In this paper, our goal is to generate synthetic data for heterogeneous (mixed-type) tabular datasets with high machine learning utility (MLu). Since the MLu performance depends on accurately approximating the conditional distributions, we focus on devising a synthetic data generation method based on conditional distribution estimation. We introduce MaCoDE by redefining the consecutive multi-class classification task of Masked Language Modeling (MLM) as histogram-based non-parametric conditional density estimation. Our approach enables the estimation of conditional densities across arbitrary combinations of target and conditional variables. We bridge the theoretical gap between distributional learning and MLM by demonstrating that minimizing the orderless multi-class classification loss leads to minimizing the total variation distance between conditional distributions. To validate our proposed model, we evaluate its performance in synthetic data generation across 10 real-world datasets, demonstrating its ability to adjust data privacy levels easily without re-training. Additionally, since masked input tokens in MLM are analogous to missing data, we further assess its effectiveness in handling training datasets with missing values, including multiple imputations of the missing entries.

In stock price forecasting, modeling the probabilistic dependence between stock prices within a time-series framework has remained a persistent and highly challenging area of research. We propose a novel model to explain the extreme co-movement in multivariate data with time-series dependencies. Our model incorporates a Hawkes process layer to capture abrupt co-movements, thereby enhancing the temporal representation of market dynamics. We introduce dynamic hypergraphs into our model adapting to higher-order (groupwise rather than pairwise) relationships within the stock market. Extensive experiments on real-world benchmarks demonstrate the robustness of our approach in predictive performance and portfolio stability.

Recent advances in a generative neural network model extend the development of data augmentation methods. However, the augmentation methods based on the modern generative models fail to achieve notable improvement in class imbalance data compared to the conventional model, Synthetic Minority Oversampling Technique (SMOTE). We investigate the problem of the generative model for imbalanced classification and introduce a framework to enhance the SMOTE algorithm using Variational Autoencoders (VAE s). Our approach systematically quantifies the density of data points in a low-dimensional latent space using the VAE, simultaneously incorporating information on class labels and classification difficulty. Then, the data points potentially degrading the augmentation are systematically excluded, and the neighboring observations are directly augmented on the data space. Empirical studies on several imbalanced datasets represent that this simple process innovatively improves the conventional SMOTE algorithm over the deep learning models. Consequently, we conclude that the selection of minority data and the interpolation in the data space are beneficial for imbalanced classification problems with a relatively small number of data points.

The Gaussianity assumption in Variational AutoEncoders (VAEs) enhances computational efficiency and provides a solid theoretical basis for estimating probability distributions. However, we have empirically found that approximating distributions with non-smooth densities using the Gaussian VAE is challenging. Therefore, we propose an approach for distributional learning in VAEs that extends to estimating the quantile function while accommodating both smooth and non-smooth densities. This is achieved by utilizing the continuous ranked probability score, a strictly proper scoring rule, as our reconstruction loss. Our method can be seen as a specialized form of a nonparametric M-estimator for estimating general quantile functions, and we establish a theoretical connection between our model and quantile estimation. Furthermore, we demonstrate that our reconstruction loss functions as the lower bound of an infinite mixture of asymmetric Laplace distributions, which allows our synthetic data generation mechanism to maintain differential privacy. We validate the effectiveness of our model in capturing the underlying distribution through experiments involving synthetic data generation on real-world tabular datasets, showing that the level of data privacy can be easily adjusted.

In recent years, there has been a growing interest in probabilistic forecasting methods that offer more comprehensive insights by considering prediction uncertainties rather than point estimates. This paper introduces a novel variational autoencoder learning framework for multivariate distributional forecasting. Our approach employs distributional learning to directly estimate the cumulative distribution function of future time series conditional distributions using the continuous ranked probability score. By incorporating a temporal structure within the latent space and utilizing versatile quantile models, such as the generalized lambda distribution, we enable distributional forecasting by generating synthetic time series data for future time points. To assess the effectiveness of our method, we conduct experiments using a multivariate dataset of real cryptocurrency prices, demonstrating its superiority in forecasting high-volatility scenarios.

Electricity is essential in daily life and it is difficult to imagine a life without it. When its supply is not sufficient, people's daily activities can be interrupted. Small sellers with batteries can form an electricity market and they can supply a supplement amount of electricity that is needed during a peak consumption time instead of building a large-scale power plant. As a case study, Saudi Arabia's electricity market is studied. The people in Saudi Arabia, especially in hot regions such as Jazan (or Jizan), experience the interruption of the electricity especially during the summer season for two reasons [1]. The first reason is Jazan has hot weather and whenever the hot weather increases, the demand of the electricity will increase. The second reason is there are few power plants that can provide enough electricity for people. This paper will explore the possibility of generating electricity using numerous solar panels with batteries by simulation.

One of the main goals of the electronic health record (EHR) system is to empower patients to access to their own medical decisions. However, medical data is largely coming from clinical institutions so there is no way for them to control or maintain their own medical record. Patients or guardians may need to keep track of their medical data such as observed symptoms or measurements that may not be available in the EHR. Additionally, clinical decision without patient medical history can be error-prone and even be detrimental. Personal medical condition history is considered as the one of the weakest links in the current healthcare systems. For those reasons, it is necessary to have an effective and efficient personal health record system (PHRS) that allows patients or guardians to constantly monitor and control the personal health record. We propose a cloud based personal health record system that allows constant monitoring capability by supporting dynamic creation of clinical document architecture (CDA) document from a mobile device. The generated CDA document may be used to assess current health against major diseases through a clinical decision support system. We provide constant monitoring capability by using easy uploading module and decision support system. Our proposed system uses medical coding standards such as ICD-9-CM, SNOMED CT, etc. to achieve interoperability between different electronic health record systems.

With development of Internet and technology, the traditional market has become an electronic market, and the value of trading has become more than a traditional market. The SME e-commerce has been grown rapidly, and more businesses have used the e-commerce for selling and buying goods or services. Consequently, most of the plastic e-marketplaces only allow the large companies to search the product and then contact the other large companies by email or phone. The e-marketplace needs an automated or intelligent agent in order to automated purchasing consortium among the SMEs. The consortium agents allow SMEs to make purchasing consortium by exchanging electronic messages in a multi-agent system. In addition, most of the e-marketplace does not support the purchasing consortium. This paper implements an automated purchasing consortium by applying Modified Knapsack Problem Algorithm, and by eliminating the broker agent in the consortium. Also, the efficiency of the simulated system in the plastic market is demonstrated by a series of simulations in which the result is Modified knapsack Algorithm Consortium and Exactly Consortium.

The extent of glacier terminus displacement is instrumental in investigations of natural or artificial geographic changes. Its importance to earth science and engineering is reflected in the considerable efforts that have been devoted to the development of several boundary displacement analysis methods. Among the methods, the buffering-based approach compares favorably with other approaches in objectivity and robustness. However, it does not consider the relative positions of boundaries, because its buffering operation cannot determine features' relative directions. This limitation incurs inaccurate calculation results – underestimation of mean shifts and overestimation of shape variations, especially when the two compared boundaries intersect. Discrete displacement analysis (DDA), an alternative method that considers given geographic objects as a set of a finite number of points, is proposed here. In a series of tests carried out, including Jakobshavn glacier's calving front, DDA was found to correctly calculate mean shift and shape variation even in cases where the conventional buffering-based method failed. Moreover, this approach is independent of the dimension of space in which it is implemented, and thus is expected to be utilized for analysis of 3D geographic object displacement.

There have been numerous suggestions for improving the quality of service (QoS) of the Internet amid network congestion. However, due to the nature of the Internet service - best-effort delivery and distributed - it has been difficult to achieve. One dominant idea of solving the congestion problem is to have varying price for Internet access. The concept of smart market, proposed by MacKie-Mason et al. (1994), can help build an efficient Internet pricing structure that can be used to provide QoS when the network is congested. Efficient pricing structure can improve QoS of the Internet service by discouraging unnecessary usage of the Internet while giving access to the user who has "need-to-use" of the Internet through the pricing mechanism. In this paper, we propose a conceptual model for the pricing structure that utilizes auction market model as an implementation of the smart market. As a proof of concept, we have implemented a network sorting algorithm in CLIPS that implements auction part of the model.

Data classification and prediction problems are prevalent in many domains. The need to predict to which class a particular data point belongs has been seen in areas such as medical diagnosis, credit rating, Web filtering, prediction, and stock rating. This has led to strong interest in developing systems that can accurately classify data and predict outcome. The classification is typically based on the feature values of objects being classified. Often, a form of ordering relation, defined by feature values, on the objects to be classified is known. For instance, the objects belonging to one class have larger (or smaller) feature values than do those in the other class. Exploiting this characteristic of isotonicity, we propose a data-classification method called isotonic separation based on linear programming, especially network programming. The paper also addresses an extension of the isotonic-separation method for continuous outcome prediction. Applications of the isotonic separation for discrete outcome prediction and its extension for continuous outcome prediction are shown to illustrate its applicability.