Dataset of vehicle images for Indonesia toll road tariff classification.

Vehicle classifications with different methods have been applied for many purposes. The data provided in this article is useful for classifying vehicle purposes following the Indonesia toll road tariffs. Indonesia toll road tariff regulations divide vehicles into five groups as follows, group-1, group-2, group-3, group-4, and group-5, respectively. Group-1 is a class of non-truck vehicles, while group-2 to group-5 are classes of truck vehicles. The non-truck class consists of the sedan, pick-up, minibus, bus, MPV, and SUV. Truck classes are grouped based on the number of truck's axles. Group-2 is a class of trucks with two axles, a group-3 truck with three axles, a group-4 truck with four axles, and a group-5 truck with five axles or more. The dataset is categorized into five classes accordingly, which are group-1, group-2, group-3, group-4, and group-5 images. The data made available in this article observes images of vehicles obtained using a smartphone camera. The vehicle images dataset incorporated with deep learning, transfer learning, fine-tuning, and the Residual Neural Network (ResNet) model can yield exceptional results in the classification of vehicles by the number of axles.

Fast Convolutional Method for Automatic Sleep Stage Classification.

OBJECTIVES: Polysomnography is essential to diagnose sleep disorders. It is used to identify a patient's sleep pattern during sleep. This pattern is obtained by a doctor or health practitioner by using a scoring process, which is time consuming. To overcome this problem, we developed a system that can automatically classify sleep stages. METHODS: This paper proposes a new method for sleep stage classification, called the fast convolutional method. The proposed method was evaluated against two sleep datasets. The first dataset was obtained from physionet.org, a physiologic signals data centers. Twenty-five patients who had a sleep disorder participated in this data collection. The second dataset was collected in Mitra Keluarga Kemayoran Hospital, Indonesia. Data was recorded from ten healthy respondents. RESULTS: The proposed method reached 73.50% and 56.32% of the F-measures for the PhysioNet and Mitra Keluarga Kemayoran Hospital data, respectively. Both values were the highest among all the machine learning methods considered in this study. The proposed method also had an efficient running time. The fast convolutional models of the PhysioNet and Mitra Keluarga Kemayoran Hospital data needed 42.60 and 0.06 seconds, respectively. CONCLUSIONS: The fast convolutional method worked well on the tested datasets. It achieved a high F-measure result and an efficient running time. Thus, it can be considered a promising tool for sleep stage classification.

A New Data Representation Based on Training Data Characteristics to Extract Drug Name Entity in Medical Text.

One essential task in information extraction from the medical corpus is drug name recognition. Compared with text sources come from other domains, the medical text mining poses more challenges, for example, more unstructured text, the fast growing of new terms addition, a wide range of name variation for the same drug, the lack of labeled dataset sources and external knowledge, and the multiple token representations for a single drug name. Although many approaches have been proposed to overwhelm the task, some problems remained with poor F-score performance (less than 0.75). This paper presents a new treatment in data representation techniques to overcome some of those challenges. We propose three data representation techniques based on the characteristics of word distribution and word similarities as a result of word embedding training. The first technique is evaluated with the standard NN model, that is, MLP. The second technique involves two deep network classifiers, that is, DBN and SAE. The third technique represents the sentence as a sequence that is evaluated with a recurrent NN model, that is, LSTM. In extracting the drug name entities, the third technique gives the best F-score performance compared to the state of the art, with its average F-score being 0.8645.

Adaptive Online Sequential ELM for Concept Drift Tackling.

A machine learning method needs to adapt to over time changes in the environment. Such changes are known as concept drift. In this paper, we propose concept drift tackling method as an enhancement of Online Sequential Extreme Learning Machine (OS-ELM) and Constructive Enhancement OS-ELM (CEOS-ELM) by adding adaptive capability for classification and regression problem. The scheme is named as adaptive OS-ELM (AOS-ELM). It is a single classifier scheme that works well to handle real drift, virtual drift, and hybrid drift. The AOS-ELM also works well for sudden drift and recurrent context change type. The scheme is a simple unified method implemented in simple lines of code. We evaluated AOS-ELM on regression and classification problem by using concept drift public data set (SEA and STAGGER) and other public data sets such as MNIST, USPS, and IDS. Experiments show that our method gives higher kappa value compared to the multiclassifier ELM ensemble. Even though AOS-ELM in practice does not need hidden nodes increase, we address some issues related to the increasing of the hidden nodes such as error condition and rank values. We propose taking the rank of the pseudoinverse matrix as an indicator parameter to detect "underfitting" condition.

Metaheuristic Algorithms for Convolution Neural Network.

A typical modern optimization technique is usually either heuristic or metaheuristic. This technique has managed to solve some optimization problems in the research area of science, engineering, and industry. However, implementation strategy of metaheuristic for accuracy improvement on convolution neural networks (CNN), a famous deep learning method, is still rarely investigated. Deep learning relates to a type of machine learning technique, where its aim is to move closer to the goal of artificial intelligence of creating a machine that could successfully perform any intellectual tasks that can be carried out by a human. In this paper, we propose the implementation strategy of three popular metaheuristic approaches, that is, simulated annealing, differential evolution, and harmony search, to optimize CNN. The performances of these metaheuristic methods in optimizing CNN on classifying MNIST and CIFAR dataset were evaluated and compared. Furthermore, the proposed methods are also compared with the original CNN. Although the proposed methods show an increase in the computation time, their accuracy has also been improved (up to 7.14 percent).