A material performance prediction method, system, electronic device and storage medium

Abstract

This invention relates to the field of artificial intelligence technology, and specifically discloses a method, system, electronic device, and storage medium for predicting material properties. The method includes: collecting first data affecting material properties; processing the first data to obtain second data; using the first and second data as sample data, and splitting the sample data into a first dataset and a second dataset; training multiple first algorithm models based on the first dataset, and obtaining a second algorithm model based on the first training results; training the second algorithm model based on the second dataset, and obtaining a third algorithm model based on the second training results; and inputting the third data into the third algorithm model to predict the material properties. This invention saves time and improves the efficiency of material performance testing, facilitating the faster selection of materials with better performance.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, system, electronic device, and storage medium for predicting material properties. Background Technology

[0002] Anomalous Hall devices are widely used in aerospace, automotive, and magnetic field detection due to their low fabrication cost, small size, and high reliability. The main factor influencing the performance of anomalous Hall devices is the magnetic thin film material that generates the anomalous Hall effect (AHE). One proposed method for predicting the anomalous Hall performance of magnetic thin film materials through experimental trial and error requires continuous manual experimentation, which is time-consuming and inefficient. Summary of the Invention

[0003] In view of this, the present invention proposes a material performance prediction method, system, electronic device and storage medium, which combines machine learning and performance prediction of magnetic thin film materials with anomalous Hall properties. By leveraging the learning capabilities of machine learning and utilizing existing experimental data, it helps to predict magnetic thin film materials with good anomalous Hall properties, thereby saving the time cost of repeated trials, improving the performance testing efficiency of magnetic thin film materials, and enabling faster screening of high-performance magnetic thin film materials.

[0004] To achieve the above objectives, one aspect of the present invention provides a method for predicting material properties, specifically including the following steps:

[0005] Collect the primary data that affects material properties;

[0006] The first data is processed to obtain the second data;

[0007] The first data and the second data are used as sample data, and the sample data is split into a first dataset and a second dataset;

[0008] Multiple first algorithm models are trained based on the first dataset, and a second algorithm model is obtained based on the first training results;

[0009] The second algorithm model is trained based on the second dataset, and the third algorithm model is obtained based on the second training results;

[0010] The third data is input into the third algorithm model to predict the material's properties.

[0011] In some embodiments, the material is a magnetic thin film with anomalous Hall properties;

[0012] The first data includes one or more of the following: the composition ratio of different types of elements contained in the thin film, the magnetization of the element, the position of the Fermi level of the element, the number of valence electrons of the element, the relative atomic mass of the element, the test temperature, the thickness of the thin film, the aspect ratio of the thin film, the anomalous Hall resistance of the thin film, and the longitudinal resistance of the thin film.

[0013] The second data includes: anomalous Hall resistivity and anomalous Hall angle.

[0014] In some embodiments, the step of processing the first data to obtain the second data includes:

[0015] The anomalous Hall resistivity is obtained by calculating the anomalous Hall resistance and the thickness.

[0016] The longitudinal resistivity is obtained by calculating the longitudinal resistance, the thickness, and the aspect ratio.

[0017] The anomalous Hall angle is obtained by calculating the anomalous Hall resistivity and the longitudinal resistivity.

[0018] In some embodiments, the step of calculating the anomalous Hall resistance based on the thickness includes:

[0019] The anomalous Hall resistivity is obtained by calculating the anomalous Hall resistance and the thickness based on the first formula, wherein the first formula is:

[0020] The anomalous Hall resistivity = the anomalous Hall resistance × the thickness;

[0021] The step of calculating the longitudinal resistivity by considering the longitudinal resistance, the thickness, and the aspect ratio includes:

[0022] Based on the second formula, the longitudinal resistivity is calculated using the anomalous Hall resistance, the thickness, and the aspect ratio. The second formula is:

[0023] The longitudinal resistivity = the longitudinal resistance × the thickness ÷ the aspect ratio;

[0024] The step of calculating the anomalous Hall angle by comparing the anomalous Hall resistivity with the longitudinal resistivity includes:

[0025] The anomalous Hall resistivity and the longitudinal resistivity are calculated based on the third formula to obtain the anomalous Hall angle, wherein the third formula is:

[0026] The anomalous Hall angle = the anomalous Hall resistivity ÷ the longitudinal resistivity.

[0027] In some implementations, the first training result includes a first mean absolute error and a first root mean square error;

[0028] The step of obtaining the second algorithm model based on the first training result includes:

[0029] The first algorithm model whose first mean absolute error is less than a first threshold and whose first root mean square error is less than a second threshold is selected as the second algorithm model.

[0030] The second training result includes the second mean absolute error, the second root mean square error, and the second coefficient of determination.

[0031] The steps for obtaining the third algorithm model based on the second training result include:

[0032] The second algorithm model, in which the second mean absolute error is less than the third threshold and the second root mean square error is less than the fourth threshold, is selected as the third algorithm model.

[0033] In some implementations, the first training result includes a first mean absolute error, a first root mean square error, and a first coefficient of determination;

[0034] The step of obtaining the second algorithm model based on the first training result includes:

[0035] The first algorithm model is selected as the second algorithm model if the first mean absolute error is less than the first threshold, the first root mean square error is less than the second threshold, and the absolute value of the difference between the first determination coefficient and 1 is less than the fifth threshold.

[0036] The second training result includes the second mean absolute error, the second root mean square error, and the second coefficient of determination.

[0037] The steps for obtaining the third algorithm model based on the second training result include:

[0038] The second algorithm model is selected as the third algorithm model if the second mean absolute error is less than the third threshold, the second root mean square error is less than the fourth threshold, and the absolute value of the difference between the second coefficient of determination and 1 is less than the sixth threshold.

[0039] In some implementations, the step of inputting third data into the third algorithm model to predict the material's properties includes:

[0040] The third data is input into the third algorithm model, and the predicted anomalous Hall resistivity and the predicted anomalous Hall angle are output.

[0041] The properties of the material are determined based on the predicted anomalous Hall resistivity and the predicted anomalous Hall angle.

[0042] In some implementations, the first algorithm model includes any one of the following: random forest algorithm model, decision tree algorithm model, support vector machine algorithm model, and artificial neural network algorithm model;

[0043] The magnetic thin film includes a ferromagnetic thin film;

[0044] After the step of inputting the third data into the third algorithm model to predict the material's properties, the method further includes:

[0045] Compare the prediction results output by the third algorithm model with the results calculated based on the third data.

[0046] Another aspect of the present invention provides a material property prediction system, comprising:

[0047] The data acquisition module is configured to acquire first data that affects the material properties;

[0048] A data processing module, configured to process the first data to obtain the second data;

[0049] The data analysis module is configured to use the first data and the second data as sample data, and to split the sample data into a first dataset and a second dataset.

[0050] The data analysis module is further configured to train multiple first algorithm models based on the first dataset, and obtain a second algorithm model based on the first training results;

[0051] The data analysis module is also configured to train the second algorithm model based on the second dataset, and obtain a third algorithm model based on the second training results;

[0052] The data analysis module is also configured to input third data into the third algorithm model to predict the material's performance.

[0053] In another aspect of the present invention, an electronic device is provided, comprising: at least one processor; and a memory storing a computer program executable on the processor, the computer program performing the steps of the method described above when executed by the processor.

[0054] In another aspect of the present invention, a storage medium is provided, which stores a computer program that, when executed by a processor, implements the above-described method steps.

[0055] The present invention has at least the following beneficial technical effects: The solution of the present invention uses the most accurate algorithm model to predict material properties, thereby replacing the experimental trial and error method used in related technologies to test the quality of material properties, saving time and costs, improving the efficiency of material performance testing, and enabling faster screening of materials with better performance. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart of an embodiment of the material property prediction method provided by the present invention;

[0058] Figure 2 A schematic diagram of yet another embodiment of the material property prediction method provided by the present invention;

[0059] Figure 3 A schematic diagram of an embodiment of the material property prediction system provided by the present invention;

[0060] Figure 4 A schematic diagram of the structure of an embodiment of the electronic device provided by the present invention;

[0061] Figure 5 A schematic diagram of the structure of an embodiment of the storage medium provided by the present invention. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to specific examples and the accompanying drawings.

[0063] It should be noted that all uses of "first" and "second" in the embodiments of the present invention are for the purpose of distinguishing two entities or parameters with the same name but different names. It is clear that "first" and "second" are only for the convenience of expression and should not be construed as limiting the embodiments of the present invention. Subsequent embodiments will not explain this in detail.

[0064] Based on the above objectives, a first aspect of the present invention proposes a method for predicting material properties. In such... Figure 1 In the illustrated embodiment, the method includes the following steps:

[0065] S10. Collect the first data that affects material properties;

[0066] S20. Process the first data to obtain the second data;

[0067] S30. The first data and the second data are used as sample data, and the sample data is split into a first dataset and a second dataset;

[0068] S40. Train multiple first algorithm models based on the first dataset, and obtain a second algorithm model based on the first training results;

[0069] S50. Train the second algorithm model based on the second dataset, and obtain the third algorithm model based on the second training results;

[0070] S60. Input the third data into the third algorithm model to predict the material's properties.

[0071] Specifically,

[0072] In step S10, the first data on the effects on material properties is obtained through previous experiments and calculations;

[0073] In step S20, the collected first data is processed to obtain second data, which is data that can reflect the material properties;

[0074] In step S30, the first data collected and the second data obtained in the data processing stage are used as sample data for the algorithm model used in machine learning. The sample data is split into a first dataset and a second dataset, and the algorithm model is evaluated in stages based on the first dataset and the second dataset, thereby obtaining the algorithm model with the optimal material property prediction results.

[0075] In step S40, multiple first algorithm models are trained based on the first dataset, and second algorithm models are obtained based on the first training results. This selects the first batch of algorithm models with good material prediction performance, which are the second algorithm models. These models can be selected from random forest, decision tree, support vector machine, artificial neural network, and other similar models.

[0076] In step S50, the second algorithm model is trained using the second dataset, and the third algorithm model is obtained based on the second training results. Thus, the optimal algorithm model, i.e. the third algorithm model, is selected from the first batch of selected second algorithm models.

[0077] In step S60, the third data is the feature value that needs to be input into the third algorithm model. The third data is input into the third algorithm model and the prediction result is output. The quality of the material properties can be determined by the prediction result. The prediction result of materials with different component ratios can be quickly output by the third algorithm model.

[0078] The above approach uses machine learning to predict material properties, replacing the experimental trial-and-error method used in related technologies to test material properties. This saves time and improves the efficiency of material performance testing, enabling faster screening of materials with better performance.

[0079] In some embodiments, the material is a magnetic thin film with anomalous Hall properties;

[0080] The first data includes one or more of the following: the composition ratio of different types of elements contained in the thin film, the magnetization of the element, the position of the Fermi level of the element, the number of valence electrons of the element, the relative atomic mass of the element, the test temperature, the thickness of the thin film, the aspect ratio of the thin film, the anomalous Hall resistance of the thin film, and the longitudinal resistance of the thin film.

[0081] The second data includes: anomalous Hall resistivity and anomalous Hall angle.

[0082] Specifically, the solution of the present invention can be applied to the performance prediction of magnetic thin film materials with anomalous Hall properties, thereby saving time and costs, improving the performance testing efficiency of magnetic thin film materials with anomalous Hall properties, and enabling faster screening of magnetic thin film materials with better performance.

[0083] When preparing magnetic thin films (e.g., ferromagnetic thin films) with multiple elemental compositions, the relationship between performance and composition is unclear due to the large number of elemental ratios. Conventional experimental trial-and-error methods are no longer able to enumerate all the various compositional materials. Therefore, the present invention can use existing experimental data as sampling data to help predict the composition of a certain thin film with good anomalous Hall performance, thereby saving the time cost of repeated attempts and improving experimental efficiency.

[0084] The selection of sampling data has a significant impact on the prediction of material properties. The physical quantity that reflects the anomalous Hall performance is the anomalous Hall resistivity ρ. xy The larger the values ​​of the anomalous Hall angle θ, the better the anomalous Hall performance of the magnetic thin film. Ferromagnetic thin films are a type of magnetic thin film; the following explanation uses ferromagnetic thin films as an example to illustrate the selection of sampling data.

[0085] Ferromagnetic thin films typically consist of one or more of the ferromagnetic elements Fe, Co, and Ni, with other elements introduced later. Therefore, the composition ratio of different elements in a ferromagnetic thin film affects its overall properties, including magnetization; stronger magnetization results in a higher anomalous Hall resistivity. Different elemental compositions also cause Fermi level shifts, affecting the Berry curvature integral and consequently the anomalous Hall resistivity. Different compositions can also lead to disordered atomic distribution, altering the local electron environment and lattice translation symmetry, thus affecting the Berry curvature and consequently the anomalous Hall resistivity. The testing temperature primarily affects the anomalous Hall resistivity by influencing electron and phonon scattering in the ferromagnetic thin film. The relative atomic masses of different elements affect the spin-orbit coupling of the thin film material; stronger spin-orbit coupling results in a higher anomalous Hall resistivity. The number of valence electrons varies among elements; introducing a particular element into the thin film affects its overall electronic and band structure, influencing the Berry curvature integral and consequently the anomalous Hall resistivity. The thickness of a ferromagnetic thin film affects its anomalous Hall resistivity by influencing interface scattering. Smaller thickness results in greater interface scattering, but smaller thickness is not always better; there is a critical value. The aspect ratio, anomalous Hall resistance, and longitudinal resistance of the ferromagnetic thin film also affect the anomalous Hall resistivity and anomalous Hall angle.

[0086] Factors affecting the anomalous Hall performance of ferromagnetic thin films include the composition ratio of different elements in the ferromagnetic film, the magnetization of the elements, the position of the Fermi level of the elements, the number of valence electrons of the elements, the relative atomic mass of the elements, the test temperature, the thickness of the film, the aspect ratio of the film, the anomalous Hall resistance of the film, and the longitudinal resistance of the film. Therefore, the determined sampling data, i.e., the first data, can be one or more of the above-mentioned data. The more types of data included, the more accurate the prediction results of the subsequent material prediction process.

[0087] It should be understood that the embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention.

[0088] Furthermore, the process parameters for preparing the thin film also affect the anomalous Hall properties of the thin film material. Therefore, process parameters can be added to the first data. For example, annealing temperature and annealing rate affect the crystallinity of the thin film. Different grain sizes, crystal orientations, and grain boundaries will cause different electron scattering intensities, which in turn affect the anomalous Hall resistivity. Therefore, annealing temperature and annealing rate are added to the first data.

[0089] In some embodiments, the step of processing the first data to obtain the second data includes:

[0090] The anomalous Hall resistivity is obtained by calculating the anomalous Hall resistance and the thickness.

[0091] The longitudinal resistivity is obtained by calculating the longitudinal resistance, the thickness, and the aspect ratio.

[0092] The anomalous Hall angle is obtained by calculating the anomalous Hall resistivity and the longitudinal resistivity.

[0093] In some embodiments, the step of calculating the anomalous Hall resistance based on the thickness includes:

[0094] The anomalous Hall resistivity is obtained by calculating the anomalous Hall resistance and the thickness based on the first formula, wherein the first formula is:

[0095] The anomalous Hall resistivity = the anomalous Hall resistance × the thickness;

[0096] The step of calculating the longitudinal resistivity by considering the longitudinal resistance, the thickness, and the aspect ratio includes:

[0097] Based on the second formula, the longitudinal resistivity is calculated using the anomalous Hall resistance, the thickness, and the aspect ratio. The second formula is:

[0098] The longitudinal resistivity = the longitudinal resistance × the thickness ÷ the aspect ratio;

[0099] The step of calculating the anomalous Hall angle by comparing the anomalous Hall resistivity with the longitudinal resistivity includes:

[0100] The anomalous Hall resistivity and the longitudinal resistivity are calculated based on the third formula to obtain the anomalous Hall angle, wherein the third formula is:

[0101] The anomalous Hall angle = the anomalous Hall resistivity ÷ the longitudinal resistivity.

[0102] Specifically, after data collection is completed, data processing begins, based on the formula ρ. xy =R xy *t,ρ xx (Longitudinal resistivity) = R xx *t / b, θ=ρ xy / ρ xx Since the parameter reflecting the anomalous Hall performance, ρ, can be calculated, the anomalous Hall resistivity can be determined. xy And the anomalous Hall angle θ. Where t represents the film thickness, b represents the aspect ratio, and R... xy Represents the anomalous Hall resistance, R xx This represents the longitudinal resistance. Data obtained during the data acquisition and processing phases are used as sample data for subsequent data analysis.

[0103] It should be understood that the embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention.

[0104] In some implementations, the first training result includes a first mean absolute error and a first root mean square error;

[0105] The step of obtaining the second algorithm model based on the first training result includes:

[0106] The first algorithm model whose first mean absolute error is less than a first threshold and whose first root mean square error is less than a second threshold is selected as the second algorithm model.

[0107] The second training result includes the second mean absolute error, the second root mean square error, and the second coefficient of determination.

[0108] The steps for obtaining the third algorithm model based on the second training result include:

[0109] The second algorithm model, in which the second mean absolute error is less than the third threshold and the second root mean square error is less than the fourth threshold, is selected as the third algorithm model.

[0110] Specifically, such as Figure 2 As shown, firstly, the feature values ​​that need to be set when designing an algorithm model for machine learning include: the composition ratio of different elements, annealing temperature, annealing rate, film thickness, test temperature, magnetization of each element, Fermi level position, number of valence electrons, relative atomic mass, etc. The anomalous Hall resistivity ρ... xy θ is used as the target value to be predicted by the algorithm model. The target value is predicted based on the set feature values, thereby improving the accuracy of the prediction results. The selection of sample data and the setting of feature values ​​both contribute to building a good algorithm model.

[0111] Next, the existing sample data is randomly split into a first dataset and a second dataset. The ratio of the first and second datasets can be freely set based on the actual application scenario; for example, the first dataset can account for 80% and the second dataset for 20%. The sample data of the first dataset is used to train an existing algorithm model. To avoid potential errors and inaccuracies in a single algorithm model, multiple algorithm models such as random forest and decision tree are used. The training results of each algorithm model are compared to evaluate them, and the algorithm models with better prediction results are selected. The sample data of the second dataset is used to evaluate the trained algorithm model, and the optimal algorithm model is further selected. The evaluation criteria can be the root mean square error (RMSE) and the mean absolute error (MAE). The smaller the RMSE and MAE values, the more accurate the model prediction results. Then, model validation is performed by inputting the composition of the magnetic thin film material and the corresponding experimental data into the optimal algorithm model to predict the anomalous Hall performance of the magnetic thin film material.

[0112] Finally, the results predicted by the optimal algorithm model are output so that experimenters can compare the predicted results with the calculated actual anomalous resistivity and anomalous Hall angle.

[0113] It should be understood that the embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention.

[0114] The above methods can effectively predict the composition of multi-element alloy thin film materials with excellent anomalous Hall properties, saving time and improving experimental efficiency. This facilitates faster screening of material components and accelerates the application of materials in real life.

[0115] In some implementations, the first training result includes a first mean absolute error, a first root mean square error, and a first coefficient of determination;

[0116] The step of obtaining the second algorithm model based on the first training result includes:

[0117] The first algorithm model is selected as the second algorithm model if the first mean absolute error is less than the first threshold, the first root mean square error is less than the second threshold, and the absolute value of the difference between the first determination coefficient and 1 is less than the fifth threshold.

[0118] The second training result includes the second mean absolute error, the second root mean square error, and the second coefficient of determination.

[0119] The steps for obtaining the third algorithm model based on the second training result include:

[0120] The second algorithm model is selected as the third algorithm model if the second mean absolute error is less than the third threshold, the second root mean square error is less than the fourth threshold, and the absolute value of the difference between the second coefficient of determination and 1 is less than the sixth threshold.

[0121] Specifically, the first, second, third, fourth, fifth, and sixth thresholds can all be freely set based on the actual application scenario. The smaller the first, second, third, and fourth thresholds are, the more accurate the prediction results of the algorithm model will be. The closer the fifth and sixth thresholds are to 0, the more accurate the prediction results of the algorithm model will be.

[0122] This indicates that the selection process for the second algorithm model is the same as that for the third algorithm model, the difference being the sample data used; the former uses the first dataset, while the latter uses the second dataset. The following example illustrates the selection process for the third algorithm model.

[0123] The evaluation criteria used in the selection process include mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R²). 2 To comprehensively evaluate, the lower the values ​​of MAE and RMSE, the better. 2 The closer the value is to 1, the better.

[0124] The formula for calculating MAE is:

[0125] The formula for calculating RMSE is:

[0126] R 2 The calculation formula is:

[0127] Where n is the number of samples, y2 is the anomalous Hall resistivity or anomalous Hall angle predicted by the algorithm model, y1 is the actual anomalous Hall resistivity or anomalous Hall angle obtained from the experiment, and y0 is the average value of the actual anomalous Hall resistivity or anomalous Hall angle among the n samples.

[0128] Using the above formulas for calculating MAE, RMSE, and R... 2 The calculation formulas can ultimately yield the MAE, RMSE, and R corresponding to the anomalous Hall resistivity and anomalous Hall angle, respectively. 2 Total R 2 R is the value of anomalous Hall resistivity and anomalous Hall angle. 2 The average of the sums of values. This is achieved by comparing the MAE, RMSE, and total R-value for each model. 2 The optimal algorithm model is selected based on the value.

[0129] It should be understood that the embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention.

[0130] In some implementations, the step of inputting third data into the third algorithm model to predict the material's properties includes:

[0131] The third data is input into the third algorithm model, and the predicted anomalous Hall resistivity and the predicted anomalous Hall angle are output.

[0132] The properties of the material are determined based on the predicted anomalous Hall resistivity and the predicted anomalous Hall angle.

[0133] In some implementations, the first algorithm model includes any one of the following: random forest algorithm model, decision tree algorithm model, support vector machine algorithm model, and artificial neural network algorithm model;

[0134] The magnetic thin film includes a ferromagnetic thin film;

[0135] After the step of inputting the third data into the third algorithm model to predict the material's properties, the method further includes:

[0136] Compare the prediction results output by the third algorithm model with the results calculated based on the third data.

[0137] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 3 As shown, embodiments of the present invention also provide a material property prediction system, comprising:

[0138] Data acquisition module 110, the data acquisition module 110 is configured to acquire first data affecting material properties;

[0139] Data processing module 120, configured to process the first data to obtain second data;

[0140] Data analysis module 130 is configured to use the first data and the second data as sample data, and to split the sample data into a first dataset and a second dataset;

[0141] The data analysis module 130 is further configured to train multiple first algorithm models based on the first dataset, and obtain a second algorithm model based on the first training results;

[0142] The data analysis module 130 is further configured to train the second algorithm model based on the second dataset, and obtain a third algorithm model based on the second training result;

[0143] The data analysis module 130 is also configured to input third data into the third algorithm model to predict the performance of the material.

[0144] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 4 As shown, an embodiment of the present invention also provides an electronic device 30, which includes a processor 310 and a memory 320. The memory 320 stores a computer program 321 that can run on the processor. When the processor 310 executes the program, it performs the steps of the method described above.

[0145] The memory, as a non-volatile storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the material property prediction method described in the embodiments of this application. The processor executes various system functions and data processing by running the non-volatile software programs, instructions, and modules stored in the memory, thereby implementing the material property prediction method described in the above embodiments.

[0146] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on system usage. Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the local module via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0147] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 5 As shown, an embodiment of the present invention also provides a storage medium 40, which stores a computer program 410 that executes the above method when executed by a processor.

[0148] Finally, it should be noted that those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium for the program can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The above computer program embodiments can achieve the same or similar effects as any of the corresponding foregoing method embodiments.

[0149] Those skilled in the art will also understand that the various exemplary logic blocks, modules, circuits, and algorithm steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the functionality of various illustrative components, blocks, modules, circuits, and steps has been generally described. Whether this functionality is implemented as software or as hardware depends on the specific application and the design constraints imposed on the system as a whole. Those skilled in the art can implement the functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the embodiments disclosed herein.

[0150] The above are exemplary embodiments disclosed in this invention. However, it should be noted that various changes and modifications can be made without departing from the scope of the embodiments of this invention as defined by the claims. The functions, steps, and / or actions of the methods according to the disclosed embodiments described herein do not need to be performed in any particular order. The sequence numbers of the disclosed embodiments of this invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. Furthermore, although the elements disclosed in the embodiments of this invention may be described or claimed individually, they may be understood as multiple unless explicitly limited to a singular number.

[0151] It should be understood that, as used herein, the singular form “a” is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, “and / or” refers to any and all possible combinations of one or more of the associated listed items.

[0152] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of the invention, technical features of the above embodiments or different embodiments can be combined, and many other variations of different aspects of the invention exist, which are not provided in the details for the sake of brevity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

Claims

1. A material performance prediction method, characterized by, include: Collect the first data affecting the performance of magnetic thin film materials; The magnetic thin film has anomalous Hall properties; the first data includes the composition ratio of different types of elements contained in the thin film, the magnetization of the elements, the position of the Fermi level of the elements, the number of valence electrons of the elements, the relative atomic mass of the elements, the test temperature, the thickness of the thin film, the aspect ratio of the thin film, the anomalous Hall resistance of the thin film, and the longitudinal resistance of the thin film. The first data is processed to obtain an anomalous Hall resistivity and an anomalous Hall angle. The processing of the first data to obtain the anomalous Hall resistivity and anomalous Hall angle includes: calculating the anomalous Hall resistance and the thickness to obtain the anomalous Hall resistivity; calculating the longitudinal resistance, the thickness, and the aspect ratio to obtain the longitudinal resistivity; and calculating the anomalous Hall resistivity and the longitudinal resistivity to obtain the anomalous Hall angle. The first data, the anomalous Hall resistivity, and the anomalous Hall angle are used as sample data, and the sample data is split into a first dataset and a second dataset. Based on the first dataset, multiple first algorithm models are trained to obtain a first mean absolute error and a first root mean square error. The first algorithm model whose first mean absolute error is less than a first threshold and whose first root mean square error is less than a second threshold is selected as the second algorithm model. The second algorithm model is trained based on the second dataset to obtain the second mean absolute error and the second root mean square error. The second algorithm model with the second mean absolute error less than the third threshold and the second root mean square error less than the fourth threshold is selected as the third algorithm model. The third data is input into the third algorithm model to predict the anomalous Hall performance of the magnetic thin film material. The process of inputting the third data into the third algorithm model to predict the anomalous Hall performance of the magnetic thin film material includes: inputting the third data into the third algorithm model and outputting the predicted anomalous Hall resistivity and the predicted anomalous Hall angle; and determining the performance of the material based on the predicted anomalous Hall resistivity and the predicted anomalous Hall angle.

2. The method of claim 1, wherein, The step of calculating the anomalous Hall resistivity by relating the anomalous Hall resistance to the thickness includes: The anomalous Hall resistivity is obtained by calculating the anomalous Hall resistance and the thickness based on the first formula, wherein the first formula is: The anomalous Hall resistivity = the anomalous Hall resistance × the thickness; The step of calculating the longitudinal resistivity by considering the longitudinal resistance, the thickness, and the aspect ratio includes: Based on the second formula, the longitudinal resistivity is calculated using the anomalous Hall resistance, the thickness, and the aspect ratio. The second formula is: The longitudinal resistivity = the longitudinal resistance × the thickness ÷ the aspect ratio; The step of calculating the anomalous Hall angle by comparing the anomalous Hall resistivity with the longitudinal resistivity includes: The anomalous Hall resistivity and the longitudinal resistivity are calculated based on the third formula to obtain the anomalous Hall angle, wherein the third formula is: The anomalous Hall angle = the anomalous Hall resistivity ÷ the longitudinal resistivity.

3. The method according to any one of claims 1 or 2, characterized in that, Training multiple first algorithm models based on the first dataset to obtain the first mean absolute error and the first root mean square error includes: training multiple first algorithm models based on the first dataset to obtain the first mean absolute error, the first root mean square error, and the first coefficient of determination; The step of selecting the first algorithm model whose first mean absolute error is less than a first threshold and whose first root mean square error is less than a second threshold as the second algorithm model includes: The first algorithm model is selected as the second algorithm model if the first mean absolute error is less than the first threshold, the first root mean square error is less than the second threshold, and the absolute value of the difference between the first determination coefficient and 1 is less than the fifth threshold. The second algorithm model is trained based on the second dataset to obtain the second mean absolute error and the second root mean square error, including: The second algorithm model is trained based on the second dataset to obtain the second mean absolute error, the second root mean square error, and the second coefficient of determination. The step of selecting the second algorithm model, in which the second mean absolute error is less than the third threshold and the second root mean square error is less than the fourth threshold, as the third algorithm model includes: The second algorithm model is selected as the third algorithm model if the second mean absolute error is less than the third threshold, the second root mean square error is less than the fourth threshold, and the absolute value of the difference between the second coefficient of determination and 1 is less than the sixth threshold.

4. The method of any one of claims 1, wherein, The first algorithm model includes any one of the following: random forest algorithm model, decision tree algorithm model, support vector machine algorithm model, and artificial neural network algorithm model; The magnetic thin film includes a ferromagnetic thin film; After the step of inputting the third data into the third algorithm model to predict the anomalous Hall performance of the magnetic thin film material, the method further includes: Compare the prediction results output by the third algorithm model with the results calculated based on the third data.

5. A material performance prediction system, characterized by, include: The data acquisition module is configured to acquire first data affecting the performance of magnetic thin film materials; The magnetic thin film has anomalous Hall properties; the first data includes the composition ratio of different types of elements contained in the thin film, the magnetization of the elements, the position of the Fermi level of the elements, the number of valence electrons of the elements, the relative atomic mass of the elements, the test temperature, the thickness of the thin film, the aspect ratio of the thin film, the anomalous Hall resistance of the thin film, and the longitudinal resistance of the thin film. The data processing module is configured to process the first data to obtain an anomalous Hall resistivity and an anomalous Hall angle. The process of processing the first data to obtain the anomalous Hall resistivity and anomalous Hall angle includes: calculating the anomalous Hall resistivity by combining the anomalous Hall resistance with the thickness; calculating the longitudinal resistivity by combining the longitudinal resistance with the thickness with the aspect ratio; and calculating the anomalous Hall angle by combining the anomalous Hall resistivity with the longitudinal resistivity. The data analysis module is configured to use the first data, the anomalous Hall resistivity, and the anomalous Hall angle as sample data, and to split the sample data into a first dataset and a second dataset. The data analysis module is further configured to train multiple first algorithm models based on the first dataset to obtain a first mean absolute error and a first root mean square error, and select the first algorithm model whose first mean absolute error is less than a first threshold and whose first root mean square error is less than a second threshold as the second algorithm model. The data analysis module is further configured to train the second algorithm model based on the second dataset to obtain the second mean absolute error and the second root mean square error, and select the second algorithm model whose second mean absolute error is less than the third threshold and whose second root mean square error is less than the fourth threshold as the third algorithm model. The data analysis module is further configured to input third data into the third algorithm model to predict the anomalous Hall performance of the magnetic thin film material. This inputting third data into the third algorithm model includes: inputting the third data into the third algorithm model and outputting the predicted anomalous Hall resistivity and the predicted anomalous Hall angle; and determining the material's performance based on the predicted anomalous Hall resistivity and the predicted anomalous Hall angle.

6. An electronic device, comprising: At least one processor; as well as A memory storing a computer program executable on the processor, characterized in that the processor executes the program and performs the steps of the method as described in any one of claims 1 to 4.

7. A storage medium storing a computer program, characterized by When the computer program is executed by a processor, it performs the steps of the method as described in any one of claims 1 to 4.

Images