Test error bounds for classifiers: A survey of old and new results

Anguita, Davide; Ghelardoni, Luca; Ghio, Alessandro; Ridella, Sandro

doi:10.1109/foci.2011.5949469

Cited by 12 publications

(7 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, the ANNs contained three hidden layers and up to eight nodes in each hidden layer. Considering the data set size and computational cost, 10-fold cross-validation was used to validate the ANN training method. , Relative root-mean-square error (RMSE) and coefficient of determination ( R 2 ) were calculated as validation metrics. Null models were constructed as an additional validation step by averaging all the inputs, kinetic parameters, and internal resistance across all samples .…”

Section: Methodsmentioning

confidence: 99%

Hybrid Modeling of Engineered Biological Systems through Coupling Data-Driven Calibration of Kinetic Parameters with Mechanistic Prediction of System Performance

2023

View full text Add to dashboard Cite

Mechanistic models can provide predictive insight into the design and optimization of engineered biological systems, but the kinetic parameters in these models need to be frequently calibrated and uniquely identified. This limitation can be addressed by hybrid modeling that integrates mechanistic models with datadriven approaches. Herein, we developed a hybrid modeling strategy using bioelectrochemical systems as a platform system. The data-driven component consisted of artificial neural networks (ANNs) trained with mechanistically derived kinetic parameters as outputs to compute error signals. The hybrid model was built using 148 samples from the literature. After 10-fold crossvalidation, the model was tested with another 28 samples. Internal resistance was accurately predicted with a relative root-meansquare error (RMSE) of 3.9%. Microbial kinetic parameters were predicted using the data-driven component and fed into the mechanistic component to simulate the system performance. The R 2 values between predicted and observed organic removal and current for systems fed with a simple substrate were 0.90 and 0.94, respectively, significantly higher than those obtained from the stand-alone data-driven model (0.51 and 0) and mechanistic model (0.07 and 0.15). This strategy can potentially be applied to engineered biological systems for in silico system design and optimization.

show abstract

Section: Methodsmentioning

confidence: 99%

Hybrid Modeling of Engineered Biological Systems through Coupling Data-Driven Calibration of Kinetic Parameters with Mechanistic Prediction of System Performance

2023

View full text Add to dashboard Cite

show abstract

“…This ensures that the validation set is distinct from the training set and undergoes training k times. This approach minimizes training bias towards specific data, providing a more accurate evaluation of the model's general performance [61,62].…”

Section: -Fold Cross-validationmentioning

confidence: 99%

Automated Debonding Characterization in Reinforced Structures Based on Finite Element Analysis and Convolutional Neural Networks

Kim,

Han

2024

IEEE Access

View full text Add to dashboard Cite

A study utilizing convolutional neural networks (CNN) has been conducted to detect and classify invisible debonding-type defects in reinforced structures. Training data for these defects is collected from the finite element models of honeycomb sandwich panels and skin-stringer systems, commonly employed reinforcement structures in aerospace applications. The excitation frequency is determined based on the amplitude of the reflected wave from the defect, and the optimal sensor array is selected. The constructed two-dimensional training image, created by vertically stacking the measured responses in the time domain, exhibits high classification performance even with a shallow neural network. The neural network undergoes optimization through adjustment to the kernel parameters and initial learning rate. To assess the general performance of the training model, k-fold cross-validation is employed. The CNN-based non-destructive evaluation algorithm demonstrates high classification performance for debonding defects in honeycomb sandwich panels and skin-stringer structures. Moreover, the suggested algorithm is robust against noise, emphasizing its effectiveness in real-world applications.

show abstract

“…The best-fit model for the particular dataset can be observed by tuning this set of additional variables. Following the model selection phase, the error estimation phase ensures the reliability of the results by assessing the performance of the chosen model [57].…”

Section: Validation Of the Climatic Data With Groundnut Yield Using T...mentioning

confidence: 99%

An Artificial Neural Network for Predicting Groundnut Yield Using Climatic Data

Sajindra,

Abekoon,

Wimalasiri

et al. 2023

AgriEngineering

View full text Add to dashboard Cite

Groundnut, being a widely consumed oily seed with significant health benefits and appealing sensory profiles, is extensively cultivated in tropical regions worldwide. However, the yield is substantially impacted by the changing climate. Therefore, predicting stressed groundnut yield based on climatic factors is desirable. This research focuses on predicting groundnut yield based on several combinations of climatic factors using artificial neural networks and three training algorithms. The Levenberg–Marquardt, Bayesian Regularization, and Scaled Conjugate Gradient algorithms were evaluated for their performance using climatic factors such as minimum temperature, maximum temperature, and rainfall in different regions of Sri Lanka, considering the seasonal variations in groundnut yield. A three-layer neural network was employed, comprising a hidden layer. The hidden layer consisted of 10 neurons, and the log sigmoid functions were used as the activation function. The performance of these configurations was evaluated based on the mean squared error and Pearson correlation. Notable improvements were observed when using the Levenberg–Marquardt algorithm as the training algorithm and applying the natural logarithm transformation to the yield values. These improvements were evident through the higher Pearson correlation values for training (0.84), validation (1.00) and testing (1.00), and a lower mean squared error (2.2859 × 10−21) value. Due to the limited data, K-Fold cross-validation was utilized for optimization, with a K value of 5 utilized for the process. The application of the natural logarithm transformation to the yield values resulted in a lower mean squared error (0.3724) value. The results revealed that the Levenberg–Marquardt training algorithm performs better in capturing the relationships between the climatic factors and groundnut yield. This research provides valuable insights into the utilization of climatic factors for predicting groundnut yield, highlighting the effectiveness of the training algorithms and emphasizing the importance of carefully selecting and expanding the climatic factors in the modeling equation.

show abstract

Test error bounds for classifiers: A survey of old and new results

Cited by 12 publications

References 23 publications

Hybrid Modeling of Engineered Biological Systems through Coupling Data-Driven Calibration of Kinetic Parameters with Mechanistic Prediction of System Performance

Hybrid Modeling of Engineered Biological Systems through Coupling Data-Driven Calibration of Kinetic Parameters with Mechanistic Prediction of System Performance

Automated Debonding Characterization in Reinforced Structures Based on Finite Element Analysis and Convolutional Neural Networks

An Artificial Neural Network for Predicting Groundnut Yield Using Climatic Data

Contact Info

Product

Resources

About