Novelty detection is a classification problem to identify abnormal patterns; therefore, it is an important task for applications such as fraud detection, fault diagnosis and disease detection. However, when there is no label that indicates normal and abnormal data, it will need expensive domain and professional knowledge, so an unsupervised novelty detection approach will be used. On the other hand, nowadays, using novelty detection on high dimensional data is a big challenge and previous research suggests approaches based on principal component analysis (PCA) and an autoencoder in order to reduce dimensionality. In this paper, we propose deep autoencoders with density based clustering (DAE-DBC); this approach calculates compressed data and error threshold from deep autoencoder model, sending the results to a density based cluster. Points that are not involved in any groups are not considered a novelty; the grouping points will be defined as a novelty group depending on the ratio of the points exceeding the error threshold. We have conducted the experiment by substituting components to show that the components of the proposed method together are more effective. As a result of the experiment, the DAE-DBC approach is more efficient; its area under the curve (AUC) is shown to be 13.5 percent higher than state-of-the-art algorithms and other versions of the proposed method that we have demonstrated.

This study proposes an efficient prediction method for coronary heart disease risk based on two deep neural networks trained on well-ordered training datasets. Most real datasets include an irregular subset with higher variance than most data, and predictive models do not learn well from these datasets. While most existing prediction models learned from the whole or randomly sampled training datasets, our suggested method draws up training datasets by separating regular and highly biased subsets to build accurate prediction models. We use a two-step approach to prepare the training dataset: (1) divide the initial training dataset into two groups, commonly distributed and highly biased using Principal Component Analysis, (2) enrich the highly biased group by Variational Autoencoders. Then, two deep neural network classifiers learn from the isolated training groups separately. The well-organized training groups enable a chance to build more accurate prediction models. When predicting the risk of coronary heart disease from the given input, only one appropriate model is selected based on the reconstruction error on the Principal Component Analysis model. Dataset used in this study was collected from the Korean National Health and Nutritional Examination Survey. We have conducted two types of experiments on the dataset. The first one proved how Principal Component Analysis and Variational Autoencoder models of the proposed method improves the performance of a single deep neural network. The second experiment compared the proposed method with existing machine learning algorithms, including Naïve Bayes, Random Forest, K-Nearest Neighbor, Decision Tree, Support Vector Machine, and Adaptive Boosting. The experimental results show that the proposed method outperformed conventional machine learning algorithms by giving the accuracy of 0.892, specificity of 0.840, precision of 0.911, recall of 0.920, f-measure of 0.915, and AUC of 0.882.

Automatic anomaly detection for time-series is critical in a variety of real-world domains such as fraud detection, fault diagnosis, and patient monitoring. Current anomaly detection methods detect the remarkably low proportion of the actual abnormalities correctly. Furthermore, most of the datasets do not provide data labels, and require unsupervised approaches. By focusing on these problems, we propose a novel deep learning-based unsupervised anomaly detection approach (RE-ADTS) for time-series data, which can be applicable to batch and real-time anomaly detections. RE-ADTS consists of two modules including the time-series reconstructor and anomaly detector. The time-series reconstructor module uses the autoregressive (AR) model to find an optimal window width and prepares the subsequences for further analysis according to the width. Then, it uses a deep autoencoder (AE) model to learn the data distribution, which is then used to reconstruct a time-series close to the normal. For anomalies, their reconstruction error (RE) was higher than that of the normal data. As a result of this module, RE and compressed representation of the subsequences were estimated. Later, the anomaly detector module defines the corresponding time-series as normal or an anomaly using a RE based anomaly threshold. For batch anomaly detection, the combination of the density-based clustering technique and anomaly threshold is employed. In the case of real-time anomaly detection, only the anomaly threshold is used without the clustering process. We conducted two types of experiments on a total of 52 publicly available time-series benchmark datasets for the batch and real-time anomaly detections. Experimental results show that the proposed RE-ADTS outperformed the state-of-the-art publicly available anomaly detection methods in most cases.

Coronary heart disease (CHD) is one of the leading causes of death worldwide; if suffering from CHD and being in its end-stage, the most advanced treatments are required, such as heart surgery and heart transplant. Moreover, it is not easy to diagnose CHD at the earlier stage; hospitals diagnose it based on various types of medical tests. Thus, by predicting high-risk people who are to suffer from CHD, it is significant to reduce the risks of developing CHD. In recent years, some research works have been done using data mining to predict the risk of developing diseases based on medical tests. In this study, we have proposed a reconstruction error (RE) based deep neural networks (DNNs); this approach uses a deep autoencoder (AE) model for estimating RE. Initially, a training dataset is divided into two groups by their RE divergence on the deep AE model that learned from the whole training dataset. Next, two DNN classifiers are trained on each group of datasets separately by combining a RE based new feature with other risk factors to predict the risk of developing CHD. For creating the new feature, we use deep AE model that trained on the only high-risk dataset. We have performed an experiment to prove how the components of our proposed method work together more efficiently. As a result of our experiment, the performance measurements include accuracy, precision, recall, F-measure, and AUC score reached 86.3371%, 91.3716%, 82.9024%, 86.9148%, and 86.6568%, respectively. These results show that the proposed AE-DNNs outperformed regular machine learning-based classifiers for CHD risk prediction.

Any bearing faults are a leading cause of motor damage and bring economic losses. Fast and accurate identification of bearing faults is valuable for preventing damaging the whole equipment and continuously running industrial processes without interruption. Vibration signals from a running motor can be utilized to diagnose a bearing health condition. This study proposes a detection method for bearing faults based on two types of neural networks from motor vibration data. The proposed method uses an autoencoder neural network for constructing a new motor vibration feature and a feed-forward neural network for the final detection. The constructed signal feature enhances the prediction performance by focusing more on a fault type that is difficult to detect. We conducted experiments on the CWRU bearing datasets. The experimental study shows that the proposed method improves the performance of the feed-forward neural network and outperforms the other machine learning algorithms.