Artificial Intelligence Approach in Aerospace for Error Mitigation

Citation: Bautista-Hernández, J.; Martín-Prats, M.Á. Artificial Intelligence Approach in Aerospace for Error Mitigation. Aerospace 2024, 11, 300. https://doi.org/10.3390/ aerospace11040300 Academic Editor: Peng Wei Received: 6 March 2024 Revised: 30 March 2024 Accepted: 9 April 2024 Published: 11 April 2024 Copyright: © 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). aerospace Article Artificial Intelligence Approach in Aerospace for Error Mitigation Jorge Bautista-Hernández 1,2,* and MaríaÁngeles Martín-Prats 1 1Department of Electronics Engineering, University of Seville, 41004 Seville, Spain; [email protected] 2Department of Electrical Engineering, Airbus Poland, 02-256 Warsaw, Poland *Correspondence: [email protected] Abstract: Many of the reports created at assembly lines, where all components of an aircraft are installed, frequently indicate that errors threaten safety. The proposed methodology in this study evaluates error prediction and risk mitigation to prevent failures and their consequences. The results linked to a typical electrical harness manufacture of a military aircraft estimated reductions of 93% in time and 90% in error during the creation of engineering manufacturing processes using AI techniques. However, traditional risk assessments methods struggle to identify and mitigate errors effectively. Thus, developing an advanced methodology to ensure systems safety is needed. This paper addresses how innovative AI technology solutions can overcome these challenges, mitigate error risks, and enhance safety in aerospace. Technologies, such as artificial intelligence, predictive algorithms, machine learning, and automation, can play a key role in enhancing safety. The aim of this study is to develop a model that considers the factors that can potentially contribute to error creation, through an artificial intelligence (AI) approach. The specific AI techniques used such as support vector machine, random forest, logistic regression, K-nearest neighbor, and XGBoost (Python 3.8.5) show good performance for use in error mitigation. We have compared the modeled values obtained in this study with the experimental ones. The results confirm that the best metrics are obtained by using support vector machine and logistic regression. The smallest deviation between the measured and modeled values for these AI methods do not exceed 5%. Furthermore, using advancements in machine learning methods can enhance error mitigation in aerospace. The use of AutoML can play a key role in automatically finding an appropriate model which provides the best performance metrics and therefore the most reliable forecast for data prediction and error mitigation. Keywords: predictive algorithm; artificial intelligence; error mitigation; machine learning; risk matrix 1. Introduction In the aerospace safety field, there are different types of data collected to analyze accidents, non-conformances, deviations, and errors. Safety databases are composed of reports with integrated tools to allow the creation of understandable data for analysts in different formats. Typically, systems for quality, security, and safety are integrated into the aerospace organization and monitored to ensure compliance with safety standards like International Civil Aviation Organization (ICAO) Annex 19 [ 1 ]. Traditional methods use indicators to define an event as being acceptable. Most of the investigation processes, which are related to the identification of these factors, aim to evaluate undesired outcomes [ 2 ]. Risk assessment processes are fundamental to identify, evaluate, and mitigate potential risks. The approach of risk mitigation by assessment is considered critical to safe operations [ 3 ]. Aviation has developed aeronautical decision-making systems, which consider factors such as requirements, aircraft performance, and safety parameters to help make the best decisions [ 4 ]. A risk matrix is used to evaluate the likelihood and consequences of an event. After defining the risk, impact mitigation techniques are applied to avoid potential errors [ 5 , 6 ]. However, the limitations of current databases in safety risk assessment prevent Aerospace 2024,11, 300. https://doi.org/10.3390/aerospace11040300 https://www.mdpi.com/journal/aerospace Aerospace 2024,11, 300 2 of 21 the provision of efficient risk mitigation measures. Therefore, ensuring safety at the highest level is compromised. Aerospace safety risk refers to the evaluation of undesired situations that may lead to accidents or incidents. Undesired outcomes and hazards may be present in all aerospace activities. There are associated uncertainties that need to be properly considered in order to avoid higher risks [ 7 ]. Thus, continuous safety monitoring is performed to manage these undesired factors. Events such as equipment damage, serious incidents, injury to people, or multiple deaths may occur. Measures for risk mitigation are directly associated with safety risk [ 8 ]. The methods used to control safety risks rely on likelihood analysis of safety outcomes, considering factors such as equipment failures or human errors. Strategies to mitigate these risks are also established for this purpose [9]. Aerospace activities result from a chain of complex events involving multiple variables and different scenarios with rapid changes that require fast decisions. The multiple dependent variables involved in the end-to-end process make aerospace activities susceptible to hazards. In particular, aerospace is affected by latent conditions such as design and manufacturer decisions, procedure writers, management, and maintenance activities [ 10 ]. These conditions are generated mainly by people who make incorrect decisions within inadequate organizational processes. Consequently, real-time hazard predictions are a cutting-edge approach to challenge. Continuous research and development in aviation technology aim to enhance safety methods, making them more effective in anticipating and preventing failures [11]. In recent years, data have grown exponentially; therefore, requirements are more difficult to meet and errors may appear more often. Electrical systems are more complex and higher data accuracy is required for efficient performance [ 12 ]. New technology has improved flight safety. However, there is evidence that human errors are the main trigger for errors [ 13 , 14 ]. Systems have increased in complexity and more interconnections are necessary for good performance [ 7 ]. The complexity and interconnection of aerospace systems can contribute to increasing the probability of error creation and failure propagation. Current approaches do not have the necessary capability to address potential risk and to establish strategies for error mitigation. Thus, the integration of advanced error mitigation using AI methods to evaluate the risk of error creation under real-time conditions is necessary. AI approaches can provide innovative solutions to overcome the limitations of existing techniques. Concerning safety issues, different approaches, such as manufacturing processes based on the latest 4.0 technology, are solutions to produce final safety products [ 15 ]. Industry 4.0 is based on a methodology that represents a smart chained network where machines and products interact with each other automatically. This technology can not only improve flight safety [ 16 ], but also has a significant impact on overcoming challenges and enhancing sustainable business performance. The safety process involves identification problems that appear in the early design stage through an approach aiming to identify future threats by prediction. The use of AI techniques provides more evidence for more accurate decision making [ 17 , 18 ]. Predictive algorithms developed in this context with application to aerospace can play a key role in preventing failures [ 19 ] and consequently in preventing accidents and saving human lives. Artificial intelligence technology can prevent undesired situations from occurring [ 20 ]. The integration of AI into design models has the potential to enhance safety and reliability. AI models can not only handle complex and nonlinear data but also have the capacity to anticipate errors and make the best decisions in case uncertainties appear. Furthermore, advanced analytics and AI algorithm development can analyze large amounts of data to identify patterns which can potentially contribute to create failures [ 21 ]. All of these factors are relevant for use as references for safety in aviation [22]. This innovative approach leverages a risk matrix function, which enables a structured evaluation and visualization of threats. This approach is driven by real-time data, which require a dynamic understanding of risk by considering multiple factors and probabilistic Aerospace 2024,11, 300 3 of 21 approaches for making real-time decisions [ 23 ]. The methodology established in this study will accelerate the development of tasks, monitor real-time data, and assess the risk of error creation and its severity. Using the algorithm’s insights, the best decision can be made in the dynamic manufacturing area and in simultaneously forecasting the risk of error creation for new datasets [24–26]. The main aim of this study is to prevent failures in existing manufacturing processes by applying techniques based on AI machine learning. Moreover, this methodology enhances efficiency and productivity by integrating this technology in the engineering processes and enabling real-time communication between interconnected devices [ 17 , 27 , 28 ]. The use of automation of existing manual tasks will not only improve the work delivered from the engineering to the manufacturing area, but will also reduce the human decision factor, thereby decreasing the risk of accidents. Vulnerabilities may compromise the data, increasing the risks of cyberattacks. Consequently, security strategies are implemented to protect the integrity of the algorithm [29]. The objectives of this research are to design an AI model to mitigate errors and evaluate their risks in the aerospace manufacturing assembly line. In particular, support vector machine (SVM), random forest (RF), logistic regression (LR), K-nearest neighbor (KNN), and XGBoost (eXtreme Gradient Boosting) have been used and evaluated using the following metrics: accuracy, precision, recall, and F 1 -score. Additionally, this approach can not only reduce errors but potentially enhance efficiency and bring improvements in the aerospace manufacturing assembly line. Overall, this study aims to enhance safety and keep aerospace at the highest level [30]. 2. Materials and Methods Aerospace frequently uses a high volume of electrical harness drawings. Threedimensional models are used to create a full-scale drawing enriched with manufacturing information, which is called a Formboard drawing. Harness assessment is performed using a computer-aided-design (CAD) model that represents an assembly with electrical components. The risk matrix is calculated according to the criteria shown in Table 1, which is based on the following input parameters: number of wires (H), terminations (Z), and electrical components (N). The most vulnerable stage, in terms of error creation, involves harness transformation from the current model to a flattened model. Electrical parameters are added to the model in order to flatten the 3D model accordingly. The generated mesh is extracted using an automatic script. The target is to flatten the entire mesh generated in the real length to produce a 2D model. The risk matrix function, which represents a modern and effective approach to safety management, evaluates and prioritizes risks based on their likelihood and impact. The inputs/outputs of the AI model developed in this study are presented in Figure 1. The inputs to the system ( X1, X2, X3, X4) are the factors related to the type of harness and the outputs include automation, predictions, and risk assessments to inform decision making in the engineering processes. Aerospace 2024, 11, x FOR PEER REVIEW 4 of 23 Figure 1. Predictive innovative algorithm structure defined by input parameters, risk matrix function, and outputs. The dataset related to C295 aircraft consists of 157 electrical harnesses, which are obtained from the bill of material of each instance. It includes information about wiring length, number of electrical components, and electrical protection length. In total, in this study, the dataset includes 18,523 m of total wiring length, 21,200 electrical components, and 250 m of protective sheath length present in the 157 instances. The functionality of the risk matrix provides features for identification of the risk in a more efficient manner. It is defined by categories such as low, medium, and high risk, which are established by assigning scores to each element of the dataset. Thus, data prioritization needs to be considered for immediate attention and mitigation. Overall, management of the risk provides information in real time in order to make the best decisions [31]. The new digital environment requires a holistic approach that integrates more automation and system interconnection. The latest implemented technology in industry 4.0 contains hi-tech strategies that enable sustainability by establishing connections between objects and interconnecting intelligent systems. The development of predictive innovative algorithms plays a key role in predicting error creation and achieving a better performance of the results [28]. Integrating AI methodologies into the manufacturing processes can enhance both efficiency and safety. The AI techniques used in this study include SVM, RF, LR, KNN, and XGBoost, which contribute to the analysis and results. SVM can be used for both data classification and regression. RF maintains accuracy from the average effect return of each individual tree. LR is used to estimate the probability of the binary outcomes. KNN presents high consistency in the results. XGBoost achieves high accuracy in predictions. Predictive algorithm functionalities, such as model evaluation, can be beneficial in assessing the performance of the model using appropriate metrics [32]. The appropriate selection of AI techniques to train the model and the continuous monitoring of the outcomes can be fundamental to make accurate predictions [33]. The predictive algorithm structure developed in this study is presented in Figure 1. 2.1. Automation Script Complete automation is achieved by using a script that automates repetitive tasks necessary to turn the 3D harness model into a 2D manufacturing drawing. The developed automatic script includes the operations such as rotation of the electrical components and labeling of the electrical components. Other, more complicated operations such as rotation of the wiring, quick rolls, bends, and/or translation of the terminations are also automatically performed. This situation enables the automatic spreading of the 3D harness model into a 2D Formboard without errors [34]. Figure 1. Predictive innovative algorithm structure defined by input parameters, risk matrix function, and outputs. Aerospace 2024,11, 300 4 of 21 The dataset related to C295 aircraft consists of 157 electrical harnesses, which are obtained from the bill of material of each instance. It includes information about wiring length, number of electrical components, and electrical protection length. In total, in this study, the dataset includes 18,523 m of total wiring length, 21,200 electrical components, and 250 m of protective sheath length present in the 157 instances. The functionality of the risk matrix provides features for identification of the risk in a more efficient manner. It is defined by categories such as low, medium, and high risk, which are established by assigning scores to each element of the dataset. Thus, data prioritization needs to be considered for immediate attention and mitigation. Overall, management of the risk provides information in real time in order to make the best decisions [31]. The new digital environment requires a holistic approach that integrates more automation and system interconnection. The latest implemented technology in industry 4.0 contains hi-tech strategies that enable sustainability by establishing connections between objects and interconnecting intelligent systems. The development of predictive innovative algorithms plays a key role in predicting error creation and achieving a better performance of the results [ 28 ]. Integrating AI methodologies into the manufacturing processes can enhance both efficiency and safety. The AI techniques used in this study include SVM, RF, LR, KNN, and XGBoost, which contribute to the analysis and results. SVM can be used for both data classification and regression. RF maintains accuracy from the average effect return of each individual tree. LR is used to estimate the probability of the binary outcomes. KNN presents high consistency in the results. XGBoost achieves high accuracy in predictions. Predictive algorithm functionalities, such as model evaluation, can be beneficial in assessing the performance of the model using appropriate metrics [ 32 ]. The appropriate selection of AI techniques to train the model and the continuous monitoring of the outcomes can be fundamental to make accurate predictions [ 33 ]. The predictive algorithm structure developed in this study is presented in Figure 1. 2.1. Automation Script Complete automation is achieved by using a script that automates repetitive tasks necessary to turn the 3D harness model into a 2D manufacturing drawing. The developed automatic script includes the operations such as rotation of the electrical components and labeling of the electrical components. Other, more complicated operations such as rotation of the wiring, quick rolls, bends, and/or translation of the terminations are also automatically performed. This situation enables the automatic spreading of the 3D harness model into a 2D Formboard without errors [34]. Additionally, industrialization involves the creation of whole engineering documentation required for harness manufacturing. In traditional methods, this process is created from the perspective of the compliance with the requirements. In this new approach, a risk matrix function, which evaluates the likelihood of errors during the manufacturing process, is used. This assessment can be performed with five main dimensions: 3D model, the ‘Formboard’ (FB) defined as X1 representing the complexity of the 3D geometry, and X2, X3, and X4. Table 1. Criteria associated with each harness category associated with the risk matrix. Description Variables Criteria Cluster Rule Category Formboard (FB) X1Number of zones (Z) Very large Z > 100 5 Large 60 < Z < 100 4 Medium 20 < Z < 60 3 Small 2 < Z < 20 2 Very small Z = 2 1 Technical Instructions (IT) X2Number of wires (H) Very large H > 600 5 Large 300 < H < 600 4 Medium 100 < H < 300 3 Small 10 < H < 100 2 Very small H < 10 1 Aerospace 2024,11, 300 5 of 21 Table 1. Cont. Description Variables Criteria Cluster Rule Category Manufacturing Bill of Material (BOM) Electrical Test (ET) X3 X4 Number of components (N) Number of wires (H) Very large N > 500 5 Large 300 < N < 500 4 Medium 100 < N < 300 3 Small 20 < N < 100 2 Very small N < 20 1 Very large H > 600 5 Large 300 < H < 600 4 Medium 100 < H < 300 3 Small 10 < H < 100 2 Very small H < 10 1 2.2. Risk Matrix The complexity matrix for each of the harnesses was generated using input scores. The parameters which were used were determined by quantification of the number of zones (Z), number of wires (H), and number of electrical components (N) present in each harness dataset. The risk matrix function in an explicit form is represented in the following equation: Φ(Z, H, N)=Σ4 i=1Xi=X1(Z)+X2(H)+X3(N)+X4(H)(1) The initiation of the kernel size, denoted as X i , for the case study of the electrical harness in a military aircraft C295 follows the steps depicted in Figure 2. It shows the algorithm structure and highlights the different values of the risk matrix for each range associated with scores assigned to the dataset. A red color is assigned for a very high and high risk; yellow for a medium risk; and green for a low and very low risk of error creation. Aerospace 2024, 11, x FOR PEER REVIEW 6 of 23 Figure 2. Algorithm diagnostic for the selection of electrical harness complexity in manufacturing plant. The final score, as a sum of these values, determines the likelihood of error creation during the manufacturing engineering processes. Scores from 4 to 20 correspond from the simplest to the most complex geometry of the 3D electrical harness shown in Table 2. Table 2. Rule established to define risk matrix assessment in each harness. Category Rule Risk Matrix 1 0 ≤ Φ ≤ 4 Very Low 2 4 < Φ ≤ 8 Low 3 9 ≤ Φ ≤ 12 Medium 4 13 ≤ Φ ≤ 17 High 5 Φ > 17 Very High The risk matrix outcomes split the dataset into three main groups. The dataset is grouped into high, medium, and low risk of creating an error during the manufacturing process. Histograms in Figure 3 show how the scores are assigned to the parameters X1, X2, X3,and X4 in the case study of C295 military aircraft. From these histograms, only a very small percent of 3.8% of the dataset had scores of 4 and 5, which could potentially generate errors. The ability to focus attention on specific data or components rather than treating the entire system uniformly improves effectiveness in the processes. Clustering data will help to identify the features and similarities between them for failure anticipation. The clusters suggest that certain combinations of parameters can occur together, indicating relationships between these variables [35]. Figure 2. Algorithm diagnostic for the selection of electrical harness complexity in manufacturing plant. The final score, as a sum of these values, determines the likelihood of error creation during the manufacturing engineering processes. Scores from 4 to 20 correspond from the simplest to the most complex geometry of the 3D electrical harness shown in Table 2. Table 2. Rule established to define risk matrix assessment in each harness. Category Rule Risk Matrix 1 0 ≤Φ≤4 Very Low 2 4 < Φ≤8 Low 3 9 ≤Φ≤12 Medium 4 13 ≤Φ≤17 High 5Φ> 17 Very High Aerospace 2024,11, 300 6 of 21 The risk matrix outcomes split the dataset into three main groups. The dataset is grouped into high, medium, and low risk of creating an error during the manufacturing process. Histograms in Figure 3show how the scores are assigned to the parameters X1 , X2 , X3 , and X4 in the case study of C295 military aircraft. From these histograms, only a very small percent of 3.8% of the dataset had scores of 4 and 5, which could potentially generate errors. The ability to focus attention on specific data or components rather than treating the entire system uniformly improves effectiveness in the processes. Clustering data will help to identify the features and similarities between them for failure anticipation. The clusters suggest that certain combinations of parameters can occur together, indicating relationships between these variables [35]. Aerospace 2024, 11, x FOR PEER REVIEW 7 of 23 Figure 3. Histograms showing the quantity of harness part numbers on ‘y’ axis and the scores assigned on ‘x’ axis for each input parameter. Histogram (a) represents parameter X1; (b) parameter X2; (c) parameter X3; (d) parameter X4. Figure 4 shows the risk matrix function evaluated for each harness within the dataset and the probability associated with the error creation during the manufacturing engineering processes in a color code. Figure 4. Risk matrix associated with electrical harness in military aircraft with level of risk of error creation in each harness part number of the dataset. Red color depicts a very high/high risk; yellow depicts a medium risk, and green depicts a low/very low risk of error creation. 2.3. Clustering Hierarchy The main types of algorithms differ in the type of input/output data and scope of the problem that should be solved. Unsupervised algorithm techniques based on clustering are useful for application within the harness dataset. These techniques can be beneficial for identifying similarities and differences among harnesses, and therefore, for a more detailed analysis [27,35]. Figure 3. Histograms showing the quantity of harness part numbers on ‘y’ axis and the scores assigned on ‘x’ axis for each input parameter. Histogram (a) represents parameter X1 ; (b) parameter X2; (c) parameter X3; (d) parameter X4. Figure 4shows the risk matrix function evaluated for each harness within the dataset and the probability associated with the error creation during the manufacturing engineering processes in a color code. Aerospace 2024, 11, x FOR PEER REVIEW 7 of 23 Figure 3. Histograms showing the quantity of harness part numbers on ‘y’ axis and the scores assigned on ‘x’ axis for each input parameter. Histogram (a) represents parameter X1; (b) parameter X2; (c) parameter X3; (d) parameter X4. Figure 4 shows the risk matrix function evaluated for each harness within the dataset and the probability associated with the error creation during the manufacturing engineering processes in a color code. Figure 4. Risk matrix associated with electrical harness in military aircraft with level of risk of error creation in each harness part number of the dataset. Red color depicts a very high/high risk; yellow depicts a medium risk, and green depicts a low/very low risk of error creation. 2.3. Clustering Hierarchy The main types of algorithms differ in the type of input/output data and scope of the problem that should be solved. Unsupervised algorithm techniques based on clustering are useful for application within the harness dataset. These techniques can be beneficial for identifying similarities and differences among harnesses, and therefore, for a more detailed analysis [27,35]. Figure 4. Risk matrix associated with electrical harness in military aircraft with level of risk of error creation in each harness part number of the dataset. Red color depicts a very high/high risk; yellow depicts a medium risk, and green depicts a low/very low risk of error creation. Aerospace 2024,11, 300 7 of 21 2.3. Clustering Hierarchy The main types of algorithms differ in the type of input/output data and scope of the problem that should be solved. Unsupervised algorithm techniques based on clustering are useful for application within the harness dataset. These techniques can be beneficial for identifying similarities and differences among harnesses, and therefore, for a more detailed analysis [27,35]. The algorithm used in this study applies clustering to analyze and to classify datasets. Figure 5presents the clustering of data points defined by the risk matrix. Thus, ‘Formboard’ (FB) is represented by cluster 0 (in blue), ‘Technical Instructions’ (IT) by cluster 1 (in orange), ‘Manufacturing Bill of Material’ (MBOM) by cluster 2 (in green), and ‘Electrical Test’ (ET) by cluster 3 (in red). Aerospace 2024, 11, x FOR PEER REVIEW 8 of 23 The algorithm used in this study applies clustering to analyze and to classify datasets. Figure 5 presents the clustering of data points defined by the risk matrix. Thus, ‘Formboard’ (FB) is represented by cluster 0 (in blue), ‘Technical Instructions’ (IT) by cluster 1 (in orange), 'Manufacturing Bill of Material' (MBOM) by cluster 2 (in green), and 'Electrical Test' (ET) by cluster 3 (in red). Figure 5. Cluster’s structure computed using Python grouping the harness dataset in four clusters. The agglomerative clustering hierarchy uses input data defined as 𝐴𝐴1∪… ∪ 𝐴𝐴𝑗𝑗 for the set for each observation in each cluster, with the following constraints: 𝐴𝐴1∪… ∪ 𝐴𝐴𝑗𝑗 represents the number of clusters. 𝐴𝐴 𝑗𝑗 ∩ 𝐴𝐴 𝑗𝑗′ = ∅ no overlapping j ≠ j’. The initialization of the clusters and merging process is as follows: • Calculation of the proximity matrix, which represents the distances between each pair of data points based on the minimum distance: dmin =�∑(qi−pi)2 n i=1 , where pi and qi are coordinates of the data points. • Identify two clusters with minimum distance. • Merge the identified clusters into a new cluster combining the elements 𝐴𝐴1∪… ∪ 𝐴𝐴𝑗𝑗. • After merging two clusters, create a new cluster (𝐴𝐴𝑛𝑛𝑛𝑛𝑛𝑛) represented as 𝐴𝐴𝑛𝑛𝑛𝑛𝑛𝑛 = 𝐴𝐴1+ 𝐴𝐴𝑗𝑗, which contains elements of both clusters 𝐴𝐴1 and 𝐴𝐴𝑗𝑗. • Recalculate the distances between the new cluster and the remainder. • Update the proximity matrix with the new distances. • Continue this process iteratively. • Repeat until the stop criterion is met with all elements merged in a cluster. Figure 6 shows the merging process depicted as a dendrogram, which provides information about the critical clusters to be considered. Therefore, the dataset is grouped in families that give information about potential anomalies to be considered in order to anticipate error creation. Figure 5. Cluster’s structure computed using Python grouping the harness dataset in four clusters. The agglomerative clustering hierarchy uses input data defined as A1∪. . . ∪Aj for the set for each observation in each cluster, with the following constraints: A1∪. . . ∪Ajrepresents the number of clusters.Aj∩Aj′=no overlapping j =j′. The initialization of the clusters and merging process is as follows: • Calculation of the proximity matrix, which represents the distances between each pair of data points based on the minimum distance: dmin =qΣn i=1(qi−pi)2 , where p i and qiare coordinates of the data points. •Identify two clusters with minimum distance. • Merge the identified clusters into a new cluster combining the elements A1∪. . . ∪Aj . • After merging two clusters, create a new cluster ( Anew ) represented as Anew =A1+Aj , which contains elements of both clusters A1and Aj. •Recalculate the distances between the new cluster and the remainder. •Update the proximity matrix with the new distances. •Continue this process iteratively. •Repeat until the stop criterion is met with all elements merged in a cluster. Figure 6shows the merging process depicted as a dendrogram, which provides information about the critical clusters to be considered. Therefore, the dataset is grouped in families that give information about potential anomalies to be considered in order to anticipate error creation. Aerospace 2024,11, 300 8 of 21 Aerospace 2024, 11, x FOR PEER REVIEW 9 of 23 Figure 6. Harness dendrogram in the hierarchical clustering illustrates a group of similar data within each node. Hierarchical clustering establishes relationships between similar datasets grouped into clusters and labeled by letters from A to P. 2.4. Logistic Regression The best fit for the outcome categorical variable is defined by a logistic regression. Figure 7 represents the relationship between the probability of error creation and the risk matrix. The risk matrix function presents different values for each harness within the dataset. The probability function associated with the logistic regression is used for new data predictions. The logistic regression passes the input through the logistic/sigmoid function and then treats the outcome result as a probability [29]. Figure 7. Logistic regression dataset fits to a sigmoid function which predicts the probability of error creation. Figure 6. Harness dendrogram in the hierarchical clustering illustrates a group of similar data within each node. Hierarchical clustering establishes relationships between similar datasets grouped into clusters and labeled by letters from A to P. 2.4. Logistic Regression The best fit for the outcome categorical variable is defined by a logistic regression. Figure 7represents the relationship between the probability of error creation and the risk matrix. The risk matrix function presents different values for each harness within the dataset. The probability function associated with the logistic regression is used for new data predictions. The logistic regression passes the input through the logistic/sigmoid function and then treats the outcome result as a probability [29]. Aerospace 2024, 11, x FOR PEER REVIEW 9 of 23 Figure 6. Harness dendrogram in the hierarchical clustering illustrates a group of similar data within each node. Hierarchical clustering establishes relationships between similar datasets grouped into clusters and labeled by letters from A to P. 2.4. Logistic Regression The best fit for the outcome categorical variable is defined by a logistic regression. Figure 7 represents the relationship between the probability of error creation and the risk matrix. The risk matrix function presents different values for each harness within the dataset. The probability function associated with the logistic regression is used for new data predictions. The logistic regression passes the input through the logistic/sigmoid function and then treats the outcome result as a probability [29]. Figure 7. Logistic regression dataset fits to a sigmoid function which predicts the probability of error creation. Figure 7. Logistic regression dataset fits to a sigmoid function which predicts the probability of error creation. 2.5. Confusion Matrix The confusion matrix considers the scores for each harness based on the number of zones (Z), number of wires (H), and number of components (N). The most important metrics in (2–5) are obtained from the confusion matrix as follows: • Precision is defined as the ratio of correct positive predictions to the total number of positive predictions. This is the ratio between the real positive predictions and the total number of positives as the sum of true and false positives. Aerospace 2024,11, 300 9 of 21 • The recall or sensitivity ratio is defined as the true positive prediction of the total number of positives. It is the ratio between real positives and real cases indicated as true positives or false negatives. • Accuracy is defined as the ratio of the true prediction to the total value. It is the ratio of the total number of true positives and negatives in relation to the total number of real positives and real negatives within the dataset. • The F 1 score is the harmonic mean of the precision and recall used to measure test accuracy. It varies from 1.0, showing perfect values of precision and recall, to 0, where either precision or recall is zero. Precision =TP TP +FP (2) Recall or sensitivity =TP P=TP TP +FN (3) Accuracy =TP +TN P+N=TP +TN TP +TN +FP +FN (4) F1=2 1 Precision +1 Recall =2 TP 2 TP +FP +FN (5) In addition, the machine learning classification report provides some additional values as outcomes. Category 0 represents the negative class that means ‘no error creation’, and category 1 represents the positive class, meaning ‘error creation’. The macro-average is the average of the precision, recall, and F 1 -score. The weighted average is the weighted average of the precision, recall, and F 1 scores, as shown in Table 3. The performance metrics are depicted in Figure 8. Table 3. Classification reports of machine learning metrics such as precision, recall, and F 1 -score, with macro- and weighted average. Category Precision Recall F1-Score 0 1.00 0.95 0.98 1 0.91 1.00 0.95 Accuracy - - 0.97 Macro-average 0.95 0.98 0.96 Weighted average 0.97 0.97 0.97 Aerospace 2024, 11, x FOR PEER REVIEW 11 of 23 Figure 8. Performance metrics representation for precision, recall, and F1 score, with macro- and weighted average. The algorithm splits the dataset into two subsets: training set and test set. The training set was used to fit the model, and the test set was used to evaluate the performance of the model through the confusion matrix as shown in Figure 9. Figure 9. Confusion matrix outcome classification in true positives, true negatives, false positives, and false negatives. From the confusion matrix, the predictions show the good performance of the algorithm. True positives (TP): Prediction is a value that matches reality. The element at (1,1) represents the six data samples as true positives classified as ‘error creation’. True negatives (TN): The prediction value also matches reality. The element at (0,0) represents the 26 data samples as true negatives classified as ‘no error creation’. The diagonal elements Figure 8. Performance metrics representation for precision, recall, and F 1 score, with macro- and weighted average. Aerospace 2024,11, 300 16 of 21 The LR has shown very good performance in comparison to other proposed, existing techniques in terms of F 1 -score. It may not have the highest accuracy but its overall performance suggests that it can be the best choice. SVM may be a good choice since it has achieved the highest accuracy and precision; however, the recall is lower than LR. 3.3. Experimental Verification In order to verify the accuracy of the proposed methodology, we have compared the performance of AI methods using model data of this study and experimental data presented by ElDali and Kumar 2021 [ 37 ]. The experimental datasets used the Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) and Prognostics and Health Monitoring (PHM08) generated by NASA Ames Research Center (Mountain View, CA, USA). The datasets contain features related to a total of 435 aircraft engines from sensor measurements. The sensor values were selected according to the metrics presenting the highest suitability to define failure patterns. The measurements were used for predictions of the remaining useful lifetime of the aircraft engines and their fault estimation. To verify the performance of our model, we have applied AI methods including SVM, RF, LR, KNN, and XGBoost to the measurement datasets of the sensors. Each AI algorithm is used to determine its effectiveness using technical data (TD) from input parameters of this study and measured data from the sensors (SM). The performance metrics are represented in Table 6. Table 6. Performance metrics comparison obtained using different AI methods between technical dataset (TD) and sensor measurement (SM). Metrics AI Methods Accuracy Precision Recall F1-Score TD SM TD SM TD SM TD SM SVM 0.98 0.95 1.00 1.00 0.85 0.88 0.92 0.94 RF 0.97 1.00 1.00 1.00 0.82 1.00 0.90 1.00 LR 0.97 1.00 1.00 1.00 0.95 1.00 0.98 1.00 KNN 0.95 0.95 0.76 1.00 1.00 0.88 0.86 0.94 XGBoost 0.96 1.00 1.00 1.00 0.80 1.00 0.88 1.00 The difference between SM and TD in accuracy is up to 4% for SVM, RF, LR, and XGBoost. The accuracy obtained using KNN is the same on the measured and model values. The measured and modeled precision is the same for all the methods, except for KNN, which shows a difference of 24%. The recall shows a difference up to 18% between SM and TD. The F 1 -score presents a difference up to 12%. The analysis reveals that SVM and LR have consistently performed, showing the smallest deviation between measured and model values for all metrics. Overall, the experimental values are aligned with the modeled ones, which proves the good performance of the developed model in this study. 3.4. Advancements in Machine Learning The process involved in the simulation of the AI algorithms includes the necessary steps to ensure optimal model performance and accurate predictions. Data preprocessing is necessary to ensure that the input data are appropriate for model training. Additionally, the most suitable AI techniques need to be selected, which depend on the characteristics of the dataset. Also, it is crucial to optimize selected hyperparameters in order to establish an appropriate set up for the AI model proposed. Moreover, visualizations of the model performance metrics are necessary to assess the model accuracy and effectiveness. Thus, Automated Machine Learning (AutoML) can be a valuable tool to perform the entire process automatically. To assess the effectiveness and suitability of the selected AI methods, we have compared the traditional machine learning approach with AutoML. We have used the same input data for both traditional and AutoML models. The AutoML frameworks include tools such as H 2 O-AutoML, DataRobot, Cloud AutoML, the Tree-based Pipeline Optimization Tool (TPOT), Auto-Keras, Auto-Weka, ML BOX, AutoSklearn, and Auto- Pytorch [ 38 – 41 ]. In this study, we have used the TPOT, which uses a Genetic Algorithm (GA) Aerospace 2024,11, 300 17 of 21 to automatically find which AI methods present the best performance for the dataset [ 42 ]. We have compared performance metrics of the AI methods including SVM, RF, LR, KNN, and XGBoost for both AutoML and traditional approaches. The results are presented in Figure 17. Aerospace 2024, 11, x FOR PEER REVIEW 19 of 23 LR 0.97 1.00 1.00 1.00 0.95 1.00 0.98 1.00 KNN 0.95 0.95 0.76 1.00 1.00 0.88 0.86 0.94 XGBoost 0.96 1.00 1.00 1.00 0.80 1.00 0.88 1.00 The difference between SM and TD in accuracy is up to 4% for SVM, RF, LR, and XGBoost. The accuracy obtained using KNN is the same on the measured and model values. The measured and modeled precision is the same for all the methods, except for KNN, which shows a difference of 24%. The recall shows a difference up to 18% between SM and TD. The F1-score presents a difference up to 12%. The analysis reveals that SVM and LR have consistently performed, showing the smallest deviation between measured and model values for all metrics. Overall, the experimental values are aligned with the modeled ones, which proves the good performance of the developed model in this study. 3.4. Advancements in Machine Learning The process involved in the simulation of the AI algorithms includes the necessary steps to ensure optimal model performance and accurate predictions. Data preprocessing is necessary to ensure that the input data are appropriate for model training. Additionally, the most suitable AI techniques need to be selected, which depend on the characteristics of the dataset. Also, it is crucial to optimize selected hyperparameters in order to establish an appropriate set up for the AI model proposed. Moreover, visualizations of the model performance metrics are necessary to assess the model accuracy and effectiveness. Thus, Automated Machine Learning (AutoML) can be a valuable tool to perform the entire process automatically. To assess the effectiveness and suitability of the selected AI methods, we have compared the traditional machine learning approach with AutoML. We have used the same input data for both traditional and AutoML models. The AutoML frameworks include tools such as H2O-AutoML, DataRobot, Cloud AutoML, the Tree-based Pipeline Optimization Tool (TPOT), Auto-Keras, Auto-Weka, ML BOX, AutoSklearn, and Auto-Pytorch [38–41]. In this study, we have used the TPOT, which uses a Genetic Algorithm (GA) to automatically find which AI methods present the best performance for the dataset [42]. We have compared performance metrics of the AI methods including SVM, RF, LR, KNN, and XGBoost for both AutoML and traditional approaches. The results are presented in Figure 17. Figure 17. Performance metrics of the AI methods including KNN, LR, SVM, RF, and XGBoost for both AutoML (darker colored bars) and traditional (lighter colored bars) approaches. Figure 17. Performance metrics of the AI methods including KNN, LR, SVM, RF, and XGBoost for both AutoML (darker colored bars) and traditional (lighter colored bars) approaches. Figure 17 shows the performance metrics obtained using AutoML and traditional ML for each of the AI methods proposed. In each graph, the darker-colored bars represent the metrics obtained using AutoML and the lighter-colored bars represent the metrics using manual ML. The highest deviation appears in recall in KNN, showing a difference of 15% between AutoML and the manual approach. LR presents the highest difference of 12% in recall. The methods SVM, RF, and XGBoost show a difference up to 3% in recall and F 1 . The results show that leveraging AutoML can find the best classifier while reducing the time invested in manual ML. We use the k-fold cross technique to validate the performance of the AI methods proposed. Such a technique splits the dataset into k-folds. The TPOT evaluates the model on each fold in order to estimate the performance more accurately. The TPOT also uses the value of k = 5, splitting the dataset into five equal parts. Each fold is used as a validation set while the remaining four are used for training. It evaluates the model five times in order to assess the performance of the different AI methods. This value balances the performance accuracy and the computational effort. Table 7presents the metrics of the best pipelines using a 10-fold cross-validation. These metrics indicate that the best pipelines proposed by the TPOT provide an average accuracy, recall, and F 1 -score of 95%, while the average for precision is 96%. Table 7. Performance metrics of the best pipelines using k-fold cross-validation. K-Folds Accuracy Precision Recall F1-Score 1st-fold 97 97 97 97 2nd-fold 96 96 96 95 3rd-fold 96 96 96 95 4th-fold 95 95 95 94 5th-fold 96 97 97 97 6th-fold 96 96 96 96 7th-fold 95 96 95 95 8th-fold 95 96 95 95 9th-fold 94 95 94 94 10th-fold 93 94 93 93 Average 95 96 95 95 Aerospace 2024,11, 300 18 of 21 3.5. Discussion Logistic regression is used for classification of the dataset and the evaluation of model performance by a confusion matrix. The logistic regression can evaluate the significance of the input parameters and the potential impact in the output model. Thus, a powerful predictive algorithm can forecast error creation in order to avoid errors. Simultaneously, errors in engineering documentation are mitigated. Therefore, validation of the results can provide evidence to enhance process efficiency and reliability at electrical manufacturing in aerospace and overall safety in aviation. The dataset from the statistical analysis shows a Spearman coefficient of 0.98, which demonstrates that the variables are highly positive correlated. The box-plot representation shows that the dataset outcomes present a deviation above the tendency line compared to an ideal normal distribution of 3.85 in a large harness cluster, 1.65 in a medium one, and 0.65 in a small one. AI results based on failure anticipation methods show that the highest probability within the dataset to create errors in engineering processes are presented by the high-risk matrix. A 90% reduction in error creation has been achieved by the AI model developed in this study. The highest correlation is presented between X3 and X4 parameters with the ‘risk matrix’ at values of 0.98 and 0.97, respectively. Additionally, a sensitivity analysis shows good prediction results, proving the robustness and reliability of the model. The parameter variation shows consistency in the model outputs and are proved to validate model reliability and robustness of the model. Furthermore, the flexibility presented by the risk matrix can accommodate other types of risk. The proposed approach can be generalized to other disciplines such as healthcare, energy, and automotives. Healthcare can gain the benefit of using the risk matrix, which offers real-time information to the surgeons about the level of risk they will face during the surgical procedure. This approach can provide the detection and prevention of failures. Additionally, it can also provide prioritization on which patient to act on in terms of mortality, in order to save their lives. In the energy sector, the risk matrix can be used to analyze the hazards associated with equipment failures to enhance safety strategies or mitigate losses by implementing more efficient processes. Furthermore, the risk matrix can play a key role in product safety in the automotive industry. The assessment of the risks can be useful in reducing the accident rate and decreasing production time. With the increasing digitalization of automotives, the risk matrix can also be useful in assessing cybersecurity threats. The use of AutoML can play a key role in automatically finding an appropriate model which provides reliable forecasts for data prediction. The accuracy and effectiveness of the predictive models used can be improved without the need of extensive human intervention. For future directions, in spacecraft systems, considering existing constraints related to limited data availability, uncertainty in values, and the need for real-time decisions can be very beneficial to use. However, continuous research and innovation will be necessary to apply AutoML to aerospace applications and reach the highest accuracy. 4. Conclusions In this study, AI techniques are applied to design predictive algorithms capable of predicting and mitigating potential failures in the aerospace manufacturing processes. The decision-making aspect implies the assessment of the risk of error creation by a risk matrix function. This approach has the potential to enhance efficiency, reduce errors, and contribute to overall improvements in the aerospace manufacturing domain. The integration of AI techniques into the manufacturing design process for electrical harnesses enhances systems reliability, improves the performance of electrical systems, and improves error mitigation, making it well suited for aerospace. The AI results can be used to identify nonlinear data behavior and complex patterns within the dataset, enhancing confidence in the models’ outputs. After comparing different AI techniques, such as SVM, RF, LR, KNN, and XGBoost, the best performance is obtained from logistic regression (LR). Therefore, LR is used for output predictions. Integrating AI techniques, logistic regression Aerospace 2024,11, 300 19 of 21 provides a rigorous analysis of the model behavior and enhances the reliability of the proposed approach. The experimental verification analysis demonstrated that LR, RF, and XGBoost show high performance. These AI methods are feasible for predicting errors and mitigating risks of aerospace systems. The experimental values are aligned with the modeled ones, which provides evidence of the good performance of the model approach in this study. This analysis indicates that SVM and LR have shown the smallest deviation up to 5% between measured and model values for all metrics. The use of AutoML has also provided evidence to obtain the best metrics of the selected AI methods. The results have shown that the performance metrics obtained using AutoML and traditional ML are comparable. However, leveraging AutoML can enhance the choice of the best classifier, while reducing the time used in manual ML. Author Contributions: Conceptualization, methodology, software Python version 3.8.5, analysis, writing, J.B.-H.; methodology review, visualization of Section 2, supervision, M.Á.M.-P. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Data Availability Statement: Data presented in this study are contained within the article. Acknowledgments: The authors acknowledge the contributions of the electronic engineering department from Seville University and Airbus Poland for supporting this work with their research facilities and resources. Conflicts of Interest: Author Jorge Bautista-Hernández was employed by the company Airbus Poland. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Abbreviations ICAO International Civil Aviation Organization AI Artificial intelligence CAD Computer-aided design AS Automatic script FB Formboard IT Technical Instructions MBOM Manufacturing Bill of Material ET Electrical Test TP True positives TN True negatives FP False positives FN False negatives EHF Electrical Harness Flattening SVM Support vector machine RF Random forest LR Logistic regression KNN K-nearest neighbor XGBoost eXtreme Gradient Boosting C-MAPSS Commercial Modular Aero-Propulsion System Simulation PHM08 Prognostics and Health Monitoring AutoML Automated Machine Learning TPOT Tree-based Pipeline Optimization Tool GA Genetic Algorithm Aerospace 2024,11, 300 20 of 21 References 1. ICAO. Safety Management Manual, Doc 9859; International Civil Aviation Organization (ICAO): Montreal, QC, Canada, 2013. 2. Barr, L.C.; Newman, R.; Ancel, E.; Belcastro, C.M.; Foster, J.V.; Evans, J.; Klyde, D.H. Preliminary risk assessment for small unmanned aircraft systems. In Proceedings of the 17th AIAA Aviation Technology, Integration, and Operations Conference, Denver, CO, USA, 5–9 June 2017; p. 3272. 3. Leveson, N. A new accident model for engineering safer systems. Saf. Sci. 2004,42, 237–270. [CrossRef] 4. O’Hare, D. Aeronautical Decision Making: Metaphors, Models, and Methods; Routledge: Abingdon, UK, 2003. 5. Stamatelatos, M.; Dezfuli, H.; Apostolakis, G.; Everline, C.; Guarro, S.; Mathias, D.; Mosleh, A.; Paulos, T.; Riha, D.; Smith, C. Probabilistic Risk Assessment Procedures Guide for NASA Managers and Practitioners; NTRS: Chicago, IL, USA, 2011. 6. Clothier, R.A.; Walker, R.A. The safety risk management of unmanned aircraft systems. In Handbook of Unmanned Aerial Vehicles; Springer Science: Dordrecht, The Netherlands, 2015; pp. 2229–2275. 7. Rutkowski, M. Safety as an element of creating competitive advantage among airlines given the example of the Airbus A350 XWB and the Boeing 787 Dreamliner aircraft. In Zeszyty Naukowe. Transport/Politechnika ´ Sl ˛aska; The University of Economics: Katowice, Poland, 2020. 8. Schober, P.; Boer, C.; Schwarte, L.A. Correlation Coefficients: Appropriate Use and Interpretation. Anesth. Analg. 2018,126, 1763–1768. [CrossRef] [PubMed] 9. Casado, R.S.G.R.; Alencar, M.H.; de Almeida, A.T. Combining a multidimensional risk evaluation with an implicit enumeration algorithm to tackle the portfolio selection problem of a natural gas pipeline. Reliab. Eng. Syst. Saf. 2022,221, 108332. [CrossRef] 10. Reason, J.; Hollnagel, E.; Paries, J. Revisiting the Swiss cheese model of accidents. J. Clin. Eng. 2006,27, 110–115. 11. Velásquez, R.M.A. Root cause analysis for inverters in solar photo-voltaic plants. Eng. Fail. Anal. 2020,118, 104856. [CrossRef] 12. Atak, A.; Kingma, S. Safety culture in an aircraft maintenance organisation: A view from the inside. Saf. Sci. 2011,49, 268–278. [CrossRef] 13. Kelly, D.; Efthymiou, M. An analysis of human factors in fifty controlled flight into terrain aviation accidents from 2007 to 2017. J. Saf. Res. 2019,69, 155–165. [CrossRef] [PubMed] 14. Ladkin, P.B.; Schepper, W. EMI, TWA 800 and Swissair 111. In The Resks Digist; University of Bielefeld: Bielefeld, Germany, 2000; Volume 21. 15. Yang, J.; Kim, J. Accident diagnosis algorithm with untrained accident identification during power-increasing operation. Reliab. Eng. Syst. Saf. 2020,202, 107032. [CrossRef] 16. Haseeb, M.; Hussain, H.I.; ´ Slusarczyk, B.; Jermsittiparsert, K. Industry 4.0: A solution towards technology challenges of sustainable business performance. Soc. Sci. 2019,8, 154. [CrossRef] 17. Jordan, M.I.; Mitchell, T.M. Machine learning: Trends, perspectives, and prospects. Science 2015,349, 255–260. [CrossRef] 18. Kaempf, G.L.; Klein, G. Aeronautical decision making: The next generation. In Aviation Psychology in Practice; Routledge: London, UK, 2017; pp. 223–254. 19. Bozhko, S.; Hill, C.I.; Yang, T. More-Electric Aircraft: Systems and Modeling. In Wiley Encyclopedia of Electrical and Electronics Engineering; Wiley: Hoboken, NY, USA, 1999; pp. 1–31. 20. Makins, N.; Kirwan, B. Keeping the Aviation Industry Safe; EU: Brussels, Belgium, 2016. 21. Sha, Y.; He, Z.; Du, J.; Zhu, Z.; Lu, X. Intelligent detection technology of flip chip based on H-SVM algorithm. Eng. Fail. Anal. 2022,134, 106032. [CrossRef] 22. Dasari, S.K.; Cheddad, A.; Andersson, P. Predictive modelling to support sensitivity analysis for robust design in aerospace engineering. Struct. Multidiscip. Optim. 2020,61, 2177–2192. [CrossRef] 23. Herwan, J.; Misaka, T.; Furukawa, Y.; Ogura, I.; Komoto, H. A proposal for improving production efficiency of existing machining line through a hybrid monitoring and optimisation process. Int. J. Prod. Res. 2023,61, 5392–5410. [CrossRef] 24. Spearman, C. The proof and measurement of association between two things. Am. J. Psychol. 1987,100, 441–471. [CrossRef] [PubMed] 25. Todinov, M. New algorithms for optimal reduction of technical risks. Eng. Optim. 2013,45, 719–743. [CrossRef] 26. Nuzzo, R.L. The Box Plots Alternative for Visualizing Quantitative Data. PM&R 2016,8, 268–272. [CrossRef] 27. Bonaccorso, G. Machine Learning Algorithms; Packt Publishing Ltd.: Birmingham, UK, 2017. 28. Gui, G.; Zhou, Z.; Wang, J.; Liu, F.; Sun, J. Machine learning aided air traffic flow analysis based on aviation big data. IEEE Trans. Veh. Technol. 2020,69, 4817–4826. [CrossRef] 29. Bautista-Hernández, J.; Martín-Prats, M.Á. The Impact of Data Injection on Predictive Algorithm Developed within Electrical Manufacturing Engineering in the Context of Aerospace Cybersecurity. Aerospace 2023,10, 984. [CrossRef] 30. Boukerche, A.; Wang, J. Machine learning-based traffic prediction models for intelligent transportation systems. Comput. Netw. 2020,181, 107530. [CrossRef] 31. Liu, C.; Yang, S.; Cui, Y.; Yang, Y. An improved risk assessment method based on a comprehensive weighting algorithm in railway signaling safety analysis. Saf. Sci. 2020,128, 104768. [CrossRef] 32. Utku, A.; Kaya, S.K. Multi-layer perceptron based transfer passenger flow prediction in Istanbul transportation system. Decis. Mak. Appl. Manag. Eng. 2022,5, 208–224. [CrossRef] 33. Feng, H.; Zhang, Y. Prediction of Hourly Airport Operational Throughput with a Multi-Branch Convolutional Neural Network. Aerospace 2024,11, 78. [CrossRef] Aerospace 2024,11, 300 21 of 21 34. Khoshoo, B.; Blank, J.; Pham, T.Q.; Deb, K.; Foster, S.N. Optimal design of electric machine with efficient handling of constraints and surrogate assistance. Eng. Optim. 2024,56, 274–292. [CrossRef] 35. Gan, G.; Ma, C.; Wu, J. Data Clustering: Theory, Algorithms, and Applications; SIAM: Bangkok, Thailand, 2020. 36. Bautista-Hernández, J.; Martín-Prats, M.Á. Monte Carlo Simulation Applicable for Predictive Algorithm Analysis in Aerospace. In Doctoral Conference on Computing, Electrical and Industrial Systems; Springer Nature: Cham, Switzerland, 2023; pp. 243–256. 37. ElDali, M.; Kumar, K.D. Fault diagnosis and prognosis of aerospace systems using growing recurrent neural networks and lstm. In Proceedings of the 2021 IEEE Aerospace Conference (50100), Big Sky, MT, USA, 6–13 March 2021; pp. 1–20. 38. Chauhan, K.; Jani, S.; Thakkar, D.; Dave, R.; Bhatia, J.; Tanwar, S.; Obaidat, M.S. Automated machine learning: The new wave of machine learning. In Proceedings of the 2020 2nd International Conference on Innovative Mechanisms for Industry Applications (ICIMIA), Bangalore, India, 5–7 March 2020; pp. 205–212. 39. Alsharef, A.; Aggarwal, K.; Sonia; Kumar, M.; Mishra, A. Review of ML and AutoML solutions to forecast time-series data. Arch. Comput. Methods Eng. 2022,29, 5297–5311. [CrossRef] [PubMed] 40. Madni, H.A.; Umer, M.; Ishaq, A.; Abuzinadah, N.; Saidani, O.; Alsubai, S.; Hamdi, M.; Ashraf, I. Water-quality prediction based on H2O AutoML and explainable AI techniques. Water 2023,15, 475. [CrossRef] 41. Tufail, S.; Riggs, H.; Tariq, M.; Sarwat, A.I. Advancements and challenges in machine learning: A comprehensive review of models, libraries, applications, and algorithms. Electronics 2023,12, 1789. [CrossRef] 42. Olson, R.S.; Moore, J.H. TPOT: A tree-based pipeline optimization tool for automating machine learning. In Proceedings of the Workshop on Automatic Machine Learning, New York, NY, USA, 24 June 2016; pp. 66–74. Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.