Work order information processing method, device, equipment and storage medium

Abstract

The present invention relates to the field of artificial intelligence technology, and is applied to the field of smart government affairs. A method, device, equipment and storage medium for processing work order information are provided to improve the efficiency of abnormal work order identification and processing. The method for processing work order information includes: selecting target abnormal work order fields and cleaning data on a historical work order information table to obtain data to be processed, and the historical work order information table includes an information piece table, an information provider table, a subject table involved and an object table involved; training a preset initial decision tree work order recognition model with the data to be processed to obtain a target decision tree work order recognition model; performing target abnormal work order field analysis on the work order information to be processed through the target decision tree work order recognition model to obtain a work order analysis result; and executing a business process corresponding to the target business process information corresponding to the work order analysis result. In addition, the present invention also relates to blockchain technology, and the historical work order information table can be stored in a blockchain.

Description

Technical Field

[0001] The present invention relates to the field of intelligent decision-making of artificial intelligence, and in particular to a method, device, equipment and storage medium for processing work order information. Background Art

[0002] With the development of information technology and the Internet of Things, the data processing of work orders has also increased, and the data processing of abnormal work orders has received particular attention. At present, the processing of abnormal work orders is generally done by distinguishing abnormal work orders from ordinary work orders, and performing text classification and analysis on the distinguished abnormal work orders.

[0003] However, due to the huge number of abnormal work orders, the fixed text classification strategy, and the dynamic changes in the characteristics of the work orders, there is a lack of flexible correlation analysis and analysis of the rules of multiple attributes. The text classification strategy is difficult to adapt to the flexible and changeable complaint strategy, and the processing efficiency is low, resulting in low efficiency in the identification and processing of abnormal work orders. Summary of the invention

[0004] The present invention provides a method, device, equipment and storage medium for processing work order information, which are used to improve the efficiency of abnormal work order identification and processing.

[0005] A first aspect of the present invention provides a method for processing work order information, comprising:

[0006] Obtain a historical work order information table, select target abnormal work order fields and perform data cleaning on the historical work order information table to obtain data to be processed, wherein the historical work order information table includes an information piece table, an information provider table, a subject table, and an object table;

[0007] Using the data to be processed, a preset initial decision tree work order recognition model is trained to obtain a target decision tree work order recognition model;

[0008] Obtaining work order information to be processed, and using the target decision tree work order recognition model, sequentially performing data preprocessing based on preset type work order fields, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on target feature fields on the work order information to be processed, to obtain work order analysis results;

[0009] Obtain target business process information corresponding to the work order analysis result, and execute the business process corresponding to the target business process information.

[0010] Optionally, in a first implementation of the first aspect of the present invention, the step of obtaining a historical work order information table, selecting target abnormal work order fields and performing data cleaning on the historical work order information table to obtain data to be processed includes:

[0011] Obtain a historical work order information table, create a full field database, identify preset type work orders, and extract feature fields for the historical work order information table to obtain associated feature fields;

[0012] The target abnormal work order field selection and data cleaning based on the preset data storage strategy are performed on the associated characteristic fields to obtain the data to be processed.

[0013] Optionally, in a second implementation of the first aspect of the present invention, the obtaining of the historical work order information table, the creation of a full field database, identification of preset type work orders and extraction of feature fields for the historical work order information table, and obtaining associated feature fields include:

[0014] Obtain a historical work order information table, associate the historical work order information table with the table, and obtain a full field database;

[0015] Matching the preset type work order fields on the full field database to obtain historical preset type work order data;

[0016] The historical field data in the historical preset type work order data is read, and correlation analysis and feature field extraction are performed on the historical field data in sequence to obtain correlation feature fields.

[0017] Optionally, in a third implementation of the first aspect of the present invention, the reading of historical field data in the historical preset type work order data, sequentially performing correlation analysis and feature field extraction on the historical field data to obtain the associated feature field includes:

[0018] Read the historical field data in the historical preset type work order data and the abnormal negative sample list data in the full field database;

[0019] Verifying the historical field data by using preset empirical rules to obtain verification field data;

[0020] By using preset correlation rules and the abnormal negative sample list data, variable correlation analysis and feature field extraction are performed on the verification field data to obtain correlation feature fields.

[0021] Optionally, in a fourth implementation of the first aspect of the present invention, the step of selecting target abnormal work order fields for the associated feature fields and performing data cleaning based on a preset data storage strategy to obtain data to be processed includes:

[0022] By using the preset abnormal work order dependent variable, a correlation analysis is performed on the associated characteristic field to obtain a correlation analysis result;

[0023] Determine whether the correlation analysis result is randomly distributed, and if the correlation analysis result is not randomly distributed, determine the corresponding correlation feature field as the field data to be cleaned;

[0024] According to the preset data storage strategy, the field data to be cleaned is deduplicated, the field meaning is converted, the data format is converted and normalized to obtain the data to be processed.

[0025] Optionally, in a fifth implementation of the first aspect of the present invention, the work order information to be processed is obtained, and through the target decision tree work order recognition model, the work order information to be processed is sequentially subjected to data preprocessing based on preset type work order fields, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on the target feature fields, to obtain a work order analysis result, including:

[0026] Acquire the work order information to be processed, and use the target decision tree work order recognition model to sequentially perform preset type work order field matching, field meaning conversion and normalization processing on the work order information to be processed, so as to obtain the preset type work order data to be processed;

[0027] Extract and statistically analyze the associated feature fields of the preset type of work order data to be processed to obtain associated analysis statistical data and associated fields of the work order to be processed;

[0028] Through the association analysis statistical data, a judgment analysis based on random distribution is performed on the association fields of the to-be-processed work orders to obtain target feature fields;

[0029] The target feature field is matched with an abnormal work order field of a preset type of work order field to obtain a work order analysis result.

[0030] Optionally, in a sixth implementation of the first aspect of the present invention, the acquiring target business process information corresponding to the work order analysis result and executing the business process corresponding to the target business process information includes:

[0031] Receiving a business execution request, parsing the business execution request and extracting key information, and obtaining target business process information corresponding to the work order analysis result;

[0032] According to the target business process information, traverse the preset business execution robot decision tree to obtain the target business execution robot;

[0033] The target business execution robot is called to execute the business process corresponding to the target business process information.

[0034] A second aspect of the present invention provides a device for processing work order information, comprising:

[0035] An acquisition and cleaning module is used to acquire a historical work order information table, select target abnormal work order fields and perform data cleaning on the historical work order information table to obtain data to be processed, wherein the historical work order information table includes an information piece table, an information provider table, a subject table involved, and an object table involved;

[0036] A training module, used to train a preset initial decision tree work order recognition model through the data to be processed to obtain a target decision tree work order recognition model;

[0037] An analysis module is used to obtain work order information to be processed, and through the target decision tree work order recognition model, sequentially perform data preprocessing based on preset type work order fields, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on target feature fields on the work order information to be processed, to obtain work order analysis results;

[0038] The execution module is used to obtain the target business process information corresponding to the work order analysis result and execute the business process corresponding to the target business process information.

[0039] Optionally, in a first implementation of the second aspect of the present invention, the acquisition and cleaning module includes:

[0040] Create an extraction submodule for obtaining a historical work order information table, create a full field database for the historical work order information table, identify preset type work orders, and extract feature fields to obtain associated feature fields;

[0041] The cleaning submodule is used to select target abnormal work order fields for the associated characteristic fields and clean the data based on a preset data storage strategy to obtain data to be processed.

[0042] Optionally, in a second implementation of the second aspect of the present invention, the creating an extraction submodule includes:

[0043] An association unit, used to obtain a historical work order information table, and perform table association on the historical work order information table to obtain a full field database;

[0044] A matching unit, configured to match the full field database based on the preset type of work order field to obtain historical preset type work order data;

[0045] The reading and analysis unit is used to read the historical field data in the historical preset type work order data, and perform correlation analysis and feature field extraction on the historical field data in sequence to obtain the correlation feature field.

[0046] Optionally, in a third implementation of the second aspect of the present invention, the reading and analyzing unit is specifically used for:

[0047] Read the historical field data in the historical preset type work order data and the abnormal negative sample list data in the full field database;

[0048] Verifying the historical field data by using preset empirical rules to obtain verification field data;

[0049] By using preset correlation rules and the abnormal negative sample list data, variable correlation analysis and feature field extraction are performed on the verification field data to obtain correlation feature fields.

[0050] Optionally, in a fourth implementation of the second aspect of the present invention, the cleaning submodule is specifically used for:

[0051] By using the preset abnormal work order dependent variable, a correlation analysis is performed on the associated characteristic field to obtain a correlation analysis result;

[0052] Determine whether the correlation analysis result is randomly distributed, and if the correlation analysis result is not randomly distributed, determine the corresponding correlation feature field as the field data to be cleaned;

[0053] According to the preset data storage strategy, the field data to be cleaned is deduplicated, the field meaning is converted, the data format is converted and normalized to obtain the data to be processed.

[0054] Optionally, in a fifth implementation of the second aspect of the present invention, the analysis module is specifically used to:

[0055] Acquire the work order information to be processed, and use the target decision tree work order recognition model to sequentially perform preset type work order field matching, field meaning conversion and normalization processing on the work order information to be processed, so as to obtain the preset type work order data to be processed;

[0056] Extract and statistically analyze the associated feature fields of the preset type of work order data to be processed to obtain associated analysis statistical data and associated fields of the work order to be processed;

[0057] Through the association analysis statistical data, a judgment analysis based on random distribution is performed on the association fields of the to-be-processed work orders to obtain target feature fields;

[0058] The target feature field is matched with an abnormal work order field of a preset type of work order field to obtain a work order analysis result.

[0059] Optionally, in a sixth implementation of the second aspect of the present invention, the execution module is specifically used to:

[0060] Receiving a business execution request, parsing the business execution request and extracting key information, and obtaining target business process information corresponding to the work order analysis result;

[0061] According to the target business process information, traverse the preset business execution robot decision tree to obtain the target business execution robot;

[0062] The target business execution robot is called to execute the business process corresponding to the target business process information.

[0063] The third aspect of the present invention provides a work order information processing device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the work order information processing device executes the above-mentioned work order information processing method.

[0064] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned method for processing work order information.

[0065] In the technical solution provided by the present invention, a historical work order information table is obtained, and target abnormal work order fields are selected and data cleansed on the historical work order information table to obtain data to be processed, wherein the historical work order information table includes an information piece table, an information provider table, a subject table involved, and an object table involved; a preset initial decision tree work order recognition model is trained through the data to be processed to obtain a target decision tree work order recognition model; the work order information to be processed is obtained, and through the target decision tree work order recognition model, the work order information to be processed is sequentially preprocessed based on data of preset type work order fields, extracted based on target feature fields of association analysis statistics, and analyzed based on abnormal work order fields of target feature fields to obtain work order analysis results; the target business process information corresponding to the work order analysis results is obtained, and the business process corresponding to the target business process information is executed. In an embodiment of the present invention, target abnormal work order fields are selected and data is cleaned for the historical work order information table to train a target decision tree work order recognition model. The target abnormal work order field analysis is performed on the work order information to be processed through the target decision tree work order recognition model, and the business process corresponding to the target business process information is executed. The work order information to be processed can be flexibly analyzed for correlation and the rules of multiple attribute combinations, and flexible and changeable work order complaint reporting strategies can be adapted. The universality and accuracy of the target decision tree work order recognition model are improved, thereby improving the efficiency of abnormal work order identification and processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic diagram of an embodiment of a method for processing work order information in an embodiment of the present invention;

[0067] Figure 2 A schematic diagram of another embodiment of a method for processing work order information in an embodiment of the present invention;

[0068] Figure 3 A schematic diagram of an embodiment of a device for processing work order information in an embodiment of the present invention;

[0069] Figure 4 A schematic diagram of another embodiment of a device for processing work order information in an embodiment of the present invention;

[0070] Figure 5 The figure is a schematic diagram of an embodiment of a device for processing work order information in an embodiment of the present invention. DETAILED DESCRIPTION

[0071] The embodiments of the present invention provide a method, device, equipment and storage medium for processing work order information, which improve the efficiency of identifying and processing abnormal work orders.

[0072] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0073] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of a method for processing work order information in an embodiment of the present invention includes:

[0074] 101. Obtain a historical work order information table, select target abnormal work order fields and perform data cleaning on the historical work order information table to obtain data to be processed. The historical work order information table includes an information piece table, an information provider table, a subject table involved, and an object table involved.

[0075] It is understandable that the execution subject of the present invention may be a device for processing work order information, or may be a terminal or a server, which is not specifically limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0076] Among them, the historical work order information table is an information table of historical work orders, and historical work orders include but are not limited to trademark work orders, advertising work orders, quality work orders and / or unfair competition work orders, etc. In this embodiment, historical work orders are preferably advertising work orders. The historical work order information table is composed of the key elements of historical work orders, including subject information, object information and behavior information based on historical work orders. The information provider table includes the personal information of the information provider, for example: taking the historical work order as a complaint and report work order as an example, the information provider table includes the name and contact information of the information provider; the subject table includes the information of the market entity complained and reported, for example: taking the historical work order as a complaint and report work order as an example, the market entity complained and reported is XX Co., Ltd., and the subject type is a domestic enterprise; the object table includes the product and commodity information involved in the historical work order, such as: commodity name massage chair, product category medical device and amount data.

[0077] The server extracts the historical work order information table from the historical database of the target city. The historical work order information table includes the information item table, the information provider table, the subject table involved and the object table involved. The information item table is a table that stores the information fields of the information item. The server creates the information provider primary key of the information provider table, the subject primary key of the subject table involved and the object primary key of the object table involved, as well as the information provider foreign key, the subject foreign key and the object foreign key involved in the information item table. By associating the information provider primary key with the information provider foreign key, associating the subject primary key with the subject foreign key, and associating the object primary key with the object foreign key, a total table is generated, that is, a full field database. The total table includes the subject characteristics, object characteristics and behavior characteristics of the historical work order. For example, Zhang San complained to the responsible unit about XX Co., Ltd. at 9:00 on December 12, 2020, believing that the massage chair provided by the company had quality defects and demanded compensation of XX yuan.

[0078] By associating the historical work order information tables, cross-correlation analysis of fields belonging to different tables is achieved. The subject type in the subject table and the product type in the object table are independent of each other in the existing business system, but association analysis and data mining can be performed in the "associated" table. After the server creates a full-field database, it calculates the similarity between the work order data in the full-field database and the preset work order data template (such as: advertising complaint and reporting work order), and determines whether the similarity is greater than the preset threshold. If so, the work order data is determined as historical preset type work order data (such as: historical advertising complaint and reporting work order data), and the historical preset type work order data is exported to a character-separated value (CSV) file, and the CSV file is imported into the value visualization tool PowerBI to facilitate the visualization and data exploration of the historical preset type work order data.

[0079] The server verifies the historical preset type work order data through preset experience rules, obtains the verification field, extracts the field corresponding to the verification field in the historical preset type work order data, and thus obtains the associated feature field, wherein the experience rules include the work order features for identifying the historical preset type work order data. For example, taking the historical preset type work order data as the historical advertising complaint and reporting work order data as an example, the grassroots law enforcement personnel have accumulated some artificially summarized work order features about professional counterfeiters and professional counterfeiting work orders in the historical advertising complaint and reporting work order data based on their daily work experience. The work order features include multiple complaints in a short period of time, familiarity with laws and regulations, complaint content citing legal provisions such as the Advertising Law and writing formatting, etc.

[0080] The server obtains the variables in the historical preset type work order data, calculates the correlation between the associated feature field and the variable through the preset Pearson correlation coefficient algorithm, and determines whether the correlation is greater than the preset correlation. If so, the corresponding field in the associated feature field is determined as the target abnormal work order field, and the target abnormal work order field is extracted to obtain the field data to be cleaned. If not, a null value is returned. The server fills the field data to be cleaned with null values ​​according to the preset data storage strategy to obtain the null-filled data, and normalizes the null-filled data through the preset batch normalization algorithm (batchnormalization, BN) to obtain the data to be processed, wherein the data to be processed includes field attributes.

[0081] 102. The preset initial decision tree work order recognition model is trained using the data to be processed to obtain a target decision tree work order recognition model.

[0082] Specifically, the server randomly distributes the data to be processed to obtain a training data set and a test data set, and trains a preset initial decision tree work order recognition model through the training data set to obtain a trained decision tree work order recognition model. For example, the server randomly distributes the data to be processed into a training data set and a test data set in a ratio of 7:3, recognizes the training data set through the preset initial decision tree work order recognition model to obtain a recognition result, calculates the loss function value of the recognition result through a preset loss function, and adjusts the learning rate and hyperparameters in the initial decision tree work order recognition model based on the loss function value through a gradient descent algorithm, thereby obtaining a trained decision tree work order recognition model;

[0083] The training decision tree work order recognition model is used to perform a judgment analysis on the test data set based on the preset abnormal type work order to obtain a test result, and the target loss function value of the test result is calculated. The target loss function value includes a vector loss function value and a classification loss function value. That is, the server judges and analyzes whether each work order in the test data set is a preset abnormal type work order (such as an advertising-related professional counterfeiting complaint and reporting work order) through the training decision tree work order recognition model. The execution process of the judgment analysis is as follows: the training decision tree work order recognition model starts from the root of the tree, and continuously selects the attribute with the largest information gain as the discrimination condition. After separating the test data set, the attribute with the largest information gain among the remaining attributes is selected as the discrimination condition by divide and conquer, and so on recursively until the purity of the leaf node reaches the maximum, and the test result is obtained. The test result includes the fields and field values ​​corresponding to "yes" or "no" respectively. The server calculates the test comprehensive central vector of all data in the test data set and the single central vector of each data in the test data set through the preset central vector formula, and calculates the absolute value of the difference between the single central vector and the test comprehensive central vector to obtain the vector loss function value, calculates the classification loss function value of the test result through the preset regression loss function, and performs weighted summation of the vector loss function value and the classification loss function value to obtain the target loss function value;

[0084] Determine whether the target loss function value is less than a preset threshold value. If the target loss function value is less than the preset threshold value, determine the trained decision tree work order recognition model as a candidate decision tree work order recognition model, that is, the server determines whether the target loss function value is less than the preset threshold value. If so, determine the trained decision tree work order recognition model as a candidate decision tree work order recognition model. If not, obtain the weight of the trained decision tree work order recognition model, iteratively adjust the weight until the target loss function value is less than the preset threshold value and converges, and obtain the error result in the test result. Through the preset residual network ResNet algorithm and multiple integrated learning weak classifiers, based on the error result, adjust the branch and hierarchical relationship of the attribute with the largest information gain as a discrimination condition in the trained decision tree work order recognition model, so as to obtain the candidate decision tree work order recognition model, thereby improving the accuracy of the candidate decision tree work order recognition model;

[0085] Acquire the historical data of project sedimentation, and train, tune and prune the candidate decision tree work order recognition model in sequence according to the historical data of project sedimentation to obtain the target decision tree work order recognition model, that is, the server searches the preset landing city database based on the information of the landing city and the complaint and reporting work order through the preset search engine to obtain the project sedimentation historical data of the corresponding landing city, and divide the project sedimentation historical data into training data and test data. Through the training data, train the candidate decision tree work order recognition model to obtain the trained candidate decision tree work order recognition model, and predict the test data based on the preset abnormal type work order (such as advertising-related professional counterfeiting complaint and reporting work order) through the trained candidate decision tree work order recognition model to obtain the prediction result, and obtain the comparison result based on the prediction result, which can be obtained by the landing city business personnel through judgment and review of the prediction result, and obtain the weight of the trained candidate decision tree work order recognition model, adjust the weight according to the comparison result, and prune and optimize the trained candidate decision tree work order recognition model through the preset pruning algorithm, so as to obtain the target decision tree work order recognition model.

[0086] 103. Obtain the work order information to be processed, and through the target decision tree work order recognition model, perform data preprocessing based on preset type work order fields, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on target feature fields on the work order information to be processed, and obtain the work order analysis results.

[0087] The server uses the target decision tree work order recognition model to perform preset type work order field matching, deduplication processing, field meaning conversion, data format conversion and normalization processing on the work order information to be processed in sequence to obtain the preset type work order data to be processed, perform correlation analysis and extract related feature fields on the preset type work order data to obtain the associated fields of the work order to be processed, and perform statistics on the preset type work order data to obtain the correlation analysis statistical data, and determine whether the correlation analysis statistical data is randomly distributed. If so, stop execution; if not, determine the corresponding associated fields of the work order to be processed as the target feature fields, call the target decision tree work order recognition model, perform judgment analysis on the target feature fields based on the preset abnormal type work orders (abnormal work order fields), and obtain the work order analysis results.

[0088] 104. Obtain target business process information corresponding to the work order analysis result, and execute the business process corresponding to the target business process information.

[0089] The server obtains the work order analysis result, which includes preset abnormal type work orders and judgment analysis results based on the preset abnormal type work orders. The work order analysis results are, for example, advertising professional counterfeiting complaint and reporting work orders and judgment analysis results based on advertising professional counterfeiting complaint and reporting work orders. The server matches the corresponding target personnel according to the work order analysis results. The target personnel are, for example, market supervision personnel who are more professional and experienced in handling advertising professional counterfeiting complaint and reporting work orders, and sends the work order analysis results to the terminal corresponding to the target personnel, so that the target personnel can perform special management of the work order analysis results according to corresponding laws and regulations (i.e., target business process information), so as to realize the execution of the corresponding business process according to the target business process information corresponding to the work order analysis results, thereby avoiding the preset abnormal type work orders from squeezing out the public resources enjoyed by ordinary work orders.

[0090] Alternatively, the server obtains the target business process information corresponding to the work order analysis result, and calls the interface corresponding to the target business process information to execute the business process corresponding to the target business process information. This improves the recognition accuracy of the optimized decision tree work order recognition model, as well as the automation and efficiency of business execution, thereby improving the efficiency of abnormal work order recognition and processing.

[0091] In the embodiment of the present invention, the target abnormal work order field selection and data cleaning of the historical work order information table are performed to train the target decision tree work order recognition model, and the target abnormal work order field analysis is performed on the work order information to be processed through the target decision tree work order recognition model, and the business process corresponding to the target business process information is executed, which can flexibly perform correlation analysis and analysis of the rules of multiple attribute composites on the work order information to be processed, adapt to flexible and changeable work order complaint reporting strategies, improve the universality and accuracy of the target decision tree work order recognition model, and thus improve the efficiency of abnormal work order recognition and processing. This solution can be applied to the field of smart government affairs, thereby promoting the construction of smart cities.

[0092] See also Figure 2 Another embodiment of the method for processing work order information in the embodiment of the present invention includes:

[0093] 201. Obtain a historical work order information table, create a full field database for the historical work order information table, identify preset type work orders, and extract feature fields to obtain associated feature fields.

[0094] Specifically, the server obtains a historical work order information table, associates the historical work order information table, and obtains a full field database; matches the full field database based on the preset type work order field to obtain historical preset type work order data; reads historical field data in the historical preset type work order data, performs correlation analysis and feature field extraction on the historical field data in turn, and obtains associated feature fields.

[0095] The server extracts a historical work order information table from the target city's historical database, where the historical work order information table is a work order information table of historical work orders. The historical work order information table includes an information piece table, an information provider table, a subject-involved table, and an object-involved table. The server calculates the Cartesian product between the information piece table and the information provider table, the subject-involved table, and the object-involved table, respectively, through a preset internal association algorithm. According to the Cartesian product, the information provider table, the subject-involved table, and the object-involved table are merged into the information piece table to obtain a full field database, where the target city's historical database may include historical work order data of one or more cities, and the target city's historical database is the city's historical database selected for training the initial decision tree work order recognition model.

[0096] The server can classify the field data in the full field database according to the preset work order field feature types and classifiers (classification networks) to obtain historical preset type work order data, such as historical preset type work order data for historical advertising complaint and reporting work order data, i.e. historical advertising professional anti-counterfeiting complaint and reporting work orders. Preset work order field feature types, for example: the work order field feature types for food include production date features, food additive information features and other features, and the work order field feature types for advertising include false propaganda features, advertising law citation features and other features. The server reads historical field data in historical preset type work order data, performs correlation analysis on the historical field data through preset experience rules, thereby obtaining corresponding valuable associated fields, and extracts valuable associated fields in the historical field data to obtain associated feature fields, wherein the preset experience rules are strategies for identifying and processing preset abnormal type work order data, and the preset abnormal type work order data include but are not limited to data abnormality work order data, occupational complaint work order data and / or other types of abnormal work order data. In this embodiment, the preset abnormal type work order data is preferably preset type occupational complaint work order data (such as advertising type occupational counterfeiting complaint and reporting work order data). For example: taking the preset abnormal type work order data as advertising type occupational counterfeiting complaint and reporting work order data as an example, the preset experience rules are judgment conditions for advertising type occupational counterfeiting complaint and reporting work order data.

[0097] Specifically, the server reads the historical field data in the historical preset type work order data and the abnormal negative sample list data in the full field database; verifies the historical field data through preset experience rules to obtain verification field data; performs variable correlation analysis and feature field extraction on the verification field data through preset correlation rules and abnormal negative sample list data to obtain associated feature fields.

[0098] The server reads the historical field data in the historical preset type work order data, and obtains the abnormal negative sample list data based on the historical preset type work order data in the full field database, such as: the historical negative sample list data of advertising professional counterfeiting work order data; through the preset experience rules, the historical field data is verified to obtain the verification field data. The preset experience rules are the strategies for identifying and processing the preset abnormal type work order data. For example, taking the preset abnormal type work order data as advertising professional counterfeiting complaint and reporting work order data as an example, the experience rule is that people with more complaints and reports are most likely professional complainants (corresponding to hot work orders). Through the preset visualization tool PowerBI, the historical field data is statistically analyzed on the average number of complaints and reports of applicants to obtain hot work orders and ordinary work orders, and the field data corresponding to the hot work orders are determined as the verification field data.

[0099] Among them, the preset correlation rules include the item-by-item analysis rules and the combined exploration rules of the preset abnormal type work order data. The item-by-item analysis rules are correlation analysis under a single variable. For example, taking the preset abnormal type work order data as the advertising-related professional anti-counterfeiting complaint and reporting work order data as an example, the item-by-item analysis rules are correlation analysis of the probability distribution (as a single variable) between the pure complained enterprise type and the number of professional complaints. The combined exploration rules are correlation analysis of multiple variables. For example, taking the preset abnormal type work order data as the advertising-related professional anti-counterfeiting complaint and reporting work order data as an example, in the advertising-related professional anti-counterfeiting complaint and reporting work order data in October, the proportion of professional complaint work orders is higher than the proportion of professional complaint work orders in the advertising-related professional anti-counterfeiting complaint and reporting work order data in September. The combined exploration rules are correlation analysis of multiple variables, namely, the month (September and October) and the proportion of professional complaint work orders.

[0100] The server uses the preset correlation rules and abnormal negative sample list data to perform a correlation analysis on the distribution of each field in the verification field data and the preset abnormal type work order (such as advertising professional counterfeiting complaint and reporting work order), that is, to determine the correspondence and correlation between each field in the verification field data and the preset abnormal type work order, so as to obtain the corresponding extractable feature fields, extract the extractable feature fields in the verification field data, and obtain the associated feature fields. For example, the associated feature fields are: IS H0T NFORM-Whether it is a hot spot reporter, NFOWARED-information piece D, PERSNAME-applicant's name, MOBLE-mobile phone, LANDTEL-landline, EMAIL-email box, ADDR-address, NVNAME-name of the object involved, NFTYPE-information piece type, NVOAM-amount involved / consumption amount, APPLBASQUE-basic problem code, TRADETYPE-industry type, REMSHOTYPE-remote shopping type, NVOBJTYPE-involved object type, REGTME-registration time, QUANTITY-quantity, EBUSINESS-whether it involves e-commerce / whether it is off-site shopping, DIVISION-involved subject jurisdiction, NCFORM-receiving method code, CONTENT / CONTEN-Internet original text.

[0101] 202. Select target abnormal work order fields for associated characteristic fields and perform data cleaning based on a preset data storage strategy to obtain data to be processed.

[0102] Specifically, the server performs correlation analysis on the associated feature fields through the preset abnormal work order dependent variables to obtain the correlation analysis results; determines whether the correlation analysis results are randomly distributed. If the correlation analysis results are not randomly distributed, the corresponding associated feature fields are determined as the field data to be cleaned; according to the preset data warehousing strategy, the field data to be cleaned is deduplicated, the field meaning is converted, the data format is converted and normalized to obtain the data to be processed.

[0103] The server performs statistical counting on the associated feature fields to obtain field count data, and uses the preset Kendall rank correlation coefficient algorithm to perform correlation analysis and statistical analysis on the associated feature fields and the preset abnormal work order dependent variables to obtain correlation analysis results. The correlation analysis results are statistical analysis display results of the correlation analysis data, and determine whether the correlation analysis results are randomly distributed. If so, stop execution; if not, determine the corresponding associated feature fields as field data to be cleaned. The server merges or deletes duplicate data items in the field data to be cleaned according to the preset data entry strategy to obtain deduplicated data, standardizes the field meaning of the deduplicated data to obtain standardized data, parses the unstructured data in the standardized data into structured data to obtain converted data, and performs function conversion on the converted data through the preset atan function or log function to obtain the data to be processed. By deduplicating, converting the field meaning, converting the data format and normalizing the field data to be cleaned, the influence of the span of the value range of multiple variables on the weight of the preset initial decision tree work order recognition model is eliminated, thereby improving the training accuracy of the preset initial decision tree work order recognition model.

[0104] 203. Train the preset initial decision tree work order recognition model using the data to be processed to obtain a target decision tree work order recognition model.

[0105] The execution process of step 203 is similar to that of step 102, and will not be repeated here.

[0106] 204. Obtain the work order information to be processed, and through the target decision tree work order recognition model, perform data preprocessing based on preset type work order fields, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on the target feature fields on the work order information to be processed, and obtain the work order analysis results.

[0107] Specifically, the server obtains the information of the work order to be processed, and through the target decision tree work order recognition model, performs preset type work order field matching, field meaning conversion and normalization processing on the work order information to be processed in turn to obtain the preset type work order data to be processed; performs associated feature field extraction and statistical analysis on the preset type work order data to obtain associated analysis statistical data and associated fields of the work order to be processed; through the associated analysis statistical data, performs judgment analysis based on random distribution on the associated fields of the work order to be processed to obtain the target feature field; matches the target feature field with the abnormal work order field of the preset type work order field to obtain the work order analysis result.

[0108] For example, the server obtains the pending work order information of the landing city, and matches the pending work order information through the preset preset type work order field decision tree (the preset type work order is such as the advertising complaint and reporting work order). The target decision tree work order recognition model includes the preset preset type work order field decision tree. The preset type work order field decision tree includes the preset type work order field. The work order information to be processed is matched to obtain the initial preset type work order data. The initial preset type work order data is deduplicated, the field meaning is converted, the data format is converted and normalized to obtain the pending preset type work order data. According to the preset dependent variable, the preset type work order data to be processed is subjected to correlation analysis and statistics to obtain correlation analysis statistical data, and the correlation feature fields obtained from the correlation analysis are extracted to obtain. Go to the field associated with the pending work order, and determine whether the statistical data of the associated analysis is randomly distributed. If so, stop executing. If not, determine the corresponding field associated with the pending work order as the target feature field (such as an advertising field). According to the target feature field, traverse and match the preset type work order field decision tree in the target decision tree work order recognition model to obtain the matching field, and determine whether there is an abnormal work order field in the matching field. If so, return the return value representing the abnormal work order. If not, return the return value representing the normal work order, so as to obtain the work order analysis result. The preset type work order field decision tree includes the preset type work order field and the abnormal work order field corresponding to the preset type work order field, and the target decision tree work order recognition model includes the preset type work order field decision tree.

[0109] 205. Obtain target business process information corresponding to the work order analysis result, and execute the business process corresponding to the target business process information.

[0110] Specifically, the server receives a business execution request, parses the business execution request and extracts key information to obtain target business process information corresponding to the work order analysis result; based on the target business process information, the server traverses the preset business execution robot decision tree to obtain the target business execution robot; and calls the target business execution robot to execute the business process corresponding to the target business process information.

[0111] The server receives a business execution request, parses the business execution request to obtain business execution information, wherein the business execution information includes the business execution requirements corresponding to abnormal work orders when the work order analysis result is a work order, and the business execution requirements corresponding to work orders when the work order analysis result is not an abnormal work order. The server matches the corresponding target business execution requirements according to the work order analysis results. According to the target business execution requirements, based on the index, the server calls a preset search engine to search the preset business process module library to obtain the corresponding multiple initial business process modules, and splices the multiple initial business process modules according to the target business execution requirements to obtain the target business process information corresponding to the work order analysis results. According to the target business process information and the preset type field data, the server traverses the preset business execution robot decision tree to obtain the corresponding target business execution robot and the target robot service address of the target business execution robot, and calls the target robot service address through the target robot service address. The target business execution robot executes the business process corresponding to the target business process information. For example, taking the preset type field data as the advertising field data as an example, if the work order analysis result is an abnormal work order, the advertising complaint reporting work order corresponding to the advertising field data is: Zhang San complained to the responsible unit about XX Co., Ltd. at 9:00 on December 12, 2020, believing that the massage chair provided by the company had quality defects and demanded compensation of XX yuan. After the server obtains the target business process information, it obtains the target business execution robot (the type of the target robot is task-type) and the target business execution robot's target robot service address based on the target business process information and the preset type field data. The server calls the target robot according to the target service address to obtain the corresponding business process node from the target business process information, and jumps the current operation page to the operation page corresponding to the business process node.

[0112] The recognition accuracy of the optimized decision tree work order recognition model, as well as the automation and efficiency of business execution, are improved, thereby improving the efficiency of abnormal work order recognition and processing.

[0113] In the embodiment of the present invention, the target abnormal work order field selection and data cleaning of the historical work order information table are performed to train the target decision tree work order recognition model, and the target abnormal work order field analysis is performed on the work order information to be processed through the target decision tree work order recognition model, and the business process corresponding to the target business process information is executed, which can flexibly perform correlation analysis and analysis of the rules of multiple attribute composites on the work order information to be processed, adapt to flexible and changeable work order complaint reporting strategies, improve the universality and accuracy of the target decision tree work order recognition model, and thus improve the efficiency of abnormal work order recognition and processing. This solution can be applied to the field of smart government affairs, thereby promoting the construction of smart cities.

[0114] The above describes the method for processing work order information in the embodiment of the present invention. The following describes the device for processing work order information in the embodiment of the present invention. Figure 3 , an embodiment of a device for processing work order information in an embodiment of the present invention includes:

[0115] The acquisition and cleaning module 301 is used to acquire a historical work order information table, select target abnormal work order fields and clean the data in the historical work order information table to obtain the data to be processed. The historical work order information table includes an information piece table, an information provider table, a subject table and an object table;

[0116] The training module 302 is used to train the preset initial decision tree work order recognition model through the data to be processed to obtain the target decision tree work order recognition model;

[0117] The analysis module 303 is used to obtain the work order information to be processed, and through the target decision tree work order recognition model, sequentially perform data preprocessing based on the preset type of work order field, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on the target feature field on the work order information to be processed, to obtain the work order analysis result;

[0118] The execution module 304 is used to obtain target business process information corresponding to the work order analysis result and execute the business process corresponding to the target business process information.

[0119] The functional implementation of each module in the above-mentioned work order information processing device corresponds to each step in the above-mentioned work order information processing method embodiment, and its functions and implementation processes will not be repeated here one by one.

[0120] In the embodiment of the present invention, the target abnormal work order field selection and data cleaning of the historical work order information table are performed to train the target decision tree work order recognition model, and the target abnormal work order field analysis is performed on the work order information to be processed through the target decision tree work order recognition model, and the business process corresponding to the target business process information is executed, which can flexibly perform correlation analysis and analysis of the rules of multiple attribute composites on the work order information to be processed, adapt to flexible and changeable work order complaint reporting strategies, improve the universality and accuracy of the target decision tree work order recognition model, and thus improve the efficiency of abnormal work order recognition and processing. This solution can be applied to the field of smart government affairs, thereby promoting the construction of smart cities.

[0121] See also Figure 4 Another embodiment of the device for processing work order information in the embodiment of the present invention includes:

[0122] The acquisition and cleaning module 301 is used to acquire a historical work order information table, select target abnormal work order fields and clean the data in the historical work order information table to obtain the data to be processed. The historical work order information table includes an information piece table, an information provider table, a subject table and an object table;

[0123] Wherein, obtaining the cleaning module 301 specifically includes:

[0124] An extraction submodule 3011 is created to obtain a historical work order information table, create a full field database for the historical work order information table, identify preset type work orders, and extract feature fields to obtain associated feature fields;

[0125] The cleaning submodule 3012 is used to select target abnormal work order fields for the associated characteristic fields and clean the data based on the preset data storage strategy to obtain the data to be processed;

[0126] The training module 302 is used to train the preset initial decision tree work order recognition model through the data to be processed to obtain the target decision tree work order recognition model;

[0127] The analysis module 303 is used to obtain the work order information to be processed, and through the target decision tree work order recognition model, sequentially perform data preprocessing based on the preset type of work order field, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on the target feature field on the work order information to be processed, to obtain the work order analysis result;

[0128] The execution module 304 is used to obtain target business process information corresponding to the work order analysis result and execute the business process corresponding to the target business process information.

[0129] Optionally, creating the extraction submodule 3011 includes:

[0130] The association unit 30111 is used to obtain the historical work order information table, associate the historical work order information table with the table, and obtain a full field database;

[0131] The matching unit 30112 is used to match the preset type work order field on the full field database to obtain the historical preset type work order data;

[0132] The reading and analysis unit 30113 is used to read the historical field data in the historical preset type work order data, perform correlation analysis and feature field extraction on the historical field data in sequence, and obtain the associated feature fields.

[0133] Optionally, the reading and analyzing unit 30113 may also be specifically used for:

[0134] Read the historical field data in the historical preset type work order data, and the abnormal negative sample list data in the full field database;

[0135] Verify historical field data through preset empirical rules to obtain verification field data;

[0136] Through the preset correlation rules and abnormal negative sample list data, variable correlation analysis and feature field extraction are performed on the verification field data to obtain the associated feature fields.

[0137] Optionally, the cleaning submodule 3012 may also be specifically used for:

[0138] Through the preset abnormal work order dependent variables, correlation analysis is performed on the associated feature fields to obtain the correlation analysis results;

[0139] Determine whether the correlation analysis result is randomly distributed. If the correlation analysis result is not randomly distributed, determine the corresponding correlation feature field as the field data to be cleaned;

[0140] According to the preset data warehousing strategy, the field data to be cleaned is deduplicated, the field meaning is converted, the data format is converted and normalized to obtain the data to be processed.

[0141] Optionally, the analysis module 303 may also be specifically used for:

[0142] Obtain the work order information to be processed, and use the target decision tree work order recognition model to perform preset type work order field matching, field meaning conversion and normalization processing on the work order information to be processed, so as to obtain the preset type work order data to be processed;

[0143] Extract and perform statistical analysis on the associated feature fields of the preset work order data to be processed, and obtain the associated analysis statistical data and the associated fields of the work order to be processed;

[0144] Through the statistical data of association analysis, the judgment analysis based on random distribution is performed on the fields associated with the work orders to be processed to obtain the target feature fields;

[0145] The target feature field is matched with the abnormal work order field of the preset type work order field to obtain the work order analysis result.

[0146] Optionally, the execution module 304 may also be specifically configured to:

[0147] Receive business execution requests, parse the business execution requests and extract key information to obtain target business process information corresponding to the work order analysis results;

[0148] According to the target business process information, the preset business execution robot decision tree is traversed to obtain the target business execution robot;

[0149] Call the target business execution robot to execute the business process corresponding to the target business process information.

[0150] The functional implementation of each module and each unit in the above-mentioned work order information processing device corresponds to each step in the above-mentioned work order information processing method embodiment, and its functions and implementation processes are no longer repeated here.

[0151] In the embodiment of the present invention, the target abnormal work order field selection and data cleaning of the historical work order information table are performed to train the target decision tree work order recognition model, and the target abnormal work order field analysis is performed on the work order information to be processed through the target decision tree work order recognition model, and the business process corresponding to the target business process information is executed, which can flexibly perform correlation analysis and analysis of the rules of multiple attribute composites on the work order information to be processed, adapt to flexible and changeable work order complaint reporting strategies, improve the universality and accuracy of the target decision tree work order recognition model, and thus improve the efficiency of abnormal work order recognition and processing. This solution can be applied to the field of smart government affairs, thereby promoting the construction of smart cities.

[0152] above Figure 3 and Figure 4 The work order information processing device in the embodiment of the present invention is described in detail from the perspective of modular functional entities, and the work order information processing device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0153] Figure 5 1 is a schematic diagram of the structure of a work order information processing device provided by an embodiment of the present invention. The work order information processing device 500 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 510 (for example, one or more processors) and a memory 520, and one or more storage media 530 (for example, one or more mass storage devices) storing application programs 533 or data 532. Among them, the memory 520 and the storage medium 530 can be short-term storage or permanent storage. The program stored in the storage medium 530 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the work order information processing device 500. Furthermore, the processor 510 may be configured to communicate with the storage medium 530 to execute a series of instruction operations in the storage medium 530 on the work order information processing device 500.

[0154] The work order information processing device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 5 The structure of the work order information processing device shown does not constitute a limitation on the work order information processing device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0155] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the method for processing work order information.

[0156] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0157] The blockchain referred to in this invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0158] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0159] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0160] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing work order information, characterized in that: The method for processing the work order information includes: Obtain a historical work order information table, select target abnormal work order fields and perform data cleaning on the historical work order information table to obtain data to be processed, wherein the historical work order information table includes an information piece table, an information provider table, a subject table, and an object table; The data to be processed are randomly allocated to obtain a training data set and a test data set; a preset initial decision tree work order recognition model is trained by the training data set to obtain a trained decision tree work order recognition model; a judgment analysis based on preset abnormal type work orders is performed on the test data set by the trained decision tree work order recognition model to obtain a test result, and a target loss function value of the test result is calculated, wherein the execution process of the judgment analysis is as follows: the training decision tree work order recognition model starts from the root of the tree, and continuously selects the attribute with the largest information gain as the discrimination condition, and recursively continues in this way until the purity of the leaf node reaches the maximum, and the test result is obtained, wherein the test result includes the fields and field values ​​corresponding to "yes" or "no", respectively, and the test comprehensive central vector of all data in the test data set is calculated by the preset central vector formula. Quantity, and a single center vector of each data in the test data set, and calculate the absolute value of the difference between the single center vector and the test comprehensive center vector to obtain a vector loss function value, calculate the classification loss function value of the test result through a preset regression loss function, and perform weighted summation on the vector loss function value and the classification loss function value to obtain a target loss function value, wherein the target loss function value includes the vector loss function value and the classification loss function value; determine whether the target loss function value is less than a preset threshold value, and if the target loss function value is less than the preset threshold value, determine the training decision tree work order recognition model as a candidate decision tree work order recognition model; obtain project sedimentation historical data, and train, tune and prune the candidate decision tree work order recognition model in turn according to the project sedimentation historical data to obtain a target decision tree work order recognition model; Obtaining work order information to be processed, and using the target decision tree work order recognition model, sequentially performing data preprocessing based on preset type work order fields, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on target feature fields on the work order information to be processed, to obtain work order analysis results; Among them, the work order information to be processed is obtained, and through the target decision tree work order recognition model, the work order information to be processed is sequentially subjected to data preprocessing based on preset type work order fields, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on target feature fields, to obtain work order analysis results, including: Acquire the work order information to be processed, and use the target decision tree work order recognition model to sequentially perform preset type work order field matching, field meaning conversion and normalization processing on the work order information to be processed, so as to obtain the preset type work order data to be processed; Extract and statistically analyze the associated feature fields of the preset type of work order data to be processed to obtain associated analysis statistical data and associated fields of the work order to be processed; Through the association analysis statistical data, a judgment analysis based on random distribution is performed on the association fields of the to-be-processed work orders to obtain target feature fields; Matching the target feature field with the abnormal work order field of the preset type of work order field to obtain a work order analysis result; Obtain target business process information corresponding to the work order analysis result, and execute the business process corresponding to the target business process information.

2. The method for processing work order information according to claim 1, characterized in that: The step of obtaining a historical work order information table, selecting target abnormal work order fields and performing data cleaning on the historical work order information table to obtain data to be processed includes: Obtain a historical work order information table, create a full field database, identify preset type work orders, and extract feature fields for the historical work order information table to obtain associated feature fields; The target abnormal work order field selection and data cleaning based on the preset data storage strategy are performed on the associated characteristic fields to obtain the data to be processed.

3. The method for processing work order information according to claim 2, characterized in that: The obtaining of the historical work order information table, creating a full field database, identifying preset type work orders and extracting characteristic fields for the historical work order information table, and obtaining associated characteristic fields include: Obtain a historical work order information table, associate the historical work order information table with the table, and obtain a full field database; Matching the preset type work order fields on the full field database to obtain historical preset type work order data; The historical field data in the historical preset type work order data is read, and correlation analysis and feature field extraction are performed on the historical field data in sequence to obtain correlation feature fields.

4. The method for processing work order information according to claim 3, characterized in that: The reading of the historical field data in the historical preset type work order data, sequentially performing correlation analysis and feature field extraction on the historical field data to obtain the associated feature fields, includes: Read the historical field data in the historical preset type work order data and the abnormal negative sample list data in the full field database; Verifying the historical field data by using preset empirical rules to obtain verification field data; By using preset correlation rules and the abnormal negative sample list data, variable correlation analysis and feature field extraction are performed on the verification field data to obtain correlation feature fields.

5. The method for processing work order information according to claim 2, characterized in that: The step of selecting target abnormal work order fields from the associated characteristic fields and performing data cleaning based on a preset data storage strategy to obtain data to be processed includes: By using the preset abnormal work order dependent variable, a correlation analysis is performed on the associated characteristic field to obtain a correlation analysis result; Determine whether the correlation analysis result is randomly distributed, and if the correlation analysis result is not randomly distributed, determine the corresponding correlation feature field as the field data to be cleaned; According to the preset data storage strategy, the field data to be cleaned is deduplicated, the field meaning is converted, the data format is converted and normalized to obtain the data to be processed.

6. The method for processing work order information according to any one of claims 1 to 5, characterized in that: The acquiring target business process information corresponding to the work order analysis result and executing the business process corresponding to the target business process information includes: Receiving a business execution request, parsing the business execution request and extracting key information, and obtaining target business process information corresponding to the work order analysis result; According to the target business process information, traverse the preset business execution robot decision tree to obtain the target business execution robot; The target business execution robot is called to execute the business process corresponding to the target business process information.

7. A device for processing work order information, characterized in that: The work order information processing device comprises: An acquisition and cleaning module is used to acquire a historical work order information table, select target abnormal work order fields and perform data cleaning on the historical work order information table to obtain data to be processed, wherein the historical work order information table includes an information piece table, an information provider table, a subject table involved, and an object table involved; The training module is used to randomly distribute the data to be processed to obtain a training data set and a test data set; the preset initial decision tree work order recognition model is trained by the training data set to obtain a training decision tree work order recognition model; the test data set is subjected to judgment analysis based on preset abnormal type work orders by the training decision tree work order recognition model to obtain a test result, and the target loss function value of the test result is calculated, wherein the execution process of the judgment analysis is as follows: the training decision tree work order recognition model starts from the root of the tree, and continuously selects the attribute with the largest information gain as the discrimination condition, and recursively continues in this way until the purity of the leaf node reaches the maximum, and the test result is obtained, wherein the test result includes the fields and field values ​​corresponding to "yes" or "no", respectively, and the test summary of all data in the test data set is calculated by the preset central vector formula. The combined central vector and the single central vector of each data in the test data set are obtained, and the absolute value of the difference between the single central vector and the test comprehensive central vector is calculated to obtain a vector loss function value. The classification loss function value of the test result is calculated through a preset regression loss function, and the vector loss function value and the classification loss function value are weighted and summed to obtain a target loss function value, wherein the target loss function value includes the vector loss function value and the classification loss function value; it is determined whether the target loss function value is less than a preset threshold value, and if the target loss function value is less than the preset threshold value, the training decision tree work order recognition model is determined as a candidate decision tree work order recognition model; the project sedimentation historical data is obtained, and according to the project sedimentation historical data, the candidate decision tree work order recognition model is sequentially trained, tuned and pruned to obtain a target decision tree work order recognition model; The analysis module is used to obtain the work order information to be processed, and through the target decision tree work order recognition model, sequentially perform data preprocessing based on the preset type of work order field, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on the target feature field on the work order information to be processed to obtain the work order analysis result; wherein, the work order information to be processed is obtained, and through the target decision tree work order recognition model, the work order information to be processed is sequentially performed data preprocessing based on the preset type of work order field, target feature field extraction based on association analysis statistics, and abnormal work order field judgment analysis based on the target feature field to obtain the work order analysis result. The result comprises: obtaining work order information to be processed, and performing preset type work order field matching, field meaning conversion and normalization processing on the work order information to be processed in sequence through the target decision tree work order recognition model to obtain preset type work order data to be processed; performing associated feature field extraction and statistical analysis on the preset type work order data to be processed to obtain associated analysis statistical data and associated fields of the work order to be processed; performing random distribution-based judgment analysis on the associated fields of the work order to be processed through the associated analysis statistical data to obtain target feature fields; matching the target feature fields with abnormal work order fields of the preset type work order fields to obtain work order analysis results; The execution module is used to obtain the target business process information corresponding to the work order analysis result and execute the business process corresponding to the target business process information.

8. A work order information processing device, characterized in that: The work order information processing device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instruction in the memory so that the work order information processing device executes the work order information processing method according to any one of claims 1 to 6.

9. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by the processor, the method for processing work order information as described in any one of claims 1 to 6 is implemented.

Images