Browsing by Subject "Artificial neural network"

Now showing 1 - 3 of 3

Open Access
Classification of leg motions by processing gyroscope signals
(IEEE, 2009) Tunçel, Orkun; Altun, Kerem; Barshan, Billur
In this study, eight different leg motions are classified using two single-axis gyroscopes mounted on the right leg of a subject with the help of several pattern recognition techniques. The methods of least squares, Bayesian decision, k-nearest neighbor, dynamic time warping, artificial neural networks and support vector machines are used for classification and their performances are compared. This study comprises the preliminary work for our future studies on motion recognition with a much wider scope.
Open Access
Machine learning and artificial neural networks-based approach to model and optimize ethyl methanesulfonate and sodium azide induced in vitro regeneration and morphogenic traits of water hyssops (Bacopa monnieri L.)
(2022-09-10) Mirza, K.; Aasim, M.; Katırcı, R.; Karataş, M.; Ali, Seyid Amjad
Application of chemical mutagens is used for artificially induced in vitro mutation to develop new cultivars with elite characteristics. However, the optimization of selecting proper mutagen, its concentration, and exposure time is of utmost importance, especially for plants containing noteworthy secondary metabolites. In this study, the effect of sodium azide (NaN3) and ethyl methanesulfonate (EMS) in different concentrations (0.025, 0.05, 0.1, and 0.2 mg l−1), and treatment time (30, 60, and 120 min) was investigated on Bacopa monnieri; an important medicinal plant. The maximum shoot counts (57.0) were achieved from the combination of 0.10 mg l−1 EMS × 60 min. Whereas, maximum shoot length (4.07 cm), node numbers (4.97) and leaf numbers (12,23) were achieved from the combination of 0.20 mg l−1 EMS × 120 min, respectively. Combination of 0.025 mg l−1 NaN3 × 120 mg/l yielded maximum shoot counts (52.30), shoot length (3.23 cm), node numbers (6.07) and leaf numbers (12.13). The trained model to predict the outputs were designed and calibrated with machine learning (ML) algorithms. Support Vector Classifier (SVC), Gaussian Process (GP), Extreme Gradient Boosting (XGBoost), Random Forest (RF) models, and Multilayer Perceptron (MLP) neural network algorithms were used to discover the best models and their hyperparameters. The RF model gave exceptional results in the prediction of the outputs. F1 scores of the RF were acquired in the range of 0.98–1.00 for different outputs. The other models’ F1 scores varied in the range of 0.65 and 0.85. The present work opens the new era of applying ML and artificial neural network (ANN) models in plant tissue culture with the possibility of application for other economic crops.
Open Access
Multifont Ottoman character recognition
(IEEE, 2000) Öztürk, Ali; Güneş, S.; Özbay, Y.
Ottoman characters from three different fonts are used character recognition problem, broadly speaking, is transferring a page that contain symbols to the computer and matching these symbols with previously known or recognized symbols after extraction the features of these symbols via appropriate preprocessing methods. Because of silent features of the characters, implementing an Ottoman character recognition system is a difficult work. Different researchers have done lots of works for years to develop systems that would recognize Latin characters. Although almost one million people use Ottoman characters, great deal of whom has different native languages, the number of studies on this field is insufficient. In this study 28 different machine-printed to train the Artificial Neural Network and a %95 classification accuracy for the characters in these fonts and a %70 classification accuracy for a different font has been found.