| 1. |
- Joloudari, Javad Hassannataj, et al.
(författare)
-
Effective Class-Imbalance Learning Based on SMOTE and Convolutional Neural Networks
- 2023
-
Ingår i: Applied Sciences. - : MDPI. - 2076-3417. ; 13:6
-
Tidskriftsartikel (refereegranskat)abstract
- Imbalanced Data (ID) is a problem that deters Machine Learning (ML) models from achieving satisfactory results. ID is the occurrence of a situation where the quantity of the samples belonging to one class outnumbers that of the other by a wide margin, making such models’ learning process biased towards the majority class. In recent years, to address this issue, several solutions have been put forward, which opt for either synthetically generating new data for the minority class or reducing the number of majority classes to balance the data. Hence, in this paper, we investigate the effectiveness of methods based on Deep Neural Networks (DNNs) and Convolutional Neural Networks (CNNs) mixed with a variety of well-known imbalanced data solutions meaning oversampling and undersampling. Then, we propose a CNN-based model in combination with SMOTE to effectively handle imbalanced data. To evaluate our methods, we have used KEEL, breast cancer, and Z-Alizadeh Sani datasets. In order to achieve reliable results, we conducted our experiments 100 times with randomly shuffled data distributions. The classification results demonstrate that the mixed Synthetic Minority Oversampling Technique (SMOTE)-Normalization-CNN outperforms different methodologies achieving 99.08% accuracy on the 24 imbalanced datasets. Therefore, the proposed mixed model can be applied to imbalanced binary classification problems on other real datasets.
|
|
| 2. |
- Katnagallu, Shyam, et al.
(författare)
-
Impact of local electrostatic field rearrangement on field ionization
- 2018
-
Ingår i: Journal of Physics D. - : IOP PUBLISHING LTD. - 0022-3727 .- 1361-6463. ; 51:10
-
Tidskriftsartikel (refereegranskat)abstract
- Field ion microscopy allows for direct imaging of surfaces with true atomic resolution. The high charge density distribution on the surface generates an intense electric field that can induce ionization of gas atoms. We investigate the dynamic nature of the charge and the consequent electrostatic field redistribution following the departure of atoms initially constituting the surface in the form of an ion, a process known as field evaporation. We report on a new algorithm for image processing and tracking of individual atoms on the specimen surface enabling quantitative assessment of shifts in the imaged atomic positions. By combining experimental investigations with molecular dynamics simulations, which include the full electric charge, we confirm that change is directly associated with the rearrangement of the electrostatic field that modifies the imaging gas ionization zone. We derive important considerations for future developments of data reconstruction in 3D field ion microscopy, in particular for precise quantification of lattice strains and characterization of crystalline defects at the atomic scale.
|
|
| 3. |
|
|
