Machine learning based optimal alloy performance prediction method and apparatus

Abstract

The application provides a kind of optimal alloy performance prediction method and device based on machine learning and electronic equipment, it is related to alloy performance optimization technical field, alleviate the technical problem that the accuracy of alloy performance prediction result in prior art is poor.The method comprises: obtaining original data set;Wherein, the composition characteristic data, heat treatment characteristic data and performance characteristic data of multiple alloys are included in the original data set;Based on the preset feature standard, the original data set is subjected to feature screening processing, to obtain the processed feature data set;Wherein, the feature standard includes composition feature standard, heat treatment feature standard and performance feature standard;Based on the feature data set, create the SVR prediction model of alloy optimal performance;Using SVR prediction model, the performance prediction of the data set to be predicted is carried out by traversing backtracking algorithm, to obtain the optimal performance prediction result corresponding to the data set to be predicted.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of alloy performance optimization, and in particular to an optimal alloy performance prediction method and device based on machine learning and an electronic device. BACKGROUND

[0002] Among alloy materials, Al-Zn-Mg-Cu alloy has a surprising strength-to-mass ratio advantage, and is therefore widely used in the aviation industry. At present, with the progress of the times and the development of aviation technology, higher requirements are also placed on the performance of Al-Zn-Mg-Cu alloy, and how to discover new high-performance alloys remains a major problem. Therefore, computational methods play an increasingly important role in the study of alloy materials. Compared with traditional computational methods, machine learning methods have their special advantages, which can quickly solve complex nonlinear relationships and easily update the machine learning model by replacing the relevant data set.

[0003] However, the prior art has the technical problem of poor accuracy of alloy performance prediction results. SUMMARY

[0004] The purpose of the present application is to provide an optimal alloy performance prediction method and device based on machine learning to alleviate the technical problem of poor accuracy of alloy performance prediction results in the prior art.

[0005] In a first aspect, an optimal alloy performance prediction method based on machine learning is provided, which comprises:

[0006] obtaining an original data set; wherein the original data set includes component feature data, heat treatment feature data and performance feature data of a plurality of alloys;

[0007] performing feature screening processing on the original data set based on a preset feature standard to obtain a processed feature data set; wherein the feature standard includes a component feature standard, a heat treatment feature standard and a performance feature standard;

[0008] creating a support vector regression (SVR) prediction model of the optimal performance of the alloy based on the feature data set;

[0009] using the SVR prediction model to perform performance prediction on a to-be-predicted data set through a traversal backtracking algorithm to obtain an optimal performance prediction result corresponding to the to-be-predicted data set.

[0010] In a possible implementation, the performance prediction of the to-be-predicted data set by traversing the backtracking algorithm using the SVR prediction model comprises the following steps.

[0011] Randomly combining and selecting the feature data in the to-be-predicted data set to obtain all to-be-predicted data.

[0012] Performing performance prediction on the all to-be-predicted data using the SVR prediction model respectively to obtain the optimal performance prediction result corresponding to the to-be-predicted data set.

[0013] In a possible implementation, the randomly combining and selecting the feature data in the to-be-predicted data set to obtain all to-be-predicted data comprises the following steps.

[0014] Randomly combining and selecting the feature data in the to-be-predicted data set using random noise rule sampling to obtain all to-be-predicted data.

[0015] In a possible implementation, the performance prediction of the to-be-predicted data set by traversing the backtracking algorithm using the SVR prediction model comprises the following steps.

[0016] Performing performance prediction on target to-be-predicted data using the SVR prediction model to obtain a target prediction result.

[0017] Performing a first judgment on whether the target prediction result is an optimal prediction result in current predicted results.

[0018] If the first judgment result is yes, performing a second judgment on whether the current predicted results cover the all to-be-predicted data.

[0019] If the second judgment result is yes, taking the target prediction result as the optimal performance prediction result corresponding to the to-be-predicted data set.

[0020] In a possible implementation, the component feature data comprises any one or more of the following:

[0021] Target element occurrence frequency and target element average content.

[0022] In a possible implementation, the heat treatment feature data comprises any one or more of the following:

[0023] Solution treatment parameters and aging treatment parameters.

[0024] In a possible implementation, the performance feature data comprises any one or more of the following:

[0025] Tensile strength, yield strength, and elongation.

[0026] In a second aspect, the embodiments of the present application provide an optimal alloy performance prediction device based on machine learning, the device comprising:

[0027] an acquisition module configured to acquire an original data set, wherein the original data set comprises component feature data, heat treatment feature data and performance feature data of a plurality of alloys;

[0028] a screening module configured to perform feature screening processing on the original data set based on a preset feature standard to obtain a processed feature data set, wherein the feature standard comprises a component feature standard, a heat treatment feature standard and a performance feature standard;

[0029] a creating module configured to create an SVR prediction model of optimal performance of the alloy based on the feature data set;

[0030] a prediction module configured to perform performance prediction on a to-be-predicted data set by using the SVR prediction model and through a traversal backtracking algorithm to obtain an optimal performance prediction result corresponding to the to-be-predicted data set.

[0031] In a third aspect, the embodiments of the present application provide an electronic device, comprising a memory and a processor, the memory storing a computer program executable on the processor, and the processor implementing the steps of the method of the first aspect when executing the computer program.

[0032] In a fourth aspect, the embodiments of the present application provide a computer readable storage medium, the computer readable storage medium storing computer executable instructions, and the computer executable instructions causing the processor to run the steps of the method of the first aspect when being called and run by the processor.

[0033] The embodiments of the present application bring the following beneficial effects:

[0034] The embodiment of the application provides a kind of optimal alloy performance prediction method, device and electronic equipment based on machine learning, first, obtain original data set, wherein the original data set includes the component feature data, heat treatment feature data and performance feature data of multiple alloys, then based on preset feature standard, feature screening processing is carried out to original data set, obtain the feature data set after processing, wherein the feature standard includes component feature standard, heat treatment feature standard and performance feature standard, to create the SVR prediction model of alloy optimal performance based on feature data set, and then using SVR prediction model, the performance prediction of the data set to be predicted is carried out by traversing backtracking algorithm, and the optimal performance prediction result corresponding to the data set to be predicted is obtained.In the scheme, the method of traversal backtracking is introduced, and the higher performance of alloy is predicted by its powerful search capability, and based on the characteristics of aluminum alloy, the data set construction mode of alloy component, heat treatment system and alloy performance trinity is proposed, so that the machine learning prediction result is more reasonable and reliable, and the technical problem that the accuracy of alloy performance prediction result in prior art is poor is solved. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the specific embodiments of the application or the technical solutions in the prior art, the drawings needed in the description of the specific embodiments or the prior art will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0036] Figure 1 A flowchart of a kind of optimal alloy performance prediction method based on machine learning provided by the embodiment of the application;

[0037] Figure 2 A schematic diagram of alloy optimal performance prediction model based on machine learning traversal backtracking method provided by the embodiment of the application;

[0038] Figure 3 A structure schematic diagram of a kind of optimal alloy performance prediction device based on machine learning provided by the embodiment of the application;

[0039] Figure 4 A structure schematic diagram of a kind of electronic equipment provided by the embodiment of the application. DETAILED DESCRIPTION

[0040] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall into the scope of the present application.

[0041] The terms "comprising" and "having" and any variations thereof in the embodiments of the present application are intended to cover the inclusion not the exclusion of one or more elements. For example, a process, method, system, product or apparatus that comprises a list of steps or units is not limited to the listed steps or units but can optionally further include other steps or units not listed or can optionally further include other steps or units inherent to such process, method, product or apparatus.

[0042] Among alloy materials, Al-Zn-Mg-Cu alloy has a surprising strength-to-mass ratio advantage, so it is widely used in the aviation industry. At present, with the progress of the times and the development of aviation technology, higher requirements have been put forward for the performance of Al-Zn-Mg-Cu alloy. Some complex technologies have already improved the performance of Al-Zn-Mg-Cu alloy to a very high level, such as severe plastic deformation (SPD), spray forming, selective laser melting and powder metallurgy. But these technologies are always affected by high cost, complex operation, limited product size, etc., so that the practical application of these technologies in industry is greatly limited. It is worth noting that although there are many preparation technologies, the composition design is always the key factor that determines the performance of the material. In addition, Al-Zn-Mg-Cu alloy has many elements, a wide composition range and complex strengthening mechanisms. In particular, the addition of some trace elements, such as Zr, Cr, Ti or rare earth elements, can enhance its performance. Therefore, the composition development space of Al-Zn-Mg-Cu alloy is huge, which leads to the time-consuming and laborious of the conventional trial-and-error method. How to discover new Al-Zn-Mg-Cu alloy and effectively improve the performance of aluminum alloy is still a major challenge. With the progress of the times and the development of aviation technology, modern industry has put forward higher requirements for the performance of alloys. However, there are many alloy elements, and the strengthening mechanism is complex and diverse. This leads to the traditional experimental trial-and-error method to become time-consuming and laborious. How to discover new high-performance alloys is still a major problem. Therefore, the calculation method plays an increasingly important role in the research of alloy materials. Compared with traditional calculation methods, machine learning methods have their special advantages. They can quickly solve complex nonlinear relationships and easily update the machine learning model by changing the relevant data set. Due to the complex nonlinear relationship between alloy composition and alloy performance, machine learning in the optimal prediction of multiple performances of alloys is still a hot topic. However, the current related researches are mostly directly using machine learning regression algorithms to fit and predict the performance, and the prediction results are difficult to exceed the range size of the original data set. In addition, the current researches are less on modeling the multi-dimensional complexity of alloy composition, preparation process and alloy performance. Most of the machine learning applications are only on the modeling of composition-performance or heat treatment system-performance. Moreover, for aluminum alloys, composition and heat treatment system are the key factors that affect the performance of the alloy.

[0043] Therefore, the embodiments of the present application provide a machine learning-based optimal alloy performance prediction method and device and electronic equipment, which alleviate the technical problem of poor accuracy of alloy performance prediction results in the prior art.

[0044] The embodiments of the present application will be further described below with reference to the accompanying drawings.

[0045] Figure 1 A flowchart of an optimal alloy performance prediction method based on machine learning provided for an embodiment of the present application. As shown in Figure 1 The method comprises:

[0046] Step S110, obtaining an original data set.

[0047] Among the original data set, there are composition feature data, heat treatment feature data and performance feature data of multiple alloys.

[0048] Illustratively, data collection and feature selection are key factors in determining the prediction ability of the machine learning model. First, the original data set is obtained, and the data in the original data set can come from related literature with a preparation method of traditional casting process. Considering the uniformity of the data, the data source needs to contain detailed data of element content, heat treatment system and alloy performance. Based on the above principle, an aluminum alloy original data set composed of element content (composition feature data), heat treatment parameters (heat treatment feature data) and material performance (performance feature data) is established.

[0049] Step S120, based on a preset feature standard, performing feature screening processing on the original data set to obtain a processed feature data set.

[0050] Among the feature standards, there are composition feature standards, heat treatment feature standards and performance feature standards.

[0051] Illustratively, as shown in Figure 2 Before machine learning modeling, the original data set needs to be cleaned and feature screened. First, for the composition feature standard, it can be set that only the elements with more than 100 occurrences and an average content of more than 0.01wt% in the original data set are selected as the composition feature data. In addition, considering the importance of heat treatment system to alloy performance, for the heat treatment feature standard, the solid solution treatment parameters and aging treatment parameters can be selected as the heat treatment features of the machine learning modeling. Moreover, considering the characteristics of aluminum alloy, the tensile strength, yield strength and elongation can be selected as the performance features of the machine learning modeling. It should be noted that the specific values of the preset feature standard can be selected according to the actual situation.

[0052] Step S130, based on the feature data set, creating an SVR prediction model of optimal alloy performance.

[0053] Illustratively, as shown in Figure 2 Using the processed feature data set, the SVR prediction model is created and optimized to obtain the SVR prediction model of optimal alloy performance.

[0054] Step S140, using the SVR prediction model, the performance prediction of the to-be-predicted data set is performed through the traversal backtracking algorithm, and the optimal performance prediction result corresponding to the to-be-predicted data set is obtained.

[0055] As shown in the example, Figure 2 As shown in the example, The traversal backtracking algorithm is a hyperparameter optimization method in machine learning, that is, after running all possible features in each random combination, the optimal result is output. The backtracking algorithm has strong search capability, and many complex and large-scale problems can use the backtracking search algorithm to obtain all feasible solutions (similar to the exhaustive method). The composition feature data and the heat treatment feature data are taken as inputs, and the performance feature data is taken as output, and then the optimal solution, that is, the optimal performance prediction result corresponding to the to-be-predicted data set, is obtained.

[0056] It should be noted that in addition to SVR, other machine learning algorithms such as linear regression, neural network, random forest, etc. can also be used for regression fitting in model construction. The optimal alloy performance prediction method based on machine learning provided in the embodiments of the application can not only be used in the design of aluminum alloys, but also be used in the design of other materials, including copper alloys, titanium alloys, high-entropy alloys, steels, semiconductor materials, batteries, etc.

[0057] In the embodiments of the application, the traversal backtracking method is introduced, the powerful search capability is used to realize the prediction of the higher performance of the alloy, and based on the characteristics of the aluminum alloy, a three-in-one data set construction mode of alloy composition, heat treatment system and alloy performance is proposed, so that the machine learning prediction result is more reasonable and reliable, and the technical problem of poor accuracy of the alloy performance prediction result in the prior art is solved.

[0058] The above steps will be described in detail below.

[0059] In some embodiments, the above step S130 can specifically include the following steps:

[0060] Step a), the feature data in the to-be-predicted data set is randomly combined and selected to obtain all to-be-predicted data.

[0061] Step b), using the SVR prediction model, the performance of all to-be-predicted data is predicted to obtain the optimal performance prediction result corresponding to the to-be-predicted data set.

[0062] Exemplarily, the feature data in the to-be-predicted data set can be randomly combined and selected, and based on different component feature data and heat treatment feature data, permutation and combination is performed to obtain all to-be-predicted data, and the SVR prediction model is used to perform performance prediction on each to-be-predicted data respectively to obtain the optimal performance prediction result corresponding to the to-be-predicted data set.

[0063] By first randomly combining and selecting the feature data in the to-be-predicted data set to obtain all to-be-predicted data, and then using the SVR prediction model to perform performance prediction on all to-be-predicted data respectively to obtain the optimal performance prediction result corresponding to the to-be-predicted data set, it can be ensured that the prediction process covers all possibilities, ensures the uniqueness of the optimal performance prediction result, and improves the accuracy of the alloy performance prediction result.

[0064] Based on the above steps a) and b), the above step a) can specifically include the following steps:

[0065] Step c), randomly combining and selecting the feature data in the to-be-predicted data set by using random noise rule sampling to obtain all to-be-predicted data.

[0066] Exemplarily, the random noise is added to the parameters of the rule sampling to optimize the traversal backtracking model. The specific modification method can be to add a uniform distribution U(-interval / 2.interval / 2) to each feature in the combination generated in the model running process, where interval represents the interval of the regular sampling. In the modeling process, when the regular sampling is performed, the number of nodes corresponding to each parameter is [15, 14, 5, 6, 6, 7, 6, 5, 6, 6, 5], so there are about 1.4e9 possible combinations.

[0067] By adding random noise to the parameters of the rule sampling in the model construction to optimize the traversal backtracking model, the accuracy of the alloy performance prediction result is improved.

[0068] Based on the above steps a) and b), the above step b) can specifically include the following steps:

[0069] Step d), using the SVR prediction model to perform performance prediction on the target to-be-predicted data to obtain a target prediction result.

[0070] Step e), first judging whether the target prediction result is the optimal prediction result in the current predicted result.

[0071] Step f), if the first judgment result is yes, second judging whether the current predicted result covers all to-be-predicted data.

[0072] Step g), if the second judgment result is yes, the target prediction result is taken as the optimal performance prediction result corresponding to the to-be-predicted data set.

[0073] As shown in the example, Figure 2 As shown in the example,

[0074] By using the traversal backtracking algorithm, all data combinations in the to-be-predicted data set are calculated and compared, so as to determine the optimal performance prediction result corresponding to the to-be-predicted data set in all data combination possibilities, thereby effectively improving the accuracy of the alloy performance prediction result.

[0075] In some embodiments, the component feature data includes any one or more of the following:

[0076] The number of occurrences of the target element, the average content of the target element.

[0077] In some embodiments, the heat treatment feature data includes any one or more of the following:

[0078] Solution treatment parameters, aging treatment parameters.

[0079] In some embodiments, the performance feature data includes any one or more of the following:

[0080] Tensile strength, yield strength, elongation.

[0081] It should be noted that the types and quantities of the above-mentioned component feature data, heat treatment feature data and performance feature data can be adjusted according to actual needs, and the embodiments of the present application do not limit this.

[0082] Figure 3 A structure schematic diagram of an optimal alloy performance prediction device based on machine learning provided by the embodiments of the present application is shown in the figure. Figure 3 As shown in the example,

[0083] The acquisition module 301 is configured to acquire an original data set, wherein the original data set comprises component feature data, heat treatment feature data and performance feature data of multiple alloys.

[0084] The screening module 302 is configured to perform feature screening processing on the original data set based on a preset feature standard to obtain a processed feature data set, wherein the feature standard comprises a component feature standard, a heat treatment feature standard and a performance feature standard.

[0085] The creation module 303 is configured to create an SVR prediction model of optimal performance of an alloy based on the feature data set.

[0086] The prediction module 304 is configured to perform performance prediction on a to-be-predicted data set by using the SVR prediction model and traversing a backtracking algorithm to obtain a prediction result of optimal performance corresponding to the to-be-predicted data set.

[0087] In some embodiments, the prediction module 304 is configured to:

[0088] randomly combine and select feature data in the to-be-predicted data set by using random noise rule sampling to obtain all to-be-predicted data.

[0089] In some embodiments, the prediction module 304 is configured to:

[0090] randomly combine and select feature data in the to-be-predicted data set to obtain all to-be-predicted data;

[0091] perform performance prediction on all to-be-predicted data by using the SVR prediction model respectively to obtain a prediction result of optimal performance corresponding to the to-be-predicted data set.

[0092] In some embodiments, the prediction module 304 is configured to:

[0093] perform performance prediction on target to-be-predicted data by using the SVR prediction model to obtain a target prediction result;

[0094] perform a first judgment on whether the target prediction result is an optimal prediction result in current predicted results;

[0095] if the first judgment result is yes, perform a second judgment on whether the current predicted results cover all to-be-predicted data;

[0096] if the second judgment result is yes, take the target prediction result as the prediction result of optimal performance corresponding to the to-be-predicted data set.

[0097] In some embodiments, the component feature data comprises any one or more of the following:

[0098] a target element occurrence frequency and a target element average content.

[0099] In some embodiments, the heat treatment feature data comprises any one or more of the following:

[0100] solution treatment parameters, aging treatment parameters.

[0101] In some embodiments, the performance feature data comprises any one or more of the following:

[0102] tensile strength, yield strength, elongation.

[0103] The apparatus provided by the embodiments of the present application has the same implementation principles and technical effects as the foregoing method embodiments. For brevity, the system embodiments are not described in detail herein, and the reader can refer to the corresponding content in the foregoing method embodiments.

[0104] The electronic device provided by the embodiments of the present application comprises a processor and a storage device. The storage device stores a computer program. When the computer program is executed by the processor, the method of any one of the above embodiments is executed.

[0105] Figure 4 The electronic device provided by the embodiments of the present application comprises a processor 401, a memory 402, a bus 403 and a communication interface 404. The processor 401, the communication interface 404 and the memory 402 are connected through the bus 403. The processor 401 is configured to execute the executable modules stored in the memory 402, such as a computer program.

[0106] The memory 402 can include a high-speed random access memory (RAM) and can also include a non-volatile memory such as at least one disk memory. The communication between the system network element and at least one other network element is realized through at least one communication interface 404 (which can be wired or wireless). The communication can be realized through the Internet, a wide area network, a local area network, a metropolitan area network, etc.

[0107] The bus 403 can be an ISA bus, a PCI bus or an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For brevity, Figure 4 Only one bidirectional arrow is used to represent the bus in the figure, but it does not mean that there is only one bus or only one type of bus.

[0108] The memory 402 is configured to store a program. The processor 401 executes the program after receiving an execution instruction. The method executed by the apparatus defined by the flow process of any one of the foregoing embodiments of the present application can be applied to the processor 401 or realized by the processor 401.

[0109] The processor 401 can be an integrated circuit chip having a signal processing capability. In the implementation process, each step of the above method can be completed by the integrated logic circuit or the instruction in the form of software in the processor 401. The processor 401 described above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component. Each method, step and logic block diagram disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as a hardware code processor for execution, or a combination of hardware and software modules in the code processor for execution. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium in the art. The storage medium is located in the memory 402, and the processor 401 reads the information in the memory 402, and combines the hardware to complete the steps of the above method.

[0110] The computer program product of the readable storage medium provided by the embodiments of the present application comprises a computer readable storage medium storing program codes, and the instructions included in the program codes can be used to execute the method in the foregoing method embodiments. For specific implementation, please refer to the foregoing method embodiments, which will not be described here.

[0111] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the part of the prior art that essentially contributes to the present application or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0112] Finally, it should be noted that the above embodiments are only specific implementations of the present application, which are used to illustrate the technical solutions of the present application, but not to limit them. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily think of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed by the present application, or make equivalent replacements to some of the technical features. These modifications, changes or replacements do not cause the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for predicting optimal alloy performance based on machine learning, characterized by, The method comprises: obtaining an original data set; wherein the original data set comprises component feature data, heat treatment feature data and performance feature data of a plurality of alloys; performing feature screening processing on the original data set based on a preset feature standard to obtain a processed feature data set; wherein the feature standard comprises a component feature standard, a heat treatment feature standard and a performance feature standard; creating an SVR prediction model of optimal performance of the alloy based on the feature data set; performing performance prediction on a to-be-predicted data set by using the SVR prediction model through a traversal backtracking algorithm to obtain an optimal performance prediction result corresponding to the to-be-predicted data set; wherein the performance prediction on the to-be-predicted data set by using the SVR prediction model through the traversal backtracking algorithm to obtain the optimal performance prediction result corresponding to the to-be-predicted data set comprises: randomly combining and selecting feature data in the to-be-predicted data set to obtain all to-be-predicted data; performing performance prediction on the all to-be-predicted data by using the SVR prediction model respectively to obtain the optimal performance prediction result corresponding to the to-be-predicted data set; the performance prediction on the all to-be-predicted data by using the SVR prediction model respectively to obtain the optimal performance prediction result corresponding to the to-be-predicted data set comprises: performing performance prediction on target to-be-predicted data by using the SVR prediction model to obtain a target prediction result; performing a first judgment on whether the target prediction result is an optimal prediction result in current predicted results; if the first judgment result is yes, performing a second judgment on whether the current predicted results cover the all to-be-predicted data; if the second judgment result is yes, taking the target prediction result as the optimal performance prediction result corresponding to the to-be-predicted data set.

2. The method of claim 1, wherein, the randomly combining and selecting the feature data in the to-be-predicted data set to obtain the all to-be-predicted data comprises: randomly combining and selecting the feature data in the to-be-predicted data set by using random noise rule sampling to obtain the all to-be-predicted data.

3. The method of claim 1, wherein, the component feature data comprises any one or more of the following: target element occurrence frequency, target element average content.

4. The method of claim 1, wherein, the heat treatment feature data comprises any one or more of the following: solution treatment parameters, aging treatment parameters.

5. The method of claim 1, wherein, the performance feature data comprises any one or more of the following: tensile strength, yield strength, elongation.

6. A machine learning based optimal alloy performance prediction apparatus, characterized by, the device comprises: an acquisition module configured to acquire an original data set; wherein the original data set comprises component feature data, heat treatment feature data and performance feature data of a plurality of alloys; a screening module configured to perform feature screening processing on the original data set based on a preset feature standard to obtain a processed feature data set; wherein the feature standard comprises a component feature standard, a heat treatment feature standard and a performance feature standard; a creating module configured to create an SVR prediction model of optimal performance of the alloy based on the feature data set; The prediction module is configured to perform performance prediction on the to-be-predicted data set by traversing a backtracking algorithm using the SVR prediction model, to obtain an optimal performance prediction result corresponding to the to-be-predicted data set. The prediction module is further configured to randomly combine and select feature data in the to-be-predicted data set to obtain all to-be-predicted data, and perform performance prediction on the all to-be-predicted data using the SVR prediction model to obtain the optimal performance prediction result corresponding to the to-be-predicted data set. The prediction module is further configured to perform performance prediction on target to-be-predicted data using the SVR prediction model to obtain a target prediction result, and perform a first judgment on whether the target prediction result is an optimal prediction result in currently predicted results, and perform a second judgment on whether the currently predicted results cover the all to-be-predicted data if the first judgment result is yes, and take the target prediction result as the optimal performance prediction result corresponding to the to-be-predicted data set if the second judgment result is yes.

7. An electronic device comprising a memory, a processor, the memory having stored therein a computer program executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer executable instructions, and the computer executable instructions, when called and executed by the processor, cause the processor to execute the method in any one of claims 1 to 5.

Images