Disclosure of Invention
The invention aims to provide a technical transaction recommendation system and method for AI big data, so as to solve the problems in the background technology.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a technical transaction recommendation method of AI big data, the method comprising:
receiving an equipment chain input by a user side, and creating a task table according to the equipment chain; the equipment chain comprises an equipment model and a logic connection relation thereof;
Extracting data characteristics of each task in the task table, and correcting the task table according to the data characteristics; the data features are used for representing the demand parameters of each task;
traversing a preset task library according to the corrected task table, and establishing a virtual task table;
Acquiring a solution table based on the virtual task table, debugging the equipment chain according to the solution table, and outputting a task flow;
Acquiring quotation information of a task flow, sending the quotation information to a user side, receiving selection information uploaded by the user side, and creating a transaction platform;
The device chain and the task table are encrypted by the user side, and the data in the answer table are encrypted by the answer side.
As a further scheme of the invention: the step of receiving the equipment chain input by the user side and creating the task list according to the equipment chain comprises the following steps:
receiving a production flow and an equipment model input by a user side, and determining a logic connection relation of the equipment model according to the production flow;
Acquiring a component framework of the equipment model, and determining a component connection relation according to the data transmission direction of the equipment model;
Establishing a flow chart consisting of nodes according to the logic connection relation and the component connection relation; the flow chart contains an identification frame for representing the membership of the component; the node and the component have a mapping relation;
based on the node statistics demand parameters input by the user side, creating a task table according to the demand parameters; the data structure of the demand parameter is a preset value.
As a further scheme of the invention: the step of extracting the data characteristics of each task in the task table and correcting the task table according to the data characteristics comprises the following steps:
sequentially extracting each task in a task table, and inputting the tasks into a preset conversion model to obtain a conversion value;
Calculating correlation coefficients among the conversion values, and determining the similarity of each task according to the correlation coefficients;
classifying each task in the task list according to the similarity;
In the classifying process, the original positions of all tasks are recorded.
As a further scheme of the invention: traversing a preset task library according to the corrected task table, and establishing a virtual task table comprises the following steps:
Calculating the task number of various tasks in the corrected task list, and determining a traversal level according to the task number; the traversal level is used for determining the traversal sequence and the traversal precision;
Extracting various tasks according to the traversal level, traversing a preset task library through the tasks, and determining a target task;
and counting the target tasks and establishing a virtual task table.
As a further scheme of the invention: the steps of extracting various tasks according to the traversal level, traversing a preset task library through the tasks, and determining target tasks comprise the following steps:
determining a preset number of task step sizes, and dividing a task library according to the task step sizes;
extracting various tasks according to the traversal level, sequentially extracting tasks in the tasks, taking the tasks as tasks to be detected, and calculating editing distances between the tasks to be detected and the segmentation tasks;
Comparing the editing distance with a preset editing threshold value, and marking a segmentation task according to the comparison result; the editing threshold is determined by the traversal accuracy;
Determining a target task according to the marking result of the segmentation task; the marking result comprises marking times and editing distances in each marking;
the calculation model of the editing distance is as follows:
Wherein dis (a, b) is the editing distance between the target task and the segmentation task, a is the character string representing the target task, and b is the character string representing the segmentation task; the |a| and the |b| are the character string lengths; tail (a) is a tail character string except the first character in the character string of the target task, and tail (b) is a tail character string except the first character in the character string of the segmentation task; a [0] is the first character of the character string of the target task, and b [0] is the first character of the character string of the segmentation task.
As a further scheme of the invention: the steps of obtaining a solution table based on the virtual task table, debugging the equipment chain according to the solution table and outputting a task flow comprise:
The virtual task table is sent to a solver, solution information fed back by the solver is obtained, and a solution table corresponding to the virtual task table is established;
inquiring the corresponding relation between the virtual task table and the original task table to obtain a solution table corresponding to the original task table, which is called an application solution table;
Debugging the equipment chain according to the application solution table, and outputting a task flow;
wherein the answer information of each virtual task is at least one.
As a further scheme of the invention: the step of obtaining quotation information of the task flow, sending the quotation information to a user side, receiving selection information uploaded by the user side and creating a transaction platform comprises the following steps:
Acquiring quotation information of each answer information in a task flow, and recording the application times of the answer information in an original task table;
determining price tables corresponding to different application answer tables according to quotation information and application times;
and sending the price list to the user side, and creating a transaction platform when receiving the selection information sent by the user side.
The technical scheme of the invention also provides a technical transaction recommendation system for AI big data, which comprises:
The task list creation module is used for receiving the equipment chain input by the user side and creating a task list according to the equipment chain; the equipment chain comprises an equipment model and a logic connection relation thereof;
The task table correction module is used for extracting the data characteristics of each task in the task table and correcting the task table according to the data characteristics; the data features are used for representing the demand parameters of each task;
the virtual table generation module is used for traversing a preset task library according to the corrected task table and establishing a virtual task table;
the task flow output module is used for acquiring a solution table based on the virtual task table, debugging the equipment chain according to the solution table and outputting a task flow;
the transaction platform creation module is used for acquiring quotation information of the task flow, sending the quotation information to the user side, receiving selection information uploaded by the user side and creating a transaction platform;
The device chain and the task table are encrypted by the user side, and the data in the answer table are encrypted by the answer side.
As a further scheme of the invention: the task table creation module includes:
The logic relation determining unit is used for receiving the production flow and the equipment model input by the user side and determining the logic connection relation of the equipment model according to the production flow;
the component relation determining unit is used for obtaining the component architecture of the equipment model and determining the component connection relation according to the data transmission direction of the equipment model;
The flow chart establishing unit is used for establishing a flow chart consisting of nodes according to the logic connection relation and the component connection relation; the flow chart contains an identification frame for representing the membership of the component; the node and the component have a mapping relation;
the creating execution unit is used for counting demand parameters input by a user side based on the nodes and creating a task table according to the demand parameters; the data structure of the demand parameter is a preset value.
As a further scheme of the invention: the task correction module includes:
The type conversion unit is used for sequentially extracting each task in the task table, and inputting the tasks into a preset conversion model to obtain conversion values;
The similarity calculation unit is used for calculating correlation coefficients among the conversion values and determining the similarity of each task according to the correlation coefficients;
the classification execution unit is used for classifying each task in the task list according to the similarity;
In the classifying process, the original positions of all tasks are recorded.
Compared with the prior art, the invention has the beneficial effects that: according to the method and the device, the task table is created according to the device chain uploaded by the user, the virtual task table is created according to the created task table, the virtual task table is sent to different answering parties, after the answering information is obtained, a code flow corresponding to the original task table is established, and the device chain is compiled, so that the original requirement of the user is hidden, and the risk probability of data leakage is greatly reduced.
Detailed Description
In order to make the technical problems, technical schemes and beneficial effects to be solved more clear, the invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Fig. 1 is a flow chart of a technical transaction recommendation method of AI big data, and in an embodiment of the invention, the method includes:
Step S100: receiving an equipment chain input by a user side, and creating a task table according to the equipment chain; the equipment chain comprises an equipment model and a logic connection relation thereof;
In the existing intelligent production workshops, each product is provided with a corresponding production line, a plurality of production devices are arranged on the production line, the production devices are all finished products, when the equipment is purchased, equipment models (used for indicating which parts the equipment contains and the relation among the parts are shown, core parameters such as the size are not provided) are synchronously provided by equipment providers, and a user side (a management side of the production workshops) arranges the equipment models corresponding to the production devices according to the production line, so that an equipment chain can be obtained; after receiving the equipment chain uploaded by the user side, the execution main body can create a task table according to the equipment chain, wherein tasks in the task table are generally code writing tasks of all modules.
It should be noted that, the device models in the device chain are ordered according to the order of the production line, and may also be referred to as a logic connection relationship.
Step S200: extracting data characteristics of each task in the task table, and correcting the task table according to the data characteristics; the data features are used for representing the demand parameters of each task;
The tasks in the task table are the control code writing tasks of the various components, and the different components of different devices can be similar, such as an image acquisition processing module (used for identifying products), and the corresponding tasks are almost similar; based on the method, the tasks in the task table can be optimized, and the data size in the task table is reduced.
Specifically, the requirement parameters of the task are described, and the requirement parameters can be directly used as task items or can be used as task items after being processed; the requirement parameters are built in an input-output format, such as image-abnormal coordinates, which are a pair of requirement parameters used for indicating that the task is written to identify the image and input the abnormal position; in this process, definition information such as images (RGB, definition) -abnormal positions (accuracy) and the like may be added at the input and output tails, specifically set by the staff autonomously.
Step S300: traversing a preset task library according to the corrected task table, and establishing a virtual task table;
According to the corrected task table, traversing is carried out in a preset task library, a virtual task table corresponding to the task table can be built, the virtual task table can hide the requirements of a user side, and privacy of the user side is protected in a phase-changing manner.
Step S400: acquiring a solution table based on the virtual task table, debugging the equipment chain according to the solution table, and outputting a task flow;
Establishing a connection channel with a answering party based on the virtual task table, acquiring answering information, and generating an answering table; compiling an initial equipment chain according to the solution information in the solution table, so as to determine a task flow; the task flow is a collection of solution information for completing the production task.
Step S500: acquiring quotation information of a task flow, sending the quotation information to a user side, receiving selection information uploaded by the user side, and creating a transaction platform;
The number of the answering parties is not unique, the answering information is also not unique, the answering table formed by combining the answering information is also not unique, the answering parties with different answering information have quotation information, and the statistical quotation information can be used for distinguishing task flows for the selection of the user party, and a transaction platform is created after the selection of the user party is completed.
In one implementation of the technical scheme of the invention, the equipment chain and the task table are encrypted by the user side, and the data in the answer table are encrypted by the answer side.
The encryption means that the equipment chain and the task list belong to private data of the user side, other parties cannot browse at will, and secret responsibility is implemented in the method execution main body; the answer information in the answer table is knowledge service of the answer party, if the answer information is exposed to the user party in advance, the user party may not pay the tail money, therefore, the execution body of the method tests the answer information first, decrypts the answer information when the tail money of the user party is received, and sends the answer information to the user party; the security responsibility of solving the information is also subject to the method execution.
Fig. 2 is a first sub-flowchart of a technical transaction recommendation method for AI big data, where the step of receiving a device chain input by a user side and creating a task table according to the device chain includes:
step S101: receiving a production flow and an equipment model input by a user side, and determining a logic connection relation of the equipment model according to the production flow;
the production process and the equipment model are input by a user side, and the production process is used for determining a logic connection relation; the production flow can be understood as a production line.
Step S102: acquiring a component framework of the equipment model, and determining a component connection relation according to the data transmission direction of the equipment model;
the component architecture of the device model is known data from which the device model can be broken into components.
Step S103: establishing a flow chart consisting of nodes according to the logic connection relation and the component connection relation; the flow chart contains an identification frame for representing the membership of the component; the node and the component have a mapping relation;
a flow chart of the component assembly can be established according to the logic connection relation and the component connection relation; the flow chart is graph data, and nodes correspond to the components; the mark box is used to improve the readability of the flow chart.
Step S104: based on the node statistics demand parameters input by the user side, creating a task table according to the demand parameters; the data structure of the demand parameters is a preset value;
counting demand parameters input by a user side based on the nodes, and then creating a task table; each task in the task table corresponds to a node.
It should be noted that the data structure of the requirement parameter needs to be preset, and reference is specifically made to the description of the requirement parameter.
FIG. 3 is a second sub-flowchart of the technical transaction recommendation method for AI big data, wherein the step of extracting the data characteristics of each task in the task table and correcting the task table according to the data characteristics includes:
Step S201: sequentially extracting each task in a task table, and inputting the tasks into a preset conversion model to obtain a conversion value;
step S202: calculating correlation coefficients among the conversion values, and determining the similarity of each task according to the correlation coefficients;
step S203: classifying each task in the task list according to the similarity;
In the classifying process, the original positions of all tasks are recorded.
The above-mentioned content defines the correction process of the task table, its core lies in, confirm the similar task, and classify the task, in the subsequent solving process, reduce the repetition solution rate; specifically, the task is converted into a numerical value which is easy to be in, then the correlation coefficient can be calculated by means of the existing correlation calculation formula, and the numerical value processing is carried out on the correlation coefficient, so that the similarity can be obtained; the correlation number is directly used as the similarity, and the method is a feasible technical scheme.
In the classifying process, the original positions of the tasks are recorded, that is, the corresponding relationship between before and after correction is established.
Fig. 4 is a third sub-flowchart of a technical transaction recommendation method for AI big data, where the step of traversing a preset task library according to the corrected task table, and establishing a virtual task table includes:
Step S301: calculating the task number of various tasks in the corrected task list, and determining a traversal level according to the task number; the traversal level is used for determining the traversal sequence and the traversal precision;
step S302: extracting various tasks according to the traversal level, traversing a preset task library through the tasks, and determining a target task;
step S303: and counting the target tasks and establishing a virtual task table.
The tasks in the corrected task table are already classified and completed, the more the number of the same type of tasks is, the higher the importance is, the processing sequence (traversing sequence) is front, and the traversing precision is higher.
Extracting each type of task in sequence, traversing in a preset task library, and matching similar tasks, namely virtual tasks; one class of tasks corresponds to one virtual task.
Further, the step of extracting various tasks according to the traversal level, traversing a preset task library through the tasks, and determining the target task includes:
determining a preset number of task step sizes, and dividing a task library according to the task step sizes;
extracting various tasks according to the traversal level, sequentially extracting tasks in the tasks, taking the tasks as tasks to be detected, and calculating editing distances between the tasks to be detected and the segmentation tasks;
Comparing the editing distance with a preset editing threshold value, and marking a segmentation task according to the comparison result; the editing threshold is determined by the traversal accuracy;
Determining a target task according to the marking result of the segmentation task; the marking result comprises marking times and editing distances in each marking;
The above-mentioned content limits the matching process of the goal task, first, need to cut the task library, cut the step length by the staff presets, under the general situation, task item and quantity in the task library are not unique; furthermore, there may be an overlap between adjacent segments (dicing tasks); then, analyzing various tasks in sequence, and calculating the editing distances between all tasks in the tasks and each segmentation, wherein the editing distances can reflect the similarity between two character strings; and finally, marking the cut segments according to the editing distance, marking one or more cut segments by one task, and selecting the cut segments with more marks or closer average editing distance to all tasks as target tasks.
The calculation model of the editing distance is as follows:
Wherein dis (a, b) is the editing distance between the target task and the segmentation task, a is the character string representing the target task, and b is the character string representing the segmentation task; the |a| and the |b| are the character string lengths; tail (a) is a tail character string except the first character in the character string of the target task, and tail (b) is a tail character string except the first character in the character string of the segmentation task; a [0] is the first character of the character string of the target task, and b [0] is the first character of the character string of the segmentation task.
Further explanation is required for the computational model:
The calculation model is a recursive model, dis (a, b) is changed continuously along with the recursion progress, and finally a numerical value can be obtained and used for reflecting the editing distance, and the smaller the editing distance is, the fewer the conversion operations (adding, deleting and replacing) between two character strings are, the more similar the two character strings are.
Fig. 5 is a fourth sub-flowchart of a technical transaction recommendation method for AI big data, wherein the steps of obtaining a solution table based on the virtual task table, debugging the equipment chain according to the solution table, and outputting a task flow include:
Step S401: the virtual task table is sent to a solver, solution information fed back by the solver is obtained, and a solution table corresponding to the virtual task table is established;
Step S402: inquiring the corresponding relation between the virtual task table and the original task table to obtain a solution table corresponding to the original task table, which is called an application solution table;
Step S403: debugging the equipment chain according to the application solution table, and outputting a task flow;
wherein the answer information of each virtual task is at least one.
Each virtual task corresponds to a plurality of answer information, so that the answer table has a plurality of combination modes, and the number of the answer tables is not unique; the corresponding relation exists between the virtual task table and the corrected task table, and the corresponding relation exists between the corrected task table and the task table before correction, so that the virtual task corresponding to each task in the original task table can be determined; and sorting the solution information of the virtual tasks based on the original task table to obtain a solution table, which is called an application solution table.
And inputting the solution information in the application solution table into a device chain and compiling, and removing the invalid application solution table, wherein the rest application solution table is called a task flow.
Fig. 6 is a fifth sub-flowchart of a technical transaction recommendation method for AI big data, where the steps of obtaining quotation information of a task flow, sending the quotation information to a user side, receiving selection information uploaded by the user side, and creating a transaction platform include:
Step S501: acquiring quotation information of each answer information in a task flow, and recording the application times of the answer information in an original task table;
Step S502: determining price tables corresponding to different application answer tables according to quotation information and application times;
step S503: and sending the price list to the user side, and creating a transaction platform when receiving the selection information sent by the user side.
The number of the obtained task flows is large, quotation information is introduced in the selection process of the user side, the user side autonomously combines different answer information according to the price table, after obtaining the answer table meeting the budget, the method executes the main body to create a transaction platform, and the subsequent tail money delivery link is completed.
In a preferred embodiment of the present invention, there is provided a technical transaction recommendation system for AI big data, the system including:
The task list creation module is used for receiving the equipment chain input by the user side and creating a task list according to the equipment chain; the equipment chain comprises an equipment model and a logic connection relation thereof;
The task table correction module is used for extracting the data characteristics of each task in the task table and correcting the task table according to the data characteristics; the data features are used for representing the demand parameters of each task;
the virtual table generation module is used for traversing a preset task library according to the corrected task table and establishing a virtual task table;
the task flow output module is used for acquiring a solution table based on the virtual task table, debugging the equipment chain according to the solution table and outputting a task flow;
the transaction platform creation module is used for acquiring quotation information of the task flow, sending the quotation information to the user side, receiving selection information uploaded by the user side and creating a transaction platform;
The device chain and the task table are encrypted by the user side, and the data in the answer table are encrypted by the answer side.
Further, the task table creation module includes:
The logic relation determining unit is used for receiving the production flow and the equipment model input by the user side and determining the logic connection relation of the equipment model according to the production flow;
the component relation determining unit is used for obtaining the component architecture of the equipment model and determining the component connection relation according to the data transmission direction of the equipment model;
The flow chart establishing unit is used for establishing a flow chart consisting of nodes according to the logic connection relation and the component connection relation; the flow chart contains an identification frame for representing the membership of the component; the node and the component have a mapping relation;
the creating execution unit is used for counting demand parameters input by a user side based on the nodes and creating a task table according to the demand parameters; the data structure of the demand parameter is a preset value.
Specifically, the task correction module includes:
The type conversion unit is used for sequentially extracting each task in the task table, and inputting the tasks into a preset conversion model to obtain conversion values;
The similarity calculation unit is used for calculating correlation coefficients among the conversion values and determining the similarity of each task according to the correlation coefficients;
the classification execution unit is used for classifying each task in the task list according to the similarity;
In the classifying process, the original positions of all tasks are recorded.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.