China
Africa
Explore chapters and articles related to this topic
Machine Learning
Published in Ian Foster, Rayid Ghani, Ron S. Jarmin, Frauke Kreuter, Julia Lane, Big Data and Social Science, 2020
Cross-validation is a more sophisticated holdout training and testing procedure that removes some of the shortcomings of the holdout set approach. Cross-validation begins by splitting a labeled data set into k partitions (called folds). Typically, k is set to 5 or 10. Then, cross-validation proceeds by iterating k times. In each iteration, one of the k folds is held out as the test set, while the other k − 1 folds are combined and used to train the model. A nice property of cross-validation is that every example is used in one test set for testing the model. Each iteration of cross-validation gives us a performance estimate that then can be aggregated (typically averaged) to generate the overall estimate.
Learning Techniques
Published in Peter Wlodarczak, Machine Learning and its Applications, 2019
A frequently used type of cross-validation is k-fold cross-validation. k-fold cross-validation is a resampling procedure that can be used when there is limited training data. Whereas normally a data set is split into training, validation and test data, k-fold cross-validation uses all observations for training and testing. The basic ideas is to split the training data into k equal sized folds randomly and use k − 1 folds for training and the remaining hold out fold for testing. This procedure is repeated k times until every fold has been used exactly once for testing. In this way, all the instances are used for training and every observation is used for validation only once. If, for instance, k = 10 it is also called 10-fold cross-validation.
Pattern Recognition Techniques for Image Super-Resolution
Published in Peyman Milanfar, Super-Resolution Imaging, 2017
Cross-validation, an important component of all learning algorithms, assesses the fit of a model prior to its usage. Empirical methods for image and video processing, like the ones discussed in this chapter, must not only be tuned numerically but visually as well. Common forms of cross-validation are leave-one-out and k-fold cross-validation. Cross-validation guards against overfitting and overgeneralizing. For example, Fig. 11.11 aids in the design of the adaptive k-NN regression techniques described by Section 11.2.3, where the parameter η must be cross-validated. Too small an η and the image becomes visually noisy whereas too large an η and the image is blurred and washed out.
A prediction model for flexural strength of corroded prestressed concrete beam using artificial neural network
Published in Cogent Engineering, 2023
Yamuna Bhagwat, Gopinatha Nayak, Radhakrishna Bhat, Muralidhar Kamath
It is observed from Figure 5a and Figure 5b that the lowest RMSE and highest R2 values are recorded when K-fold is three and training repetition is one. The K-fold cross-validation is a technique of running the training phase (K-fold)-1 times by randomly choosing (1-(1/K-fold)) records for training and remaining (1/K-fold) records for validating the training results. The K-fold cross-validation process will have a direct impact on the prediction results. The RMSE value has been decreasing till K-fold = 3 and starts increasing when K-fold in all the proposed models. Conversely, the R2 value has been increasing till K-fold = 3 and starts decreasing when K-fold in all the proposed models. It is equally essential to run the model with a higher number of repetitions to fit the model. The training process of the proposed models involves 1 to 10 training repetitions.
Deep learning for industrial image: challenges, methods for enriching the sample space and restricting the hypothesis space, and possible issue
Published in International Journal of Computer Integrated Manufacturing, 2022
Tianyuan Liu, Jinsong Bao, Junliang Wang, Jiacheng Wang
In the process of DL, the dataset is generally divided into three parts: the training set, the validation set and the testing set. The validation set is mainly used to observe the training process to optimize the hyper-parameters. There are three types of cross validation (Arlot and Celisse 2010) in deep learning: k-fold cross validation, leave-one-out cross validation, and stratified k-fold cross validation. k-fold cross-validation method is the most commonly used in industry. As shown in Figure 6, k-fold cross validation generally divides the original dataset into k sub-samples. A separate sub-sample is reserved for model validation, and the other sub-samples are used for training. Each sub-sample will be used for validation, and a single estimate can be obtained ultimately by averaging the k results. The advantage of this method is that the limited sample space can be used as much as possible and data interference is eliminated to the greatest extent.
Benchmark examples for data-driven site characterisation
Published in Georisk: Assessment and Management of Risk for Engineered Systems and Geohazards, 2022
Kok-Kwang Phoon, Takayuki Shuku, Jianye Ching, Ikumasa Yoshida
The purpose of cross-validation is to test the performance of a data-driven method using a validation dataset that is distinct (or independent) from the training dataset. The generalisation error can be quantified in this way. All data-driven methods are susceptible to overfitting. Hence, cross-validation is mandatory. A reliable estimate of the generalisation error requires a sufficiently large validation dataset. This is not a problem for synthetic data, as any number of training and validation datasets can be “measured”. It is a problem for sparse real-world site data. The strategies for dividing between training and validation datasets to balance training and mitigation of overfitting are not covered in this paper. Ultimately, the validation dataset and the associated performance metrics should be selected to demonstrate how a method is “fit for practice”.