China
Africa
Explore chapters and articles related to this topic
Multivariate Approach
Published in Atsushi Kawaguchi, Multivariate Analysis for Neuroimaging Data, 2021
The discrimination boundary of the logistic discrimination introduced earlier is a straight line. There are also quadratic discriminants in which the discriminant boundary is a quadratic curve. If you draw a more complicated discriminant boundary, you may be able to discriminate more accurately. Support vector machine (SVM) is one of the non-linear discriminants and has been studied and applied as a powerful discriminating method. As features, only observation values that are effective for discrimination (support vectors) are used. We introduce the concept of maximizing margins. The margin is the “distance between the discrimination boundary and the data.” Also, consider projection to higher dimensions. It can be imagined that the discrimination ability is higher when viewing from the sky (three-dimensional) than when viewing on flat ground (two-dimensional). At that time, the discriminant boundary can be estimated by using the dual problem and the kernel trick as optimization problems. To give flexibility to the discriminant boundary, a soft margin (which may be a little inside the opponent) is introduced and is controlled by a parameter called a slack variable.
A Levenshtein distance-based method for word segmentation in corpus augmentation of geoscience texts
Published in Annals of GIS, 2023
Jinqu Zhang, Lang Qian, Shu Wang, Yunqiang Zhu, Zhenji Gao, Hailong Yu, Weirong Li
The third method relies on the neural network, which is essentially a sequence labelling task, but makes full use of the powerful feature representation capabilities of the neural network (Vinotheni and LakshmanaPandian 2021; Ma et al. 2021). Owing to the excellent performance of the cyclic neural network and long short-term memory networks (LSTM) in sequential tagging tasks, various neural networks have been widely applied in Chinese word segmentation tasks (Zhao et al. 2018). Many studies have explored the use of neural networks to automatically learn better representations (Wu et al. 2018; Al-Ayyoub et al. 2018; Qiu et al. 2020). For example, the max-margin tensor neural network was proposed to model the relationship between labels and characters (Huang, Sun, and Wang 2017). Meanwhile, Chen et al. (2015) exploited a gated recursive neural network (GRNN) to model the combination of characters for word segmentation and presented four different architectures of LSTM networks to test the effectiveness. Wang and Xu (2017) proposed a convolutional neural network combined with word embedding for Chinese word segmentation. Although different word segmentation methods are constantly proposed, due to the limitation of labelled corpora, the existing word segmentation methods will significantly decrease for the geoscience domain (Zhang et al. 2016).
The moisture migration behavior of wheat starch/gluten blended powders and extrudates
Published in Drying Technology, 2021
Gong Yanfei, Yingquan Zhang, Bo Zhang, Guo Boli, Wei Yimin
Water activity had a significant effect on the sample density (P < 0.05). At a constant wheat starch/gluten ratio, the density of the extrudates decreased when aw was increased from 0.11 to 0.35 and fluctuated within a small margin when aw varied between 0.35 and 0.76. When aw was from 0.11 to 0.35, the range of the changes was the largest. The density of 100% starch blended extrudate was reduced by 57 kg m−3. The density was reduced by 5 kg m−3 when the starch content was 0%. The change rate of density per unit mass of starch was higher than that of gluten indicating that when aw was between 0.11 and 0.35, and the shrinkage of the starch was higher than that of gluten.
Multiobjective non linear model predictive control of transesterification and lipid oil production
Published in Biofuels, 2022
Figures 13a and 13b show the biomass (X) profiles for the single objective optimal control and the MNLMPC. The final amount of biomass as a result of just minimizing X reduces the amount to zero, while the MNLMPC does reduce the amount by more than 50% thereby cutting the cost of purification by 50%. Figures 15a and 15b show the effect of single objective optimal control and MNLMPC on the glucose concentration While the single objective optimal control eliminates almost all of the glucose, the MNLMPC does reduce it albeit by a small margin.