Machine learning and artificial intelligence have achieved a human-level performance in many application domains, including image classification, speech recognition and machine translation. However, in the financial domain expert-based credit risk models have still been dominating. Establishing meaningful benchmark and comparisons on machine-learning approaches and human expert-based models is a prerequisite in further introducing novel methods. Therefore, our main goal in this study is to establish a new benchmark using real consumer data and to provide machine-learning approaches that can serve as a baseline on this benchmark. We performed an extensive comparison between the machine-learning approaches and a human expert-based model—FICO credit scoring system—by using a Survey of Consumer Finances (SCF) data. As the SCF data is non-synthetic and consists of a large number of real variables, we applied two variable-selection methods: the first method used hypothesis tests, correlation and random forest-based feature importance measures and the second method was only a random forest-based new approach (NAP), to select the best representative features for effective modelling and to compare them. We then built regression models based on various machine-learning algorithms ranging from logistic regression and support vector machines to an ensemble of gradient boosted trees and deep neural networks. Our results demonstrated that if lending institutions in the 2001s had used their own credit scoring model constructed by machine-learning methods explored in this study, their expected credit losses would have been lower, and they would be more sustainable. In addition, the deep neural networks and XGBoost algorithms trained on the subset selected by NAP achieve the highest area under the curve (AUC) and accuracy, respectively.

An accurate exchange rate forecasting and its decision-making to buy or sell are critical issues in the Forex market. Short-term currency rate forecasting is a challenging task due to its inherent characteristics, which include high volatility, trend, noise, and market shocks. We propose a novel deep learning architecture consisting of an adaptive activation function selection mechanism to achieve higher predictive accuracy. The proposed architecture is composed of seven neural networks that have different activation functions as well as softmax layer and multiplication layer with a skip connection, which are used to generate the dynamic importance weights that decide which activation function is preferred. In addition, we introduce an extended Min-Max smoothing technique to further normalize financial time series that have non-stationary properties. In our experimental evaluation, the results showed that our proposed model not only outperforms deep neural network baselines but also other classic machine learning approaches. The extended Min-Max smoothing technique is step towards forecasting non-stationary financial time series with deep neural networks.

A multivariate time series forecasting is critical in many applications, such as signal processing, finance, air quality forecasting, and pattern recognition. In particular, determining the most relevant variables and proper lag length from multivariate time series is challenging. This paper proposes an end-to-end recurrent neural network framework equipped with an adaptive input selection mechanism to improve the prediction performance for multivariate time series forecasting. The proposed model, named AIS-RNN, consists of two main components: the first neural network learns to generate context-dependent importance weights to dynamically select the input. The selected input is then fed into the second module for predicting the target variable. The experimental results show that our proposed end-to-end approach outperforms machine learning-based baselines on several public benchmark datasets. The AIS-LSTM model achieves higher performance on a public M3 dataset than the M3-specialized models. Furthermore, the AIS-RNN gives a beneficial advantage to interpret variable importance.

Early detection of infectious disease outbreaks is one of the important and significant issues in syndromic surveillance systems. It helps to provide a rapid epidemiological response and reduce morbidity and mortality. In order to upgrade the current system at the Korea Centers for Disease Control and Prevention (KCDC), a comparative study of state-of-the-art techniques is required. We compared four different temporal outbreak detection algorithms: the CUmulative SUM (CUSUM), the Early Aberration Reporting System (EARS), the autoregressive integrated moving average (ARIMA), and the Holt-Winters algorithm. The comparison was performed based on not only 42 different time series generated taking into account trends, seasonality, and randomly occurring outbreaks, but also real-world daily and weekly data related to diarrhea infection. The algorithms were evaluated using different metrics. These were namely, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), F1 score, symmetric mean absolute percent error (sMAPE), root-mean-square error (RMSE), and mean absolute deviation (MAD). Although the comparison results showed better performance for the EARS C3 method with respect to the other algorithms, despite the characteristics of the underlying time series data, Holt–Winters showed better performance when the baseline frequency and the dispersion parameter values were both less than 1.5 and 2, respectively.

Explaining dynamic relationships between input and output variables is one of the most important issues in time dependent domains such as economic, finance and so on. In this work, we propose a novel locally adaptive interpretable deep learning architecture that is augmented by recurrent neural networks to provide model explainability and high predictive accuracy for time-series data. The proposed model relies on two key aspects. First, the base model should be a simple interpretable model. In this step, we obtain our base model using a simple linear regression and statistical test. Second, we use recurrent neural networks to re-parameterize our base model to make the regression coefficients adaptable for each time step. Our experimental results on public benchmark datasets showed that our model not only achieves better predictive performance than the state-of-the-art baselines, but also discovers the dynamic relationship between input and output variables.