A multi-dimensional visual analysis method based on titanium alloy composition and processing technology

Abstract

This invention relates to the fields of visualization analysis and materials science, and provides a multi-dimensional visualization analysis method based on the composition and processing technology of titanium alloys. The main aim is to solve the problems of difficult data integration, insufficient visualization effects, and low prediction accuracy. The method includes data collection and preprocessing, model training and evaluation, multi-dimensional visualization, and performance prediction and optimization analysis. The data collection and preprocessing step involves collecting titanium alloy data from multiple sources, cleaning and normalizing it to ensure data quality. The model training and evaluation stage uses machine learning algorithms to train the feature data and evaluates the model using leave-one-out cross-validation. The multi-dimensional visualization utilizes various tools such as scatter plots, heatmaps, and parallel coordinate plots to visually display data relationships. The performance prediction and optimization analysis, based on the trained model and multi-dimensional visualization tools, predicts and optimizes the properties of titanium alloys, seeking the optimal combination of composition and processing technology.

Description

Technical Field

[0001] This invention relates to the fields of visualization analysis and materials science, and provides a multi-dimensional visualization analysis method based on the composition and processing technology of titanium alloys. Background Technology

[0002] Titanium alloys are widely used in aerospace, medical devices, and chemical industries due to their superior properties. The properties of titanium alloys are influenced by their composition and processing techniques, and the relationships between these factors are complex and difficult to visualize. Traditional titanium alloy design and optimization primarily rely on experimental trial-and-error methods, which are time-consuming, labor-intensive, and costly. However, with continuous technological advancements, data analysis and machine learning techniques are increasingly being applied to materials science to predict material properties and optimize material design.

[0003] Currently, some studies have applied data analysis and machine learning to the field of materials science. For example, regression analysis and neural network algorithms are used to predict material properties, and principal component analysis and clustering algorithms are used to classify and optimize material composition and processing techniques. The implementation steps of these technical solutions are roughly as follows: data collection, data preprocessing, model training, model evaluation, and optimization analysis. However, existing technical solutions have the following drawbacks: Data integration difficulties: Data on titanium alloy composition, processing technology, and performance come from diverse sources, making integration challenging, and existing tools lack dedicated processing modules; Insufficient visualization effects: Existing technologies mostly use general-purpose data visualization tools, lacking specific designs for titanium alloy composition and processing technology. The performance of titanium alloys is influenced by various microstructural factors and chemical compositions, and these factors exhibit highly nonlinear relationships. This makes traditional visualization methods unable to fully express all information and potential patterns in the data, and difficult to intuitively display complex multidimensional data relationships; Low prediction accuracy: Due to the limited data samples in materials science, the prediction accuracy and generalization ability of existing models are limited, making it difficult to meet practical needs. Data augmentation techniques are needed to expand the limited dataset, deeply mine and construct features with high predictive power, combine domain knowledge for feature selection and extraction, and combine the advantages of multiple machine learning models through ensemble learning to improve the model's robustness and prediction accuracy, thereby improving the model's training effect and generalization ability. Summary of the Invention

[0004] This invention solves the technical problems of data integration difficulties, insufficient visualization effects, and low prediction accuracy in the prediction and optimization analysis of titanium alloy properties.

[0005] To achieve the above objectives, the present invention employs the following technical means:

[0006] Because the present invention employs the above-mentioned technical means, it has the following beneficial effects:

[0007] 1. Systematized Data Collection and Preprocessing Steps: The data collection and preprocessing steps in this invention are systematically designed to ensure data quality and analytical accuracy. First, data on the composition, processing technology, and properties of titanium alloys are collected from multiple sources, such as literature, experiments, and the Matminer database. Then, the collected data is cleaned, including noise removal, handling missing and outlier values, to ensure accuracy. Next, the dataset is normalized to the (0, 1) interval, making data from different dimensions comparable. This series of steps ensures that the dataset used for modeling is of high quality and accuracy, laying a solid foundation for subsequent analysis and model training.

[0008] 2. Feature Selection and Model Training: This invention improves the predictive accuracy and stability of the model by combining Pearson correlation coefficients to select features highly correlated with performance indicators and utilizing advanced machine learning algorithms for model training. First, correlation coefficient analysis is performed on the preprocessed dataset to select features highly correlated with yield strength or tensile strength. Then, various machine learning algorithms, such as random forest, XGBoost, and gradient boosting regression, are used to train the model on the feature data. Leave-one-out cross-validation is used to evaluate the predictive accuracy and stability of different models, and the optimal model is selected based on the evaluation results. This process not only improves the model's predictive accuracy but also enhances its generalization ability.

[0009] 3. Multi-dimensional Visualization: This invention employs various visualization techniques, such as scatter plots, heatmaps, and parallel coordinate plots, to intuitively display the complex relationships between titanium alloy composition, processing technology, and performance, enhancing the user's data analysis experience and efficiency. For example, through dimensionality-reduced scatter plots derived from principal component analysis, users can intuitively see the clustering of different data samples; heatmaps can demonstrate the correlation between different features; and parallel coordinate plots help users analyze data from multiple dimensions. The use of these visualization tools makes complex data relationships intuitive and easy to understand, greatly improving the user's data analysis efficiency.

[0010] 4. Performance Prediction and Optimization Analysis: This invention not only predicts performance based on a trained model, but also uses multi-dimensional visualization tools and optimization algorithms to find the optimal combination of composition and processing technology, providing a scientific basis for improving the application performance of titanium alloys. Users can input new titanium alloy composition and processing parameters and use the trained model to perform performance prediction. Then, new scatter plots and line graphs are generated using multi-dimensional visualization tools. By combining the intervals of the predicted performance values ​​in the line graphs with the scatter plots showing similar performance within the same cluster, features with positive or negative correlations to performance can be identified. Based on this information, users can try adding components that are positively correlated with performance or reducing components that are negatively correlated, forming new combinations, and then re-perform performance prediction and comparative analysis. Through multiple iterations, users can finally determine the optimal combination of titanium alloy composition and processing technology to achieve the expected performance indicators. This process not only improves the interactivity and operability of the analysis, but also provides a scientific basis for improving the application performance of titanium alloys. Attached Figure Description

[0011] Figure 1 This is a flowchart of the present invention;

[0012] Figure 2 This is a screenshot of the interface.

[0013] Where a is the control area, b is the data sample cluster scatter plot, c is the custom scatter style, d is the custom scatter card, e is the performance data classification line chart, f is the table of specific values ​​of the evaluation indicators, and g is the evaluation indicator radar chart. Detailed Implementation

[0014] The embodiments of the present invention will be described in detail below. Although the present invention will be described and illustrated in conjunction with some specific embodiments, it should be noted that the present invention is not limited to these embodiments. On the contrary, any modifications or equivalent substitutions made to the present invention should be covered within the scope of the claims of the present invention.

[0015] Furthermore, to better illustrate the present invention, numerous specific details are set forth in the following detailed embodiments. Those skilled in the art will understand that the present invention can be practiced without these specific details.

[0016] A multi-dimensional visualization analysis tool based on titanium alloy composition and processing technology includes the following steps:

[0017] Step 1, Data Collection and Preprocessing: After collecting titanium alloy data from multiple sources, the data is cleaned and normalized to construct a high-quality dataset suitable for modeling.

[0018] Step 2, Model Training and Evaluation: Extract effective features, use machine learning algorithms to train the model on the feature data, and evaluate and optimize the model.

[0019] Step 3, Multi-dimensional visualization: Using a self-designed multi-dimensional visualization tool, the composition, processing technology and performance data of titanium alloys are displayed in multiple dimensions to help users intuitively analyze data relationships.

[0020] Step 4, Performance Prediction and Optimization Analysis: Based on the trained model and multi-dimensional visualization tools, the performance of titanium alloys is predicted and optimized to find the optimal combination of titanium alloy composition and processing technology.

[0021] Step 1 above specifically includes the following steps:

[0022] Step 1.1: Search and collect data on the composition, processing technology and performance of titanium alloys from literature, experiments and Matminer database, and organize and construct a dataset suitable for modeling.

[0023] Step 1.2: Clean the constructed dataset, including removing noise, handling missing values ​​and outliers, to ensure data accuracy;

[0024] Step 1.3: Normalize the dataset to the (0, 1) interval to make data from different dimensions comparable.

[0025] Step 2 above specifically includes the following steps:

[0026] Step 2.1 Feature Selection: For the dataset obtained after preprocessing in Step 1, use the Pearson correlation coefficient to select features that are highly correlated with the performance indicators to obtain the feature dataset;

[0027] Step 2.1.1: For the yield strength of titanium alloy, a correlation coefficient analysis was performed to determine 38 characteristic parameters with V alloy composition as the main feature set for the yield strength performance prediction model.

[0028] Step 2.1.1.1: Calculate the Pearson correlation coefficient between each feature in the dataset and the yield strength of the titanium alloy;

[0029] Step 2.1.1.2: Select features with an absolute value of Pearson correlation coefficient greater than 0.5 and determine them as features highly correlated with yield strength;

[0030] Step 2.1.1.3: Select 38 characteristic parameters, mainly V alloy composition, from Step 2.1.1.2 to construct the feature set of the yield strength performance prediction model;

[0031] Step 2.1.2: For the tensile strength of titanium alloy, correlation coefficient analysis was performed to determine 44 characteristic parameters with Fe alloy composition as the main feature set as the feature set of the tensile strength performance prediction model;

[0032] Step 2.1.2.1: Calculate the Pearson correlation coefficient between each feature in the dataset and the tensile strength of the titanium alloy;

[0033] Step 2.1.2.2: Select features with an absolute value of Pearson correlation coefficient greater than 0.5 and identify them as features highly correlated with tensile strength;

[0034] Step 2.1.2.3: Select 44 characteristic parameters, mainly Fe alloy composition, from Step 2.1.2.2 to construct the feature set of the tensile strength performance prediction model;

[0035] Step 2.1.3: Integrate the feature sets obtained in Step 2.1.1 and Step 2.1.2 to form two independent feature datasets, which are used for subsequent training of yield strength and tensile strength performance prediction models, respectively.

[0036] Step 2.2: Divide the feature dataset obtained in Step 2.1 into a training set and a test set in a 9:1 ratio;

[0037] Step 2.3: Construct a set of base learners, consisting of the following three parts: a random forest regressor with 50 trees, an XGBoost regressor with 50 iterations, and a gradient boost regressor with 550 iterations and a maximum depth of 5. The predictions generated by the base learners will be used to construct the meta-feature matrix.

[0038] Step 2.4: Construct a meta-learner, using the linear regression algorithm as the meta-learner, with the meta-features generated by the base learner in Step 2.3 as the input;

[0039] Step 2.5: Construct the “Superlearner” model, which consists of the base learner from Step 2.3 and the meta-learner from Step 2.4;

[0040] Step 2.6: Select a variety of machine learning algorithms for model training, including random forest, XGBoost, gradient boosting regression, decision tree, support vector machine, multiple linear regression, ridge regression based on kernel tricks, and the "Superlearner" model;

[0041] Step 2.7: Evaluate the prediction accuracy and stability of different models using leave-one-out cross-validation, and obtain the R-values ​​of each model after prediction. 2The values ​​of four evaluation indicators, MAEf, RMSE and EVS, are used to compare the performance of each model on the above evaluation indicators. Radar charts and tables are drawn to intuitively show the performance of different models. For the yield strength performance prediction model, the "Superlearner" model with the best performance on all indicators is selected to predict the yield strength performance of alloy materials. For the tensile strength performance prediction model, the support vector machine (SVR) model with the best performance on all indicators is selected to predict the tensile strength performance of alloy materials.

[0042] Step 2.8: Conduct further experimental verification and evaluation of the model determined in Step 2.7, optimize the model based on the evaluation results, and adjust the parameters to improve the model performance.

[0043] The technical problems solved in steps 2.1-2.8, and the results:

[0044] 1. This study solves the problem of feature selection when predicting the performance of titanium alloy materials. By combining Pearson correlation coefficient analysis, a set of features that are highly correlated with different performance indicators is selected.

[0045] 2. It solves the problem of how to comprehensively evaluate the predictive performance of the model. It uses multiple machine learning algorithms to train the model, which can comprehensively evaluate the performance of each algorithm on different performance indicators, thereby selecting the optimal algorithm. This multi-algorithm training method can identify patterns and relationships that a single algorithm may overlook, thereby improving the overall predictive performance and stability of the model.

[0046] 3. This study addressed the problem of how to objectively evaluate and improve the predictive performance of the model, established evaluation indicators for the performance prediction model of titanium alloy materials, and utilized R... 2 The four evaluation metrics—MAE, RMSE, and EVS—can evaluate the model's predictive accuracy, stability, and bias from different perspectives. Combined with tabular views and radar charts, a comprehensive understanding of the model's performance can be obtained. These metrics help identify the strengths and weaknesses of models, guide the selection of the optimal predictive model, and ensure the reliability and effectiveness of the model in practical applications.

[0047] Step 3 above specifically includes the following steps:

[0048] Step 3.1: Construct a scatter plot, use principal component analysis to reduce the dimensionality of the feature dataset, and then use the K-means algorithm to cluster it into three different clusters. Each cluster represents a set of data with similar properties, and each point represents a data sample. The similarity between scatter points is measured by the distance between their physical locations.

[0049] Step 3.2: Based on Step 3.1, customize the scatter plot style. Each scatter plot consists of four parts: the outermost 38 gray bars represent the features most related to yield strength, the middle 44 black bars represent the features most related to tensile strength, and the length of the bars indicates the magnitude of the values. The two arc blocks at the bottom of the outermost layer represent yield strength and tensile strength, and their relative size is indicated by their length and thickness. The central circle displays the titanium alloy chemical formula of the current scatter plot, and its different categories are indicated by pink, green and blue.

[0050] Step 3.3: Based on Steps 3.1 and 3.2, implement the display of custom scatter cards by clicking on scatter points. Each card contains two parts: one is to use bar charts of different colors to represent the features most related to yield strength and tensile strength, respectively; the other is to display the chemical formula of the titanium alloy represented by the current card and the specific parameter values ​​of the processing technology that affects the performance. Different colored cards represent different clusters, which facilitates comprehensive evaluation and horizontal and vertical comparison.

[0051] Step 3.4: Create a line chart and classify the performance data into three categories. Use a blue line to represent the yield strength and a yellow line to represent the tensile strength. Divide the values ​​into three intervals: [0, 800], [800, 1600], and [1600, 2600], which correspond to the three endpoints of 0, 1, and 2, respectively. The endpoint of the line corresponds to its corresponding value range.

[0052] Step 3.5: Based on steps 3.1, 3.2 and 3.4, implement the display of a magnified scatter point style when hovering over the scatter point and mark two corresponding polylines in red on the plane coordinate graph;

[0053] Step 3.6: Create a table showing the specific values ​​of the machine learning model used in Step 2.3 on the four evaluation metrics of R2, MAEf, EVS and RMSE, to visually demonstrate and compare the model's performance;

[0054] Step 3.7: Draw a radar chart based on Step 3.6. The four vertices of the radar chart represent the four evaluation indicators in Step 3.6. The projection axis from the center to the four vertices represents the scale from 0 to 1. To maintain consistency, the original values ​​at vertices MAEf and RMSE are subtracted from 1 to obtain the projection results. Different colors are used to represent different models. Each model is an irregular quadrilateral. The closer it is to a complete quadrilateral, the better the model performance.

[0055] The technical problems solved in steps 3.1-3.7, and the results:

[0056] 1. It solves the problem of how to reduce dimensionality and cluster feature datasets. Principal Component Analysis (PCA) is used to reduce the dimensionality of the dataset and simplify the data structure. The K-means algorithm is used to divide the data into three clusters, so that data samples with similar properties are grouped together.

[0057] 2. It solves the problem of how to intuitively display the characteristics related to yield strength and tensile strength. By customizing the scatter plot style, it displays the correlation and magnitude of each characteristic with the performance index, providing intuitive visual information. At the same time, it can find the commonalities and differences between different scatter plots based on their distribution.

[0058] 3. It solves the problem of how to display the specific information of each data sample in detail and intuitively. By using custom scatter cards, it displays the characteristics related to performance indicators, the chemical formula of titanium alloys and processing parameters, which facilitates comprehensive evaluation and comparison.

[0059] 4. It solves the problem of how to classify and visualize performance data. By designing line charts, the yield strength and tensile strength data are classified into three categories and displayed through color and numerical range. At the same time, the line charts are linked with scatter plots to enhance the correlation between data and the visualization effect.

[0060] 5. Solved the problem of how to comprehensively evaluate and compare the performance of different models, using tables to visually display the performance of each machine learning model in R. 2 The specific values ​​of MAE, EVS, and RMSE metrics are provided to facilitate comparison of model performance; radar charts are used to display the performance of each model on the four evaluation metrics, with models that are closer to a complete quadrilateral performing better, providing a clear visual comparison.

[0061] Step 4 above specifically includes the following steps:

[0062] Step 4.1: Input the new titanium alloy composition and processing parameters, and use the model trained in Step 2 to perform performance prediction and obtain the prediction results;

[0063] Step 4.2: Generate new scatter plots and line plots using the multidimensional visualization analysis tool built in Step 3. Combine the intervals to which the performance prediction values ​​belong in the line plots with the scatter plots, and compare the scatter plots with similar performance within the same clusters to find the features that are positively or negatively correlated with performance, thereby obtaining optimization schemes that can improve performance.

[0064] Step 4.3: Based on the optimization scheme obtained in Step 4.2, try to increase components that are positively correlated with performance or decrease components that are negatively correlated to form new combinations;

[0065] Step 4.4: Use the new combination obtained in Step 4.3 as input for Step 4.1 to obtain new prediction results. Compare these new prediction results with the original prediction results. Based on the comparison results, analyze and evaluate the effectiveness of the improvement scheme. Iterate and optimize the titanium alloy composition and processing parameters repeatedly to determine the optimal combination of titanium alloy composition and processing technology, thereby achieving the expected performance indicators.

[0066] The technical problems solved in steps 4.1-4.4, and the results:

[0067] 1. This paper addresses the problem of predicting the performance of titanium alloy materials. By using customized visualization tools to analyze the performance prediction results, and by comparing scatter plots and line graphs generated within the same cluster with similar performance, the paper identifies features that have a positive or negative correlation with the performance and proposes optimization schemes.

[0068] 2. The problem of how to optimize the performance of titanium alloy materials was solved. Based on the positive and negative correlation between performance and the model, the composition of titanium alloy and processing parameters were adjusted to form a new combination. The new combination was re-input into the model to obtain new prediction results. The results were compared with the original prediction results. The effectiveness of the improvement scheme was continuously analyzed and evaluated, and the scheme was repeatedly adjusted to achieve the expected performance indicators.

[0069] In summary, the present invention has the following advantages over the prior art:

[0070] 1. Data Processing and Visualization Tools Specifically for Titanium Alloys: This invention provides a set of data processing and visualization tools specifically designed for titanium alloys. These tools can intuitively display complex multidimensional data relationships. Traditional data visualization tools often lack customized designs for specific materials (such as titanium alloys), resulting in poor performance when displaying the complex relationships between titanium alloy composition, processing technology, and properties. The tools in this invention, such as scatter plots, heatmaps, and parallel coordinate plots, are designed based on the characteristics of titanium alloys, enabling a better display of the relationships between their composition, processing technology, and properties. For example, scatter plots can distinguish different titanium alloy compositions and processing technologies using different colors and shapes of markers; heatmaps can show the correlation thermal distribution between different features; and parallel coordinate plots can simultaneously display data features from multiple dimensions. The use of these tools makes complex data relationships intuitive and easy to understand, greatly improving the user's data analysis efficiency.

[0071] 2. Efficient Model Training and Evaluation Methods: This invention employs efficient model training and evaluation methods, which improve the model's predictive accuracy and stability. During the model training phase, this invention utilizes various advanced machine learning algorithms, such as Random Forest, XGBoost, and Gradient Boosting Regression, which can effectively learn complex patterns from data. During the model evaluation phase, this invention employs leave-one-out cross-validation, which evaluates the model's predictive accuracy and stability by using each sample sequentially as the validation set and the remaining samples as the training set. This evaluation method comprehensively assesses the model's performance, avoiding overfitting and underfitting issues. Through this efficient model training and evaluation method, this invention obtains a model with both high predictive accuracy and stability.

[0072] 3. Comprehensive Data Processing Flow: This invention employs a comprehensive data processing flow to ensure data quality and analytical accuracy. First, in the data collection phase, this invention gathers a large amount of titanium alloy data from multiple sources, including composition, processing technology, and performance data. Then, in the data preprocessing phase, this invention cleans and normalizes the collected data to ensure its quality. Next, in the feature selection phase, this invention selects features with high correlation to performance indicators using the Pearson correlation coefficient. Finally, in the model training and evaluation phase, this invention uses various machine learning algorithms for model training and evaluates the model's predictive accuracy and stability using leave-one-out cross-validation. This comprehensive data processing flow ensures data quality and analytical accuracy.

[0073] 4. Performance Prediction and Optimization Analysis: This invention not only predicts performance based on a trained model but also uses multi-dimensional visualization tools and optimization algorithms to find the optimal combination of composition and processing technology. In the performance prediction stage, users can input new titanium alloy composition and processing parameters and use the trained model to predict performance. Then, in the optimization analysis stage, users can generate new scatter plots and line graphs using multi-dimensional visualization tools. By combining the intervals of the predicted performance values ​​in the line graphs with the scatter plots showing similar performance within the same cluster, users can identify features that are positively or negatively correlated with performance. Based on this information, users can try adding components that are positively correlated with performance or reducing negatively correlated components to form new combinations and then re-perform performance prediction and comparative analysis. Through multiple iterations, users can ultimately determine the optimal combination of titanium alloy composition and processing technology to achieve the expected performance indicators. This process not only improves the interactivity and operability of the analysis but also provides a scientific basis for improving the application performance of titanium alloys.

Claims

1. A multi-dimensional visualization analysis method based on titanium alloy composition and processing technology, characterized in that, Includes the following steps: Step 1, Data Collection and Preprocessing: After collecting titanium alloy data from multiple sources, the data is cleaned and normalized to construct a high-quality dataset suitable for modeling; Step 2, Model Training and Evaluation: Extract effective features, use machine learning algorithms to train the model on the feature data, and evaluate and optimize the model; Step 3, Multi-dimensional visualization: Using a self-designed multi-dimensional visualization tool, the composition, processing technology and performance data of titanium alloys are displayed in multiple dimensions to help users intuitively analyze data relationships; Step 4, Performance Prediction and Optimization Analysis: Based on the trained model and multi-dimensional visualization tools, the performance of titanium alloys is predicted and optimized to find the optimal combination of titanium alloy composition and processing technology. Step 3 specifically includes the following steps: Step 3.1: Construct a scatter plot, use principal component analysis to reduce the dimensionality of the feature dataset, and then use the K-means algorithm to cluster it into three different clusters. Each cluster represents a set of data with similar properties, and each point represents a data sample. The similarity between scatter points is measured by the distance between their physical locations. Step 3.2: Based on Step 3.1, customize the scatter plot style. Each scatter plot consists of four parts: the outermost 38 gray bars represent the features most related to yield strength, the middle 44 black bars represent the features most related to tensile strength, and the length of the bars indicates the magnitude of the values. The two arc blocks at the bottom of the outermost layer represent yield strength and tensile strength, and their relative size is indicated by their length and thickness. The central circle displays the titanium alloy chemical formula of the current scatter plot, and its different categories are indicated by pink, green and blue. Step 3.3: Based on Steps 3.1 and 3.2, implement the display of custom scatter cards by clicking on scatter points. Each card contains two parts: one is to use bar charts of different colors to represent the features most related to yield strength and tensile strength, respectively; the other is to display the chemical formula of the titanium alloy represented by the current card and the specific parameter values ​​of the processing technology that affects the performance. Different colored cards represent different clusters, which facilitates comprehensive evaluation and horizontal and vertical comparison. Step 3.4: Create a line chart to classify the performance data into three categories. Use a blue line to represent yield strength and a yellow line to represent tensile strength. Divide the values ​​into three intervals: [0, 800], [800, 1600], and [1600, 2600], which correspond to the three endpoints of 0, 1, and 2, respectively. The endpoint of the line corresponds to its corresponding value range. Step 3.5: Based on steps 3.1, 3.2 and 3.4, implement the display of a magnified scatter point style when hovering over the scatter point and mark two corresponding polylines in red on the plane coordinate graph; Step 3.6: Create a table showing the specific values ​​of the machine learning model used in Step 2.3 on the four evaluation metrics of R2, MAEf, EVS and RMSE, to visually demonstrate and compare the model's performance; Step 3.7: Draw a radar chart based on Step 3.

6. The four vertices of the radar chart represent the four evaluation indicators in Step 3.

6. The projection axis from the center to the four vertices represents the scale from 0 to 1. To maintain consistency, the original values ​​at vertices MAEf and RMSE are subtracted from 1 to obtain the projection results. Different colors are used to represent different models. Each model is an irregular quadrilateral. The closer it is to a complete quadrilateral, the better the model performance.

2. The method according to claim 1, characterized in that, Step 1 specifically includes the following steps: Step 1.1: Query and collect data on the composition, processing technology and performance of titanium alloys from literature, experiments and Matminer database, and organize and construct a dataset suitable for modeling; Step 1.2: Clean the constructed dataset, including removing noise, handling missing values ​​and outliers, to ensure data accuracy; Step 1.3: Normalize the dataset to the (0, 1) interval to make data from different dimensions comparable.

3. The method according to claim 1, characterized in that, Step 2 specifically includes the following steps: Step 2.1 Feature Selection: For the dataset obtained after preprocessing in Step 1, use the Pearson correlation coefficient to select features that are highly correlated with the performance indicators to obtain the feature dataset; Step 2.2: Divide the feature dataset obtained in Step 2.1 into a training set and a test set in a 9:1 ratio; Step 2.3: Construct a set of base learners, consisting of the following three parts: a random forest regressor with 50 trees, an XGBoost regressor with 50 iterations, and a gradient boost regressor with 550 iterations and a maximum depth of 5. The predictions generated by the base learners will be used to construct the meta-feature matrix. Step 2.4: Construct a meta-learner, using the linear regression algorithm as the meta-learner, with the meta-features generated by the base learner in Step 2.3 as the input; Step 2.5: Construct the "Superlearner" model, which consists of the base learner from Step 2.3 and the meta-learner from Step 2.4; Step 2.6: Select a variety of machine learning algorithms for model training, including random forest, XGBoost, gradient boosting regression, decision tree, support vector machine, multiple linear regression, ridge regression based on kernel tricks, and the "Superlearner" model; Step 2.7: Evaluate the prediction accuracy and stability of different models using leave-one-out cross-validation. Obtain the values ​​of four evaluation indicators, R², MAEf, RMSE, and EVS, after each model's prediction. Compare the performance of each model on the above evaluation indicators, and draw radar charts and tables to visually display the performance of different models. For the yield strength performance prediction model, select the "Superlearner" model, which performs best on all indicators, to predict the yield strength performance of the alloy material. For the tensile strength performance prediction model, select the support vector machine model, which performs best on all indicators, to predict the tensile strength performance of the alloy material. Step 2.8: Conduct further experimental verification and evaluation of the model determined in Step 2.7, optimize the model based on the evaluation results, and adjust the parameters to improve the model performance.

4. The method according to claim 3, characterized in that, Step 2.1 specifically includes the following steps: Step 2.1.1: For the yield strength of titanium alloy, a correlation coefficient analysis was performed to determine 38 characteristic parameters with V alloy composition as the main feature set for the yield strength performance prediction model. Step 2.1.1.1: Calculate the Pearson correlation coefficient between each feature in the dataset and the yield strength of the titanium alloy; Step 2.1.1.2: Select features with an absolute value of Pearson correlation coefficient greater than 0.5 and determine them as features highly correlated with yield strength; Step 2.1.1.3: Select 38 characteristic parameters, mainly V alloy composition, from Step 2.1.1.2 to construct the feature set of the yield strength performance prediction model; Step 2.1.2: For the tensile strength of titanium alloy, correlation coefficient analysis was performed to determine 44 characteristic parameters with Fe alloy composition as the main feature set as the feature set of the tensile strength performance prediction model; Step 2.1.2.1: Calculate the Pearson correlation coefficient between each feature in the dataset and the tensile strength of the titanium alloy; Step 2.1.2.2: Select features with an absolute value of Pearson correlation coefficient greater than 0.5 and identify them as features highly correlated with tensile strength; Step 2.1.2.3: Select 44 characteristic parameters, mainly Fe alloy composition, from Step 2.1.2.2 to construct the feature set of the tensile strength performance prediction model; Step 2.1.3: Integrate the feature sets obtained in Step 2.1.1 and Step 2.1.2 to form two independent feature datasets, which are used for subsequent training of yield strength and tensile strength performance prediction models, respectively.

5. The method according to claim 1, characterized in that, Step 4 specifically includes the following steps: Step 4.1: Input the new titanium alloy composition and processing parameters, and use the model trained in Step 2 to perform performance prediction and obtain the prediction results; Step 4.2: Generate new scatter plots and line plots using the multidimensional visualization analysis tool built in Step 3. Combine the intervals to which the performance prediction values ​​belong in the line plots with the scatter plots, and compare the scatter plots with similar performance within the same clusters to find the features that are positively or negatively correlated with performance, thereby obtaining optimization schemes that can improve performance. Step 4.3: Based on the optimization scheme obtained in Step 4.2, try to increase components that are positively correlated with performance or decrease components that are negatively correlated to form new combinations; Step 4.4: Use the new combination obtained in Step 4.3 as input for Step 4.1 to obtain new prediction results. Compare these new prediction results with the original prediction results. Based on the comparison results, analyze and evaluate the effectiveness of the improvement scheme. Iterate and optimize the titanium alloy composition and processing parameters repeatedly to determine the optimal combination of titanium alloy composition and processing technology, thereby achieving the expected performance indicators.

Images