model: Use the extracted features and corresponding address labels (such as correct, missing, irregular) to train the deep learning model. This may require a large amount of labeled data so that the model can learn how to classify and locate.

[0064] Model evaluation: Use the test data set to evaluate the performance of the model, including accuracy, recall, F1 score, etc. Adjust and optimize the model based on the evaluation results.

[0065] Model application: Apply the trained model to actual scenarios to intelligently locate and classify new Chinese addresses. Based on the output of the model, provide corresponding suggestions and corrections for addresses with missing or irregular information.

[0066] Continuous optimization: In the actual application process, new data is continuously collected, and the model is continuously optimized and updated to adapt to changes in address information and improve model performance.

[0067] During use, the pre-processed data is input into the intelligent diagnosis program, and the system will automatically perform operations such as identification, standardization, error correction, matching, inference and analysis of address elements, including automatic completion of the input address information, correction of typos, grammatical errors and other error information in the address, and conversion of addresses in various formats and different expressions into a unified standardized format. At the same time, it has the ability to match the address with the data in the address database, and find the matching postal code and administrative division. In the process of intelligent diagnosis, it has the function of inferring the relevant address information based on the known address information and performing accuracy verification.

[0068] S4, standard address intelligent diagnosis;

[0069] include:

[0070] S4-1, 11-level standard address intelligent error correction;

[0071] Based on the 11-level standard address knowledge graph and the collected 11-level standard address data, the deep learning algorithm is used to intelligently identify, classify and correct massive address data. The NLP technology is combined to improve the understanding and analysis capabilities of address data, automatically discover and correct these errors, and display the processing results to users in the form of GIS maps, reports, etc., and accept user feedback and command input;

[0072] S4-2, 11-level standard address intelligent cutting;

[0073] When entering or using an address, the address is entered or selected in sequence multiple times according to the level; the address names are highly similar, making it difficult to enter or select the address. Based on the intelligent diagnosis capability of standard addresses, the hierarchical presentation of standard addresses is achieved.

[0074] Furthermore, in step S5, the standard address database is automatically updated based on the newly added and existing standard address data that have been intelligently diagnosed.

[0075] S5, standard address is put into storage after intelligent diagnosis;

[0076] Based on the newly added and existing standard address data that has been intelligently diagnosed this time, it can be automatically updated into the standard address library.

[0077] Taking address resource input as an example, based on intelligent diagnosis of standard addresses, hierarchical addresses that meet 11 standards are automatically output and explained in detail.

[0078] S1. Construction and management of standard address database;

[0079] Establish stock standard addresses. Address information includes home, wireless, customer, space and other resource address information.

[0080] S2, preset rules;

[0081] Preset standard address intelligent diagnosis command operation template. Split various operation instructions of standard address intelligent diagnosis into the smallest granularity operation element for management. The split element operation is used as a set of instructions to complete the atomic operation for a specific logical unit or logical unit.

[0082] Taking the input address with typos and logical errors as an example, it can be divided into the following operation units:

[0083] (1) Input text address data;

[0084] (2) 11-level standard address intelligent error correction;

[0085] (3) 11-level standard address intelligent cutting;

[0086] (4) 11-level standard addresses are stored in the database.

[0087] S3, pre-processing before intelligent diagnosis;

[0088] A new 11-level standard address knowledge graph is created to collect address data from professional resources such as home customers, gathering customers, space, and wireless. Through address elements and based on the input text information, the elements contained in the address are identified, such as country, province, city, district, street, building number, etc.

[0089] S4, standard address intelligent diagnosis;

[0090] During resource access and adjustment process application, when entering / using addresses, if an incorrect address (an address with typos) or a duplicate standard address is entered, the standard address intelligent diagnosis capability will be triggered.

[0091] include:

[0092] S4-1, 11-level standard address intelligent error correction;

[0093] like Figure 2 , 3 As shown, based on the 11-level standard address knowledge graph and the collected 11-level standard address data, deep learning algorithms are used to intelligently identify, classify and correct massive address data. NLP technology is combined to improve the understanding and analysis capabilities of address data, automatically discover and correct these errors, and display the processing results to users in the form of GIS maps, reports, etc., and accept user feedback and command input.

[0094] The HMM and N-Gram deep learning prediction model algorithms are:

[0095] P(y|x)=1 / Z(x)exp[∑ K w k (f k (x,y))]

[0096] Where: Z(x) = ∑ y exp[∑ k w k (f k (x,y))]

[0097] f k (x,y)=∑ i w k (f k (y i-1 ,y i x,i))

[0098] Text length: s;

[0099] The position of the word i in the text;

[0100] The label of the word at the current position i ;

[0101] The label of the previous word l i -1;

[0102] The main codes of the logical algorithm error correction capability are as follows:

[0103]

[0104] The main codes for typos correction capabilities are as follows:

[0105]

[0106]

[0107] S4-2, 11-level standard address intelligent cutting;

[0108] like Figure 4 As shown, when entering / using an address, the address is entered / selected in sequence multiple times according to the level; the address names are highly similar, and address entry / selection is difficult. Based on the intelligent diagnosis capability of standard addresses, the hierarchical presentation of standard addresses is achieved.

[0109] S5, standard address is put into storage after intelligent diagnosis;

[0110] Based on the newly added and existing standard address data that has been intelligently diagnosed this time, it can be automatically updated into the standard address library.

[0111] Based on the above method, an intelligent diagnostic device in this embodiment includes: at least one memory and at least one processor;

[0112] The at least one memory is used to store a machine-readable program;

[0113] The at least one processor is used to call the machine-readable program to execute an intelligent diagnosis method.

[0114] The above-mentioned specific implementations are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementations. Any technical solutions that conform to the above-mentioned specific implementations of the present invention and any appropriate changes or substitutions made by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.

[0115] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent diagnosis method, characterized in that: The steps are as follows: S1. Standard address database construction and management, management of professional resource addresses, establishment of 11-level standard address database and 11-level field attributes; S2. Preset rules to split various operation instructions of standard address intelligent diagnosis into operation units of minimum granularity for management. The split unit operations are used as a set of instructions to complete atomic operations on specific logical units or logical units. S3: Pre-processing before intelligent diagnosis: Establish a feature data model based on the features of standard address elements, extract features through deep learning algorithms, train and evaluate models based on features and address labels, continuously adjust and optimize the model based on training and evaluation results, and intelligently locate and classify missing information and irregular input problems of Chinese addresses according to the optimized model; S4, standard address intelligent diagnosis; S5. Standard address is put into storage after intelligent diagnosis.

2. An intelligent diagnosis method according to claim 1, characterized in that: In step S1, the 11-level field attributes are as follows: 1: province; 2: prefecture-level city; 3: district and county; 4: township and street; 5: road / lane / street; 6: house number; 7: community / administrative village / unit / building; 8: building; 9: unit; 10: Floor; 11: Room number.

3. An intelligent diagnosis method according to claim 2, characterized in that: In step S2, various operation instructions of the standard address intelligent diagnosis are split into operation units of the smallest granularity for management, and the split unit operations are used as a set of instructions to complete atomic operations on specific logical units or logical units.

4. The intelligent diagnosis method according to claim 3, characterized in that: In step S3, it includes: S3-1. Establish a standard address knowledge graph; S3-2, collect data and establish address database; S3-3. Training on accurate understanding and recognition of addresses.

5. An intelligent diagnosis method according to claim 4, characterized in that: In step S3-1, a standard address knowledge graph is constructed according to the 11-level standard address structure; In step S3-2, a unified and standardized address library is established, and the existing addresses are included in the standardized address library. Newly added and non-standardized addresses are classified and identified.

6. An intelligent diagnosis method according to claim 5, characterized in that: In step S3-3, the preprocessed data is input into the intelligent diagnosis program, and the system will automatically perform address element recognition, standardization, error correction, matching, inference and analysis operations to convert addresses in various formats and expressions into a unified standardized format. At the same time, it has the address matching capability, that is, matching the address with the data in the address database to find the matching postal code and administrative division function.

7. The intelligent diagnosis method according to claim 6, characterized in that: In step S4, it includes: S4-1, 11-level standard address intelligent error correction; Based on the 11-level standard address knowledge graph and the collected 11-level standard address data, the deep learning algorithm is used to intelligently identify, classify and correct massive address data. The NLP technology is combined to improve the understanding and analysis capabilities of address data, automatically discover and correct these errors, and display the processing results to users in the form of GIS maps, reports, etc., and accept user feedback and command input; S4-2, 11-level standard address intelligent cutting; When entering or using an address, the address is entered or selected in sequence multiple times according to the level; the address names are highly similar, making it difficult to enter or select the address. Based on the intelligent diagnosis capability of standard addresses, the hierarchical presentation of standard addresses is achieved.

8. An intelligent diagnosis method according to claim 7, characterized in that: In step S5, the standard address database is automatically updated based on the newly added and existing standard address data that have been intelligently diagnosed.

9. An intelligent diagnostic device, characterized in that: include: at least one memory and at least one processor; The at least one memory is used to store a machine-readable program; The at least one processor is configured to call the machine-readable program to execute the method according to any one of claims 1 to 8.

Images