A business model construction method and system based on low-code modularization, electronic device and computer-readable storage medium

Abstract

The present invention relates to a method for building a business model based on low-code modularization, which forms a business rule module by extracting the complete algorithm post-processing process running in each business, and splitting the business model into a model composed of a basic algorithm and a business rule module. The basic algorithm in the model performs basic analysis and identification on the data input to the business model, and the business rule module performs processing operations such as filtering, screening, and conversion on the analysis results of the basic algorithm, and finally obtains the business expected results of the business model. By selecting and combining different basic algorithms and different business rule modules, a new business model that meets the business scenario can be flexibly and quickly formed. This way of building a business model improves the applicability and reuse rate of the algorithm, reduces the participation of professional developers, lowers the threshold for algorithm use, reduces the workload of business model development, and improves the speed of algorithm delivery.

Description

Technical Field

[0001] The present invention relates to the field of computer business modeling technology, and in particular to a business model construction method and system based on low-code modularization, an electronic device and a computer-readable storage medium. Background Art

[0002] With the development of deep learning neural network technology, it has been widely used in finance, healthcare, transportation, security, manufacturing and other industries. It has achieved remarkable results in computer vision, natural language processing, speech recognition, data mining, etc., especially its advantages in segmented data processing such as image recognition, image classification, target detection, text analysis, and speech recognition.

[0003] Currently, the same deep learning neural network algorithm can be applied to multiple business scenarios after performing different post-processing on the algorithm's identification results. However, the business scenarios in different fields have significantly different requirements. If corresponding business rules were developed for each business scenario, the workload of developing business models would be large and inefficient. Furthermore, deep learning neural network algorithms require programming by professional algorithm engineers, which makes it difficult to quickly build new neural network algorithms suitable for current business needs.

[0004] Chinese patent CN107977706B discloses a modular distributed artificial neural network. The neural network module of this method does not adapt to business scenario requirements, is not suitable for general AI inference equipment, and cannot personalize the algorithm application conditions. Summary of the Invention

[0005] Based on this, the purpose of the present invention is to provide a business model construction method based on low-code modularization. This method extracts the post-processing process of the neural network basic algorithm to form a business rule module, which can flexibly configure data processing rules that meet the business scenarios for each neural network basic algorithm, reduce workload and improve algorithm delivery speed.

[0006] A method for building a business model based on low-code modularization includes the following steps:

[0007] S10 obtains algorithm information of the deep learning algorithm module selected and set by the user, and generates a basic algorithm configuration file according to the algorithm information;

[0008] S20 completes basic algorithm code generation according to the basic algorithm configuration file, and records basic information of the basic algorithm as well as the basic algorithm configuration file path and the basic algorithm code file path;

[0009] S30 obtains the business rule information of the business rule module selected and set by the user, and the connection relationship between the business rule module and the deep learning algorithm module, and generates a business model configuration file based on the business rule information, the connection relationship and the basic algorithm configuration file;

[0010] S40 completes the business model code generation according to the business model configuration file, and records the basic information of the business model as well as the business model configuration file path and the business model code file path.

[0011] Compared with the existing technology, the present invention extracts the complete algorithm post-processing process running in each business to form a business rule module, and splits the business model into a model composed of a basic algorithm and a business rule module. The basic algorithm in the model performs basic analysis and identification on the data input to the business model, and the business rule module performs filtering, screening, conversion and other processing operations on the analysis results of the basic algorithm, and finally obtains the business expected results of the business model. By selecting and combining different basic algorithms and different business rule modules, a new business model that meets the business scenario can be flexibly and quickly formed. This way of building a business model improves the applicability and reuse rate of the algorithm, reduces the participation of professional developers, lowers the threshold for algorithm use, reduces the workload of business model development, and improves the speed of algorithm delivery.

[0012] Furthermore, the deep learning algorithm module includes a data source submodule, a feature extraction submodule and a classifier submodule:

[0013] The data source submodule is used to pre-process the input raw data of the business model established by the user;

[0014] The feature extraction submodule is used to perform feature extraction operations on the preprocessed data to obtain high-dimensional features of the data. The feature extraction operations include pooling, convolution and attention mechanisms;

[0015] The classifier submodule is used to analyze and process the high-dimensional features of the data to obtain analysis results of the data.

[0016] Furthermore, the basic algorithm configuration file is used to reproduce the algorithm information of the deep learning algorithm module selected and set by the user, and the algorithm information includes:

[0017] The composition information of the deep learning algorithm module: the module name of each submodule;

[0018] Configuration information of the deep learning algorithm module: setting parameters of each submodule;

[0019] Basic information of deep learning algorithms: algorithm category, algorithm name, and algorithm description selected in each submodule;

[0020] Structural information of deep learning algorithm modules: connection relationships between submodules;

[0021] Global configuration information of the deep learning algorithm: the initial module and entry function of the global configuration composed of each submodule.

[0022] Furthermore, step S20 includes the following sub-steps:

[0023] S21 imports the classes corresponding to each submodule according to the composition information of the deep learning algorithm module, and generates an instance for the class of each submodule;

[0024] S22 initializes the class instances of each submodule according to the configuration information of the deep learning algorithm module;

[0025] S23 calls the pre-module of the class instance of each sub-module according to the structural information of the deep learning algorithm module, and sets the class data flow function;

[0026] S24 sets the initial module and entry function for the algorithm runtime based on the global configuration information of the deep learning algorithm, and completes the basic algorithm code generation;

[0027] S25 records the basic information of the basic algorithm as well as the basic algorithm configuration file path and the basic algorithm code file path.

[0028] Furthermore, the business model configuration file in step S30 is generated in the following manner:

[0029] Query the basic algorithm configuration file according to the algorithm name, add the parameter weight version of the deep learning algorithm module based on the basic algorithm configuration file in json format, and integrate the business rule information of the business rule module and the connection relationship between the business rule module and the deep learning algorithm module to generate a new json file as the configuration file of the business model.

[0030] Furthermore, the step S40 includes the following sub-steps:

[0031] S41 imports the basic algorithm and specifies the parameter weight file loaded by the basic algorithm according to the parameter weight version of the deep learning algorithm module;

[0032] S42 imports the corresponding business rule module class according to the business rule module name;

[0033] S43 creates an instance of the business rule module class and initializes it according to the business rule module setting information;

[0034] S44 sets data input for an instance of the business rule module class according to the connection relationship between the business rule module and the deep learning algorithm module, completing business model code generation;

[0035] S45 records the basic information of the business model as well as the business model configuration file path and the business model code file path.

[0036] Furthermore, it also includes step S60 to obtain information on the user's instance creation and computing resource configuration for the constructed business model, generate web service code on the corresponding server based on the information, query the basic information of the business model based on the business model ID, import the business model code file, business model configuration file, and the parameter weight file specified in step S41, and then complete the web service startup.

[0037] At the same time, the present invention also provides a business model construction system based on low-code modularization.

[0038] The beneficial effects of the low-code modular business model construction system proposed in the present invention are the same as the above-mentioned low-code modular business model construction method, and will not be repeated here.

[0039] At the same time, the present invention also provides an electronic device, including:

[0040] one or more processors;

[0041] A memory for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement any of the above-mentioned methods for building a business model based on low-code modularization.

[0042] At the same time, the present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements any of the above-mentioned methods for building a business model based on low-code modularization.

[0043] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of an optional hardware architecture of the electronic device of the present invention;

[0045] Figure 2 A schematic diagram of a system for building a low-code modular business model in an electronic device of the present invention;

[0046] Figure 3 This is a flow chart of the method for building a business model based on low-code modularization of the present invention. DETAILED DESCRIPTION

[0047] The technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiments of the present invention.

[0048] In order to solve the problem that when building a business model based on the existing deep learning neural network basic algorithm, either the recognition results of the neural network basic algorithm are post-processed differently to suit the needs of different business scenarios, or corresponding business rules are developed for each business scenario to adapt to the construction method of the deep learning neural network basic algorithm, resulting in a large workload and low development efficiency of the business model, the present invention proposes a business model construction system based on low-code modularization. The construction system includes a construction method, which obtains a basic algorithm configuration file by obtaining the algorithm information of the deep learning algorithm module selected and set by the user, obtains the business rule information of the business rule module selected and set by the user, and the relationship between the business rule module and the deep learning algorithm module to obtain the business model configuration file, and generates a complete algorithm code for the business model through the basic algorithm configuration file and the business model configuration file. The present invention can flexibly and quickly form a new business model that meets the business scenario through the selection and combination of different basic algorithms and different business rule modules, realize the automatic generation of business models, improve the applicability and reuse rate of the algorithm, reduce the participation of professional developers, lower the threshold for algorithm use, reduce the workload of business model development, and improve the speed of algorithm delivery.

[0049] For details, please refer to Figure 1 and Figure 2 , Figure 1 is a schematic diagram of an optional hardware architecture of an electronic device including a low-code modular business model construction system proposed by the present invention, Figure 2 This is a schematic diagram of the architecture of the low-code modular business model construction system proposed in the present invention.

[0050] In this embodiment, the electronic device 10 includes but is not limited to a memory 11, a processor 12, a network interface 13, and a business model construction system 14 based on a low-code modularization that can be interconnected through a system bus.

[0051] The electronic device 10 may be a computing device such as a rack server, a blade server, a tower server, or a cabinet server, etc. The electronic device 10 may be an independent server or a server cluster composed of multiple servers.

[0052] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 10, such as a hard disk or memory of the electronic device 10. In other embodiments, the memory 11 can also be an external storage device of the electronic device 10, such as a plug-in hard disk equipped on the electronic device 10, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. The memory 11 can also include both the internal storage unit of the electronic device 10 and its external storage device. In this embodiment, the memory 11 is generally used to store the operating system and various application software installed on the electronic device 10, such as the program code of the low-code modular business model construction system 14. In addition, the memory 11 can also be used to temporarily store various types of data that have been output or are to be output.

[0053] In some embodiments, the processor 12 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 12 is generally used to control the overall operation of the electronic device 10, such as performing control and processing related to data interaction or communication with the electronic device 10. In this embodiment, the processor 12 is used to run the program code stored in the memory 11 or process data, such as running the low-code modular business model construction method.

[0054] The network interface 13 may include a wireless network interface or a wired network interface, and the network interface 13 is generally used to establish a communication connection between the electronic device 10 and other electronic devices. For example, the network interface 13 is used to connect the electronic device 10 to an external data platform via a network, and to establish a data transmission channel and a communication connection between the electronic device 10 and the external data platform. The network may be a wireless or wired network such as an intranet, the Internet, a global system of mobile communications (GSM), wideband code division multiple access (WCDMA), a 4G network, a 5G network, Bluetooth, or Wi-Fi.

[0055] Based on the above-mentioned hardware structure and functions, the low-code modular business model construction system 14 of this application is set on it. The construction system includes but is not limited to: an algorithm configuration acquisition unit 141, an algorithm creation unit 142, a business rule configuration acquisition unit 143 and a business model creation unit 144.

[0056] The algorithm configuration acquisition unit 141 is configured to execute step S10: acquire algorithm information of the deep learning algorithm module selected and set by the user, and generate a basic algorithm configuration file according to the algorithm information.

[0057] The deep learning algorithm module is composed of algorithms selected and set by the user from a deep learning algorithm library pre-stored in the low-code modular business model construction system. Specifically, the deep learning algorithm module includes a data source submodule 201, a feature extraction submodule 202, and a classifier submodule 203.

[0058] The data source submodule 201 is used to pre-process the input raw data of the business model established by the user.

[0059] The preprocessing includes loading, decoding, cleaning, etc. Specifically, the data is standardized to make it conform to the system's usage specifications; dirty data is cleaned to remove null values and outliers; and the data format is converted to the specific format used by the system.

[0060] The feature extraction submodule 202 is used to perform feature extraction on the preprocessed data to obtain high-dimensional features of the data. Feature extraction operations include pooling, convolution, and attention mechanisms. Feature extraction submodule 202 includes multiple existing feature extraction neural networks, such as convolutional neural networks, recurrent neural networks, graph convolutional networks, and transformers.

[0061] The classifier submodule 203 is used to analyze and process the high-dimensional features of the data to obtain data analysis results. The classifier submodule 203 includes multiple existing classification processing neural networks, such as CapsuleNetworks, MobileNets, EfficientNets, Deep Belief Networks, Deep Boltzmann Machines, etc.

[0062] The algorithm information includes:

[0063] The composition information of the deep learning algorithm module: the module name of each submodule;

[0064] Configuration information of the deep learning algorithm module: setting parameters of each submodule;

[0065] Basic information of deep learning algorithms: algorithm category, algorithm name, and algorithm description selected in each submodule;

[0066] Structural information of deep learning algorithm modules: connection relationships between submodules;

[0067] Global configuration information of the deep learning algorithm: the initial module and entry function of the global configuration composed of each submodule.

[0068] The basic algorithm configuration file stores the acquired configuration information in a json format file, and the basic algorithm configuration file is used to reproduce and reconstruct the deep learning algorithm selected and set by the user.

[0069] The algorithm creation unit 142 is configured to execute step S20: complete basic algorithm code generation according to the basic algorithm configuration file, and record basic information of the basic algorithm as well as the basic algorithm configuration file path and the basic algorithm code file path.

[0070] Specifically, the method includes the following sub-steps:

[0071] S21 imports the classes corresponding to each submodule according to the composition information of the deep learning algorithm module, and generates an instance for the class of each submodule;

[0072] S22 initializes the class instances of each submodule according to the configuration information of the deep learning algorithm module;

[0073] S23 calls the pre-module of the class instance of each sub-module according to the structural information of the deep learning algorithm module, and sets the class data flow function;

[0074] S24 sets the initial module and entry function for the algorithm runtime based on the global configuration information of the deep learning algorithm, and completes the basic algorithm code generation;

[0075] S25 records the basic information of the basic algorithm as well as the basic algorithm configuration file path and the basic algorithm code file path.

[0076] The business rule configuration acquisition unit 143 is used to execute step S30: obtain the business rule information of the business rule module selected and set by the user, and the connection relationship between the business rule module and the deep learning algorithm module, and generate a business model configuration file based on the business rule information, connection relationship and basic algorithm configuration file.

[0077] Specifically, the business model configuration file is generated in the following manner.

[0078] Query the basic algorithm configuration file according to the algorithm name, add the parameter weight version of the deep learning algorithm module based on the basic algorithm configuration file in json format, and integrate the business rule information of the business rule module and the connection relationship between the business rule module and the deep learning algorithm module to generate a new json file as the configuration file of the business model.

[0079] The business rule module includes but is not limited to warning label category, warning confidence, spatial filtering, and temporal filtering.

[0080] The business rule information includes the business rule module name and business rule setting information.

[0081] The connection relationship between the business rule module and the deep learning algorithm module is to establish a connection between the algorithm and the business rule module through a connection line, and the arrow of the connection line points to the business rule module, indicating that the data processing flows to the business rule module.

[0082] The business model creation unit 144 is configured to execute step S40: complete business model code generation according to the business model configuration file, and record basic information of the business model as well as the business model configuration file path and the business model code file path.

[0083] Specifically, the method includes the following sub-steps:

[0084] S41 imports the basic algorithm and specifies the parameter weight file loaded by the basic algorithm according to the parameter weight version of the deep learning algorithm module;

[0085] S42 imports the corresponding business rule module class according to the business rule module name;

[0086] S43 creates an instance of the business rule module class and initializes it according to the business rule module setting information;

[0087] S44 sets data input for an instance of the business rule module class according to the connection relationship between the business rule module and the deep learning algorithm module, completing business model code generation;

[0088] S45 records the basic information of the business model as well as the business model configuration file path and the business model code file path.

[0089] Furthermore, the low-code modularized business model construction system 14 also includes a web service unit 145.

[0090] The web service unit 145 is used to execute step S60: obtain information on instance creation and computing resource configuration of the constructed business model by the user, generate web service code on the corresponding server based on the information, query the basic information of the business model based on the business model ID, import the business model code file, business model configuration file, and the parameter weight file specified in step S41, and then complete the web service startup.

[0091] In the deep learning algorithm module, users use a visual drag-and-drop method to select different neural networks from the data source submodule, feature extraction submodule, and classifier submodule, and define the relationships between these neural networks to build an algorithm module. Next, users select different business scenario modules from the business rule module and configure them with the algorithm module to form a business model. When a business model instance is launched, the complete algorithm code is built and the algorithm web service is provided.

[0092] This application extracts the complete algorithm post-processing process running in each business to form a business rule module, and splits the business model into a model composed of a basic algorithm and a business rule module. The basic algorithm in the model performs basic analysis and identification on the data input to the business model, and the business rule module performs filtering, screening, conversion and other processing operations on the analysis results of the basic algorithm, and finally obtains the business expected results of the business model. By selecting and combining different basic algorithms and different business rule modules, a new business model that meets the business scenario can be formed flexibly and quickly. This way of building a business model improves the applicability and reuse rate of the algorithm, reduces the participation of professional developers, lowers the threshold for algorithm use, reduces the workload of business model development, and improves the speed of algorithm delivery.

[0093] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the embodiments of the present application. The singular forms of "a", "said" and "the" used in the embodiments of the present application and the claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that, unless otherwise specified, "multiple" and "several" refer to two or more; "and / or" refers to and includes any or all possible combinations of one or more associated listed items; "first", "second", "third" and the like are only used to distinguish, and are not used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. When the above description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of the present application, for those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0094] The above-described embodiments merely represent several implementations of the present invention. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous modifications and improvements without departing from the spirit of the present invention, and the present invention is intended to encompass such modifications and variations.

Claims

1. A method for building a business model based on low-code modularization, characterized in that: The following steps are involved: S10 obtains algorithm information of the deep learning algorithm module selected and set by the user, and generates a basic algorithm configuration file according to the algorithm information; S20 completes basic algorithm code generation according to the basic algorithm configuration file, and records basic information of the basic algorithm as well as the basic algorithm configuration file path and the basic algorithm code file path; S30 obtains the business rule information of the business rule module selected and set by the user, and the connection relationship between the business rule module and the deep learning algorithm module, and generates a business model configuration file based on the business rule information, the connection relationship and the basic algorithm configuration file; S40 completes the business model code generation according to the business model configuration file, and records the basic information of the business model as well as the business model configuration file path and the business model code file path.

2. The method for building a business model based on low-code modularization according to claim 1 is characterized in that: The deep learning algorithm module includes a data source submodule, a feature extraction submodule and a classifier submodule: The data source submodule is used to pre-process the input raw data of the business model established by the user; The feature extraction submodule is used to perform feature extraction operations on the preprocessed data to obtain high-dimensional features of the data. The feature extraction operations include pooling, convolution, and attention mechanisms. The classifier submodule is used to analyze and process the high-dimensional features of the data to obtain analysis results of the data.

3. The method for building a business model based on low-code modularization according to claim 2 is characterized in that: The basic algorithm configuration file is used to reproduce the algorithm information of the deep learning algorithm module selected and set by the user. The algorithm information includes: The composition information of the deep learning algorithm module: the module name of each submodule; Configuration information of the deep learning algorithm module: setting parameters of each submodule; Basic information of deep learning algorithms: algorithm category, algorithm name, and algorithm description selected in each submodule; Structural information of deep learning algorithm modules: connection relationships between submodules; Global configuration information of the deep learning algorithm: the initial module and entry function of the global configuration composed of each submodule.

4. The method for building a business model based on low-code modularization according to claim 3 is characterized in that: The step S20 includes the following sub-steps: S21 imports the classes corresponding to each submodule according to the composition information of the deep learning algorithm module, and generates an instance for the class of each submodule; S22 initializes the class instances of each submodule according to the configuration information of the deep learning algorithm module; S23 calls the pre-module of the class instance of each sub-module according to the structural information of the deep learning algorithm module, and sets the class data flow function; S24 sets the initial module and entry function for the algorithm runtime based on the global configuration information of the deep learning algorithm, and completes the basic algorithm code generation; S25 records the basic information of the basic algorithm as well as the basic algorithm configuration file path and the basic algorithm code file path.

5. The method for building a business model based on low-code modularization according to any one of claims 1 to 4, characterized in that: The business model configuration file in step S30 is generated in the following manner: Query the basic algorithm configuration file according to the algorithm name, add the parameter weight version of the deep learning algorithm module based on the basic algorithm configuration file in json format, and integrate the business rule information of the business rule module and the connection relationship between the business rule module and the deep learning algorithm module to generate a new json file as the configuration file of the business model.

6. The method for building a business model based on low-code modularization according to claim 5 is characterized in that: The step S40 includes the following sub-steps: S41 imports the basic algorithm and specifies the parameter weight file loaded by the basic algorithm according to the parameter weight version of the deep learning algorithm module; S42 imports the corresponding business rule module class according to the business rule module name; S43 creates an instance of the business rule module class and initializes it according to the business rule module setting information; S44 sets data input for an instance of the business rule module class according to the connection relationship between the business rule module and the deep learning algorithm module, completing business model code generation; S45 records the basic information of the business model as well as the business model configuration file path and the business model code file path.

7. The method for building a business model based on low-code modularization according to claim 6 is characterized in that: The step S60 is also included: S60 obtains the user's information on instance creation and computing resource configuration for the constructed business model, generates web service code on the corresponding server based on the information, queries the basic information of the business model based on the business model ID, imports the business model code file, business model configuration file, and the parameter weight file specified in step S41, and then completes the web service startup.

8. A business model construction system based on low-code modularization, characterized in that: The system includes: an algorithm configuration acquisition unit, an algorithm creation unit, a business rule configuration acquisition unit, and a business model creation unit: The algorithm configuration acquisition unit is used to obtain algorithm information of the deep learning algorithm module selected and set by the user, and generate a basic algorithm configuration file according to the algorithm information; The algorithm creation unit is used to complete the basic algorithm code generation according to the basic algorithm configuration file, and record the basic information of the basic algorithm as well as the basic algorithm configuration file path and the basic algorithm code file path; The business rule configuration acquisition unit is used to obtain the business rule information of the business rule module selected and set by the user, and the connection relationship between the business rule module and the deep learning algorithm module, and generate a business model configuration file based on the business rule information, connection relationship and basic algorithm configuration file; The business model creation unit is used to complete business model code generation according to the business model configuration file, and record basic information of the business model as well as the business model configuration file path and the business model code file path.

9. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the low-code modular business model construction method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the low-code modular business model construction method as described in any one of claims 1 to 7.

Images