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.

Background: With advances in next-generation sequencing technologies, the bisulfite conversion of genomic DNA followed by sequencing has become the predominant technique for quantifying genome-wide DNA methylation at single-base resolution. A large number of computational approaches are available in literature for identifying differentially methylated regions in bisulfite sequencing data, and more are being developed continuously. Results: Here, we focused on a comprehensive evaluation of commonly used differential methylation analysis methods and describe the potential strengths and limitations of each method. We found that there are large differences among methods, and no single method consistently ranked first in all benchmarking. Moreover, smoothing seemed not to improve the performance greatly, and a small number of replicates created more difficulties in the computational analysis of BS-seq data than low sequencing depth. Conclusions: Data analysis and interpretation should be performed with great care, especially when the number of replicates or sequencing depth is limited.

Hematopoietic cancer is a malignant transformation in immune system cells. Hematopoietic cancer is characterized by the cells that are expressed, so it is usually difficult to distinguish its heterogeneities in the hematopoiesis process. Traditional approaches for cancer subtyping use statistical techniques. Furthermore, due to the overfitting problem of small samples, in case of a minor cancer, it does not have enough sample material for building a classification model. Therefore, we propose not only to build a classification model for five major subtypes using two kinds of losses, namely reconstruction loss and classification loss, but also to extract suitable features using a deep autoencoder. Furthermore, for considering the data imbalance problem, we apply an oversampling algorithm, the synthetic minority oversampling technique (SMOTE). For validation of our proposed autoencoder-based feature extraction approach for hematopoietic cancer subtype classification, we compared other traditional feature selection algorithms (principal component analysis, non-negative matrix factorization) and classification algorithms with the SMOTE oversampling approach. Additionally, we used the Shapley Additive exPlanations (SHAP) interpretation technique in our model to explain the important gene/protein for hematopoietic cancer subtype classification. Furthermore, we compared five widely used classification algorithms, including logistic regression, random forest, k-nearest neighbor, artificial neural network and support vector machine. The results of autoencoder-based feature extraction approaches showed good performance, and the best result was the SMOTE oversampling-applied support vector machine algorithm consider both focal loss and reconstruction loss as the loss function for autoencoder (AE) feature selection approach, which produced 97.01% accuracy, 92.60% recall, 99.52% specificity, 93.54% F1-measure, 97.87% G-mean and 95.46% index of balanced accuracy as subtype classification performance measures.

Proper demand forecasting for postal delivery service can be used for optimal logistic management, staff scheduling and topology planning. More especially, during special holidays, such as the Lunar New Year and the Chuseok (Mid-autumn day), the demand for delivery service increases sharply in South Korea. It makes a challenge to forecast demand to provide a normal delivery schedule for the Korean mail center. To address this problem, we propose a novel deep learning model equipped with selection and update layers (MLP-SUL) to achieve high predictive performance. The proposed model consists of three main parts: the first part of the model learns to generate context-dependent weights to decide which input feed to the next layer; the second part updates the weighted input to prepare encoded input, and the third part is a prediction layer that consists of a linear layer. A linear layer takes encoded input for forecasting demand. We also introduce a special data preprocessing step for our task that requires long-term forecasting. The experimental results show that our proposed deep learning model outperforms state-of-the-art baselines on Korean mail center datasets.