ABSTRACT
Fraudulent activities especially in auto insurance and credit card transactions impose significant financial losses on businesses and individuals. To overcome this issue, we propose a novel approach for fraud detection, combining convolutional neural networks (CNNs) with support vector machine (SVM), k nearest neighbor (KNN), naive Bayes (NB), and decision tree (DT) algorithms. The core of this methodology lies in utilizing the deep features extracted from the CNNs as inputs to various machine learning models, thus significantly contributing to the enhancement of fraud detection accuracy and efficiency. Our results demonstrate superior performance compared to previous studies, highlighting our model's potential for widespread adoption in combating fraudulent activities.
ABSTRACT
Infrared small target detection is widely applied in military and civilian fields. Due to the small size of infrared targets, textural detail is missing. Common target detection methods extract semantic feature by narrowing down the feature map several times, which may lead to the small targets lost in deep layers and are not effective for infrared small target detection. To solve this problem, we propose a novel network called deep asymmetric extraction and aggregation. The network mainly consists of two processes - the vertical feature extraction and the horizontal feature aggregation, both of which are enhanced by an asymmetric attention mechanism. In the vertical process, the use of asymmetric attention mechanism combined with the reduction of down-sampling makes the small target better retained in the deep layers. Then through the horizontal process, shallow spatial feature and deep semantic feature are aggregated to further highlight the small targets while suppressing background noise. Experiments on the public datasets NUAA-SISRT, NUDT-SISRT and MDvsFA-cGan show that our proposed network outperforms the state-of-the-art methods in terms of detection accuracy and parameter efficiency.
ABSTRACT
BACKGROUND: Infectious diseases such as SARS and H1N1 can significantly impact people's lives and cause severe social and economic damages. Recent outbreaks have stressed the urgency of effective research on the dynamics of infectious disease spread. However, it is difficult to predict when and where outbreaks may emerge and how infectious diseases spread because many factors affect their transmission, and some of them may be unknown. METHODS: One feasible means to promptly detect an outbreak and track the progress of disease spread is to implement surveillance systems in regional or national health and medical centres. The accumulated surveillance data, including temporal, spatial, clinical, and demographic information can provide valuable information that can be exploited to better understand and model the dynamics of infectious disease spread. The aim of this work is to develop and empirically evaluate a stochastic model that allows the investigation of transmission patterns of infectious diseases in heterogeneous populations. RESULTS: We test the proposed model on simulation data and apply it to the surveillance data from the 2009 H1N1 pandemic in Hong Kong. In the simulation experiment, our model achieves high accuracy in parameter estimation (less than 10.0 % mean absolute percentage error). In terms of the forward prediction of case incidence, the mean absolute percentage errors are 17.3 % for the simulation experiment and 20.0 % for the experiment on the real surveillance data. CONCLUSION: We propose a stochastic model to study the dynamics of infectious disease spread in heterogeneous populations from temporal-spatial surveillance data. The proposed model is evaluated using both simulated data and the real data from the 2009 H1N1 epidemic in Hong Kong and achieves acceptable prediction accuracy. We believe that our model can provide valuable insights for public health authorities to predict the effect of disease spread and analyse its underlying factors and to guide new control efforts.
