A data mining system, method and program product

Abstract

The application discloses a data mining system, method and program product, and belongs to the technical field of artificial intelligence. In the system, a task distributor is configured to distribute data mining tasks to N intelligent data mining models, a decision maker is configured to judge data results mined by the intelligent data mining models, to transmit true data results to a data fuser for data fusion, and to discard false data results, a global parameter calculator calculates global parameters according to data results reported by the N intelligent data mining models, a data mining strategy and a reward, the nth intelligent data mining model updates parameters of the intelligent data mining model according to the global parameters, and performs data mining from a data source according to the distributed task, to generate data results for the distributed task, and reports the data results, the data mining strategy and the reward to a central processor. The application does not need to use an experience pool to store historical samples, saves storage space, improves data sampling efficiency, and improves training speed.

Description

TECHNICAL FIELD

[0001] The present application relates to a data mining system, method and program product, belonging to the field of artificial intelligence. BACKGROUND

[0002] The Chinese patent application with the publication number CN109597839A discloses a data mining method based on an aviation electronic combat situation, which includes the following three aspects: establishing a multi-platform aviation electronic system resource cloud platform; generating fire control attack training data based on STK; and establishing a fire control attack flight path planning algorithm based on deep reinforcement learning; the details are as follows: 1.1) the multi-platform aviation electronic system resource cloud platform is equipped with different system functions and resources on each aircraft platform, and the aviation electronic systems of each aircraft form a "cyber" aviation electronic system through a data link, the aviation electronic system of each aircraft can be divided into multiple integrated application function areas, and different application function areas on different platforms can be combined dynamically according to the requirements of the combat task, and system information integration is realized to form a "logical aviation electronic system" oriented to the task; the multi-platform aviation electronic system resource cloud platform provides platform support for the generation and application of aviation electronic data; 1.2) the generation of fire control attack training data based on STK uses the STK tool to model a complex battlefield environment to realize the scenario of the combat task in a cooperative combat scene; and the STK tool is used to call a fire control program in real time to record the attitude of each missile launch of the aircraft and whether the target is hit in real time, and to accumulate training data to prepare for the data mining part of the data mining system based on the aviation electronic combat situation; 1.3) the establishment of a fire control attack flight path planning algorithm based on deep reinforcement learning trains a deep reinforcement learning neural network based on the acquisition of battlefield situation data to extract data features of the flight path planning; at the same time, a flight simulation data oriented to the attack effect is generated based on the STK tool to establish a flight path planning evaluation model based on deep reinforcement learning; the current battlefield situation data is read on the trained flight path planning evaluation model, and the flight path planning is realized by using the deep reinforcement learning neural network to output the flight state parameters of the aircraft.

[0003] However, the patent needs a playback memory unit, and requires a large storage space. SUMMARY

[0004] In order to overcome the shortcomings in the prior art, the present application aims to provide a data mining system, method and computer program product, which does not need to use an experience pool to store historical samples, saves storage space, improves the sampling efficiency of data, and thus improves the training speed.

[0005] To achieve the object of the present application, the present application provides a data mining system comprising a central processor and N intelligent data mining models, wherein the central processor comprises a task allocator, a decision maker, a data fusioner and a global parameter calculator, wherein the task allocator is configured to allocate data mining tasks to the N intelligent data mining models, the decision maker is configured to judge the data results mined by the intelligent data mining models, transmit the true data results to the data fusioner for data fusion, and discard the false data results; the data fusioner is configured to perform data fusion on the true data results mined by the N intelligent data mining models; the global parameter calculator calculates global parameters according to the data results reported by the N intelligent data mining models, data mining strategies and rewards; the nth intelligent data mining model updates its own parameters according to the global parameters, performs data mining from a data source according to the allocated task, generates data results for the allocated task, and reports the data results, data mining strategies and rewards to the central processor, n = 1, 2, …, N.

[0006] To achieve the object of the present application, the present application further provides a data mining method comprising the following steps:

[0007] Step 1: allocating data mining tasks to the N intelligent data mining models by the task allocator of the central processor;

[0008] Step 2: the nth intelligent data mining model performs data mining from a data source according to the allocated task, generates data results for the allocated task, and reports the data results, data mining strategies and rewards to the central processor, n = 1, 2, …, N

[0009] Step 3: judging the data results mined by the intelligent data mining models by the decision maker of the central processor, transmitting the true data results to the data fusioner for data fusion, and discarding the false data results;

[0010] Step 4: performing data fusion on the true data results mined by the N intelligent data mining models by the data fusioner;

[0011] Step 5: calculating global parameters according to the data results reported by the N intelligent data mining models, data mining strategies and rewards by the global parameter calculator;

[0012] Step 6: the nth intelligent data mining model updates its own parameters according to the global parameters.

[0013] To achieve the object of the present application, the present application further provides a computer program product comprising computer program code capable of calling a processor to execute the above-mentioned method.

[0014] Compared with the prior art, the data mining system, method and computer program product provided by the present application have the following beneficial effects:

[0015] No need to use experience pool to store historical samples, save storage space, improve data sampling efficiency, and thus improve training speed. At the same time, samples are collected in multiple different training environments, and the distribution of samples is more uniform, which is more conducive to the training of neural networks. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a component block diagram of the data mining system provided by the first embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to make the technical problems, technical solutions and beneficial effects of the present application clearer, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0018] First embodiment

[0019] Figure 1 is a component block diagram of the data mining system provided by the first embodiment of the present application, as Figure 1 shown, the data mining system provided by the first embodiment of the present application includes a central processor and N intelligent data mining models, wherein the central processor includes a task allocator, a decision maker, a data fusion device and a global parameter calculator, wherein the task allocator is configured to allocate data mining tasks to the N intelligent data mining models, the decision maker is configured to judge the data results mined by the intelligent data mining models, and the true data results are transmitted to the data fusion device for data fusion, and the false data results are discarded; the data fusion device is configured to fuse the true data results mined by the N intelligent data mining models; the global parameter calculator calculates the global parameters according to the data results reported by the N intelligent data mining models, the data mining strategy and the reward; the nth intelligent data mining model updates its parameters according to the global parameters, and performs data mining from the data source according to the allocated task, generates data results for the allocated task, and reports the data results, the data mining strategy and the reward to the central processor, n = 1, 2, …, N.

[0020] In the first embodiment, the nth intelligent data mining model includes a data mining strategy network and an action value function , is the to-be-optimized parameter of the data mining strategy network of the nth intelligent data mining model, and p n is the parameter of the value function, and the training of the nth intelligent data mining model includes the following processes:

[0021] The nth intelligent data mining model performs a series of data mining actions based on the data source. A series of data results were obtained. Received a series of awards and will , and The global parameter calculator, which reports to the central processing unit, shows that t is the number of times and T is a positive integer greater than or equal to 2; based on the data source, the system performs data mining, including data collection, web crawling, and application programming interfaces.

[0022] The global parameter calculator calculates the global parameter w according to Formula 1 and broadcasts the global parameter w to N intelligent data mining models:

[0023] Formula 1:

[0024] ,

[0025] In the formula, This is the discount factor;

[0026] The nth intelligent data mining model updates the parameters of the data mining strategy network using Formula 2:

[0027] Formula 2:

[0028] ,

[0029] In the formula, These are the current parameters of the data mining strategy network of the nth intelligent data mining model; These are the modified parameters of the data mining strategy network for the nth intelligent data mining model; Yes gradient function, It is the learning coefficient.

[0030] In the first embodiment, the nth intelligent data mining model updates the action value function using Formula 3 and the following formula: Parameters:

[0031] Formula 3:

[0032] .

[0033] Optionally, the nth intelligent data mining model includes a data mining strategy network. and action value function , ρ is the parameter to be optimized in the data mining strategy network of the nth intelligent data mining model. nFor the parameters of the value function, the training of the nth intelligent data mining model includes the following processes:

[0034] The nth intelligent data mining model performs a series of data mining actions according to the data source , obtains a series of data results , obtains a series of rewards , and reports , and to the global parameter calculator of the central processor;

[0035] The global parameter calculator calculates the global parameter w according to formula 1, and broadcasts the global parameter w to the N intelligent data mining models;

[0036] The nth intelligent data mining model updates the parameters of the data mining strategy network through formula 4:

[0037] Formula 4:

[0038] , ,

[0039] In the formula, is the learning coefficient, , is the two-norm distance of and , is the distance threshold;

[0040] The nth intelligent data mining model updates the parameters of the action value function through formula 3.

[0041] The data mining system provided by the first embodiment further includes an input unit, through which a user inputs instructions, data, etc.

[0042] The data mining system provided by the first embodiment further includes an output unit, which includes a display and the like, and the display is used to display data strategies, data results, data resources, etc.

[0043] In the first embodiment of the present application, data mining is performed through intelligent data mining models, which can learn autonomously and have the ability to adapt to new data resources and new data mining tasks.

[0044] The first embodiment of the present application does not need to use an experience pool to store historical samples, saving storage space and improving data sampling efficiency, thereby improving training speed. At the same time, samples are collected in multiple different training environments, and the distribution of samples is more uniform, which is more conducive to the training of neural networks.

[0045] The second embodiment

[0046] The second embodiment of the present application only describes the different content from the first embodiment, and the same content is not described repeatedly.

[0047] The second embodiment of the present application provides a data mining method, comprising the following steps:

[0048] Step 1: assigning a data mining task to N intelligent data mining models by a task distributor of a central processor;

[0049] Step 2: the nth intelligent data mining model performs data mining from a data source according to the assigned task, generates a data result for the assigned task, and reports the data result, a data mining strategy and a reward to the central processor, n = 1, 2, …, N

[0050] Step 3: judging the data result mined by the intelligent data mining model by a decision maker of the central processor, transmitting a true data result to a data fusion device for data fusion, and discarding a false data result;

[0051] Step 4: performing data fusion on the true data result mined by the N intelligent data mining models by the data fusion device;

[0052] Step 5: calculating a global parameter according to the data result, the data mining strategy and the reward reported by the N intelligent data mining models by a global parameter calculator;

[0053] Step 6: the nth intelligent data mining model updates its own parameters according to the global parameter.

[0054] In the second embodiment of the present application, the intelligent data mining model is used for data mining, which can autonomously learn and has the ability to adapt to new data resources and new data mining tasks.

[0055] In the second embodiment of the present application, the experience pool is not required to store historical samples, the storage space is saved, the sampling efficiency of data is improved, and the training speed is improved. At the same time, samples are collected in multiple different training environments, the distribution of samples is more uniform, and the training of the neural network is more beneficial.

[0056] Third embodiment

[0057] To achieve the object of the present application, the present application further provides a computer program product comprising computer program code, which can be called by a processor to execute the method described in the second embodiment.

[0058] The beneficial effects of the third embodiment of the present application are the same as those of the second embodiment, which are not described repeatedly here.

[0059] Furthermore, the terms "first", "second", and the like, do not denote any quantity or order, but are used to distinguish a certain feature from another, and are not necessarily used to describe a number of or a sequence. Accordingly, a feature defined with "first", "second", etc. can implicitly or explicitly include one or more of the features. In the description of the application, the meaning of "plurality" is two or more, unless otherwise expressly specified. The meaning of "several" is one or more, unless otherwise expressly specified.

[0060] The basic principles and main features of the application and the advantages of the application have been shown and described above. It should be understood by those skilled in the art that the application is not limited by the above-described embodiments, which are only illustrative of the principles of the application. Various changes and modifications can be made to the application without departing from the spirit and scope of the application, and all such changes and modifications fall within the scope of the application claimed. The scope of the application claimed is defined by the appended claims and their equivalents.

Claims

1. A data mining system, characterized in that, The system includes a central processing unit (CPU) and N intelligent data mining models. The CPU comprises a task allocator, a decision maker, a data fusion maker, and a global parameter calculator. The task allocator is configured to assign data mining tasks to the N intelligent data mining models. The decision maker is configured to judge the data results mined by the intelligent data mining models, sending true data results to the data fusion maker for data fusion and discarding false data results. The data fusion maker is configured to fuse the true data results mined by the N intelligent data mining models. The global parameter calculator calculates global parameters based on the data results, action value functions, and rewards reported by the N intelligent data mining models. The nth intelligent data mining model updates its own parameters based on the global parameters, performs data mining from the data source according to the assigned task, generates data results specific to the assigned task, and reports the data results, action value functions, and rewards to the CPU. n = 1, 2, ..., N, where N is a positive integer greater than or equal to 2. The nth intelligent data mining model includes a data mining strategy network. and action value function , ρ is the parameter to be optimized in the data mining strategy network of the nth intelligent data mining model. n The training of the nth intelligent data mining model, with parameters for the value function, includes the following process: The nth intelligent data mining model performs a series of data mining actions based on the data source. Obtain a series of data results Received a series of awards and will , , The global parameter calculator reported to the central processing unit, where t is the number of times and T is a positive integer greater than or equal to 2; The global parameter calculator calculates the global parameter w according to the following formula and broadcasts the global parameter w to N intelligent data mining models: ， In the formula, This is the discount factor; The nth intelligent data mining model updates the parameters of the data mining policy network using the following formula: ， In the formula, These are the current parameters of the data mining strategy network of the nth intelligent data mining model; These are the modified parameters of the data mining strategy network for the nth intelligent data mining model; Yes The gradient function; It is the learning coefficient.

2. The data mining system according to claim 1, characterized in that, The nth intelligent data mining model updates the action value function using the following formula. Parameters: 。 3. A data mining system, characterized in that, The system includes a central processing unit (CPU) and N intelligent data mining models. The CPU comprises a task allocator, a decision maker, a data fusion maker, and a global parameter calculator. The task allocator is configured to assign data mining tasks to the N intelligent data mining models. The decision maker is configured to judge the data results mined by the intelligent data mining models, sending true data results to the data fusion maker for data fusion and discarding false data results. The data fusion maker is configured to fuse the true data results mined by the N intelligent data mining models. The global parameter calculator calculates global parameters based on the data results, action value functions, and rewards reported by the N intelligent data mining models. The nth intelligent data mining model updates its own parameters based on the global parameters, performs data mining from the data source according to the assigned task, generates data results specific to the assigned task, and reports the data results, action value functions, and rewards to the CPU. n = 1, 2, ..., N, where N is a positive integer greater than or equal to 2. The nth intelligent data mining model includes a data mining strategy network. and action value function , ρ is the parameter of the data mining strategy network of the nth intelligent data mining model. n The training of the nth intelligent data mining model, with parameters for the value function, includes the following process: The nth intelligent data mining model performs a series of data mining operations based on the data source. A series of data results were obtained. Received a series of awards and will , and The global parameter calculator reported to the central processing unit, where t is the number of times and T is a positive integer greater than or equal to 2; The global parameter calculator calculates the global parameter w according to the following formula and broadcasts the global parameter w to N intelligent data mining models: In the formula, This is the discount factor; The nth intelligent data mining model updates the parameters of the data mining policy network using the following formula: ， ， In the formula, It is the learning coefficient. , yes and The second norm distance, This is the distance threshold.

4. The data mining system according to claim 3, characterized in that, The nth intelligent data mining model updates the action value function using the following formula. Parameters: 。 5. A data mining method, characterized in that, Includes the following steps: Step 1: Assign data mining tasks to N intelligent data mining models through the task allocator of the central processing unit; Step 2: The nth intelligent data mining model performs data mining from the data source according to the assigned task, generates data results for the assigned task, and reports the data results, data mining strategy and reward to the central processing unit, n=1,2,…,N; Step 3: The decision-maker of the central processing unit makes a judgment on the data results mined by the intelligent data mining model, and sends the true data results to the data fusion unit for data fusion, while discarding the false data results. Step 4: Use a data fusion tool to fuse the real data results mined by N intelligent data mining models; Step 5: Calculate global parameters using the global parameter calculator based on the data results, data mining strategies, and rewards reported by the N intelligent data mining models; Step 6: The nth intelligent data mining model updates its own parameters based on the global parameters; The training of the nth intelligent data mining model includes the following process: S2-01: The nth intelligent data mining model executes a series of data mining actions based on the data source. A series of data results were obtained. Received a series of awards and will , and The global parameter calculator reported to the central processing unit, the nth intelligent data mining model includes a data mining strategy network. and action value function , ρ is the parameter to be optimized in the data mining strategy network of the nth intelligent data mining model. n Here, t is the parameter of the value function, t is the number of times, and T is a positive integer greater than or equal to 2; S2-02: The global parameter calculator calculates the global parameter w according to the following formula and broadcasts the global parameter w to N intelligent data mining models: ， In the formula, This is the discount factor; S2-03: The nth intelligent data mining model updates the parameters of the data mining strategy network using the following formula: ， In the formula, These are the current parameters of the data mining strategy network of the nth intelligent data mining model; These are the modified parameters of the data mining strategy network for the nth intelligent data mining model; Yes gradient function, It is the learning coefficient; S2-04: The nth intelligent data mining model updates the action value function using the following formula. Parameters: 。 6. A data mining method, characterized in that, The process includes the following: Step 1: Assign data mining tasks to N intelligent data mining models through the task allocator of the central processing unit; Step 2: The nth intelligent data mining model performs data mining from the data source according to the assigned task, generates data results for the assigned task, and reports the data results, data mining strategy and reward to the central processing unit, n=1,2,…,N; Step 3: The decision-maker of the central processing unit makes a judgment on the data results mined by the intelligent data mining model, and sends the true data results to the data fusion unit for data fusion, while discarding the false data results. Step 4: Use a data fusion tool to fuse the real data results mined by N intelligent data mining models; Step 5: Calculate global parameters using the global parameter calculator based on the data results, data mining strategies, and rewards reported by the N intelligent data mining models; Step 6: The nth intelligent data mining model updates its own parameters based on the global parameters. The training of the nth intelligent data mining model includes the following process: S3-01: The nth intelligent data mining model executes a series of data mining actions based on the data source. A series of data results were obtained. Received a series of awards and will , and The global parameter calculator reported to the central processing unit, the nth intelligent data mining model includes a data mining strategy network. and action value function , ρ is the parameter of the data mining strategy network of the nth intelligent data mining model. n Here, t is the parameter of the value function, t is the number of times, and T is a positive integer greater than or equal to 2; S3-02: The global parameter calculator calculates the global parameter w according to the following formula and broadcasts the global parameter w to N intelligent data mining models: In the formula, This is the discount factor; S3-03: The nth intelligent data mining model updates the parameters of the data mining strategy network using the following formula: ， ， In the formula, It is the learning coefficient. , yes and The second norm distance, Distance threshold; S3-04: The nth intelligent data mining model updates the action value function using the following formula: Parameters: 。 7. A computer program product, characterized in that, It includes computer program code that can be called by a processor to execute the data mining method according to any one of claims 5-6.

Images