Limited Historical Data Research Articles

Does an Empirical Economic Relation Have a Life?

Does an Empirical Economic Relation Have a Life?

Stock Market Crisis Forecasting Using Neural Networks with Input Factor Selection

Artificial neural networks have gained increasing importance in many fields, including quantitative finance, due to their ability to identify, learn and regenerate non-linear relationships between targets of investigation. We explore the potential of artificial neural networks in forecasting financial crises with micro-, macroeconomic and financial factors. In this application of neural networks, a huge amount of available input factors, but limited historical data, often leads to over-parameterized and unstable models. Therefore, we develop an input variable reduction method for model selection. With an iterative walk-forward forecasting and testing procedure, we create out-of-sample predictions for crisis periods of the S&P 500 and demonstrate that the model selected with our method outperforms a model with a set of input factors taken from the literature.

A fuzzy Bayesian network-based approach for modeling and analyzing factors causing process variability

PurposeIn order to minimize the impact of variability on performance of the process, proper understanding of factors interdependencies and their impact on process variability (PV) is required. However, with insufficient/incomplete numerical data, it is not possible to carry out in-depth process analysis. This paper presents a qualitative framework for analyzing factors causing PV and estimating their influence on overall performance of the process.Design/methodology/approachFuzzy analytic hierarchy process is used to evaluate the weight of each factor and Bayesian network (BN) is utilized to address the uncertainty and conditional dependencies among factors in each step of the process. The weighted values are fed into the BN for evaluating the impact of each factor to the process. A three axiom-based approach is utilized to partially validate the proposed model.FindingsA case study is conducted on fused filament fabrication (FFF) process in order to demonstrate the applicability of the proposed technique. The result analysis indicates that the proposed model can determine the contribution of each factor and identify the critical factor causing variability in the FFF process. It can also helps in estimating the sigma level, one of the crucial performance measures of a process.Research limitations/implicationsThe proposed methodology is aimed to predict the process quality qualitatively due to limited historical quantitative data. Hence, the quality metric is required to be updated with the help of empirical/field data of PV over a period of operational time. Since the proposed method is based on qualitative analysis framework, the subjectivities of judgments in estimating factor weights are involved. Though a fuzzy-based approach has been used in this paper to minimize such subjectivity, however more advanced MCDM techniques can be developed for factor weight evaluation.Practical implicationsAs the proposed methodology uses qualitative data for analysis, it is extremely beneficial while dealing with the issue of scarcity of experimental data.Social implicationsThe prediction of the process quality and understanding of difference between product target and achieved reliability before the product fabrication will help the process designer in correcting/modifying the processes in advance hence preventing substantial amount of losses that may happen due to rework and scraping of the products at a later stage.Originality/valueThis qualitative analysis will deal with the issue of data unavailability across the industry. It will help the process designer in identifying root cause of the PV problem and improving performance of the process.

A data-enhanced distributionally robust optimization method for economic dispatch of integrated electricity and natural gas systems with wind uncertainty

With growing penetrations of wind power in electricity systems, the coordinated dispatch of integrated electricity and natural gas systems is becoming a popular research topic. Distributionally robust optimization can cope with the wind uncertainty of integrated electricity and natural gas systems by providing optimal solutions for the worst-case probability distribution. However, limited historical wind data hinder the estimation of worst-case probability distribution. As a breakthrough in artificial intelligence, generative adversarial networks can be established to approximate a complex uncertain probability distribution from raw data and generate realistic data subject to the identical distribution. This paper proposes a data-driven optimization method for economic dispatch of integrated electricity and natural gas systems with wind uncertainty, whose probability distribution is free. Based on limited historical data, the data-driven generative adversarial network generates artificial wind power data, which helps to improve the estimation of worst-case probability distribution in distributionally robust optimization. Moreover, the robustness of optimization solutions can be adjusted cost-effectively by controlling the auxiliary data number. In a case study, optimization solutions of the proposed method are shown to achieve a lower probability of chance constraint violation at a nearly negligible cost increase compared with those from four typical optimization methods.

Probabilistic hazard assessment: Application to geomagnetic activity

Probabilistic Hazard Assessment (PHA) provides an appropriate methodology for assessing space weather hazards and their impact on technology. PHA is widely used in geosciences to determine the probability of exceedance of critical thresholds, caused by one or more hazard sources. PHA has proved useful with limited historical data to estimate the likelihood of specific impacts. PHA has also driven the development of empirical and physical models, or ensembles of models, to replace measured data. Here we aim to highlight the PHA method to the space weather community and provide an example of how it could be used. In terms of space weather impact, the critical hazard thresholds might include the Geomagnetically Induced Current in a specific high voltage power transformer neutral, or the local pipe-to-soil potential in a particular metal pipe. We illustrate PHA in the space weather context by applying it to a twelve-year dataset of Earth-directed solar Coronal Mass Ejections (CME), which we relate to the probability that the global three-hourly geomagnetic activity index Kp exceeds specific thresholds. We call this a “Probabilistic Geomagnetic Hazard Assessment”, or PGHA. This provides a simple but concrete example of the method. We find that the cumulative probability of Kp > 6−, > 7−, > 8− and Kp = 9o is 0.359, 0.227, 0.090, 0.011, respectively, following observation of an Earth-directed CME, summed over all CME launch speeds and solar source locations. According to the historical Kp distribution, this represents an order of magnitude increase in the a priori probability of exceeding these thresholds. For the lower Kp thresholds, the results are somewhat distorted by our exclusion of coronal hole high-speed stream effects. The PHGA also reveals useful probabilistic associations between solar source location and subsequent maximum Kp for operational forecasters.

An Empirical Study of Software Metrics Diversity for Cross-Project Defect Prediction

Cross-project defect prediction (CPDP) is a mainstream method estimating the most defect-prone components of software with limited historical data. Several studies investigate how software metrics are used and how modeling techniques influence prediction performance. However, the software’s metrics diversity impact on the predictor remains unclear. Thus, this paper aims to assess the impact of various metric sets on CPDP and investigate the feasibility of CPDP with hybrid metrics. Based on four software metrics types, we investigate the impact of various metric sets on CPDP in terms of F-measure and statistical methods. Then, we validate the dominant performance of CPDP with hybrid metrics. Finally, we further verify the CPDP-OSS feasibility built with three types of metrics (orient-object, semantic, and structural metrics) and challenge them against two current models. The experimental results suggest that the impact of different metric sets on the performance of CPDP is significantly distinct, with semantic and structural metrics performing better. Additionally, trials indicate that it is helpful for CPDP to increase the software’s metrics diversity appropriately, as the CPDP-OSS improvement is up to 53.8%. Finally, compared with two baseline methods, TCA+ and TDSelector, the optimized CPDP model is viable in practice, and the improvement rate is up to 50.6% and 25.7%, respectively.

An Empirical Study of Training Data Selection Methods for Ranking-Oriented Cross-Project Defect Prediction.

Ranking-oriented cross-project defect prediction (ROCPDP), which ranks software modules of a new target industrial project based on the predicted defect number or density, has been suggested in the literature. A major concern of ROCPDP is the distribution difference between the source project (aka. within-project) data and target project (aka. cross-project) data, which evidently degrades prediction performance. To investigate the impacts of training data selection methods on the performances of ROCPDP models, we examined the practical effects of nine training data selection methods, including a global filter, which does not filter out any cross-project data. Additionally, the prediction performances of ROCPDP models trained on the filtered cross-project data using the training data selection methods were compared with those of ranking-oriented within-project defect prediction (ROWPDP) models trained on sufficient and limited within-project data. Eleven available defect datasets from the industrial projects were considered and evaluated using two ranking performance measures, i.e., FPA and Norm(Popt). The results showed no statistically significant differences among these nine training data selection methods in terms of FPA and Norm(Popt). The performances of ROCPDP models trained on filtered cross-project data were not comparable with those of ROWPDP models trained on sufficient historical within-project data. However, ROCPDP models trained on filtered cross-project data achieved better performance values than ROWPDP models trained on limited historical within-project data. Therefore, we recommended that software quality teams exploit other project datasets to perform ROCPDP when there is no or limited within-project data.

Is Inflammatory Bowel Disease a Contraindication for Pelvic XRT in Patients With Gynecologic Cancers? Efficacy and Toxicity Outcomes in the Modern Era

Pelvic radiation therapy (XRT) is an integral part of treatment for many patients with gynecologic malignancies to help decrease the risk of morbid recurrences and allow for organ preservation as an alternative to radical surgery. Historically, providers have been reluctant to offer pelvic XRT to inflammatory bowel disease (IBD) patients due to concerns over increased acute and late toxicities of treatment based on limited historical data, precluding many patients from receiving optimal therapy. This retrospective study investigated the use of pelvic XRT in patients with gynecologic cancers and IBD treated in the modern era to assess toxicity and efficacy.Patients with IBD and a gynecologic malignancy receiving pelvic XRT or chemo-XRT between January 2014 and June 2020 were included. Data abstracted included diagnosis, stage, XRT technique/dose, use of concurrent chemotherapy, and acute and late CTCAE v4.0 toxicity scores. Statistics were performed using statistical software.Six patients with IBD and FIGO stage I-III or recurrent cancer of the gynecological tract who received definitive or adjuvant XRT to the pelvis were identified. Median follow-up was 29 months (range 6-66 months). Four patients (66%) had Crohn's Disease and two (33%) had ulcerative colitis. Primary disease locations were uterus (2), vulva (1), vagina (2), and cervix (1). Four patients received concurrent platinum-based chemotherapy with pelvic XRT. Pelvic XRT was delivered using IMRT/VMAT in all patients with doses of 45-57.6 Gy in 25-32 fractions. 100% of patients completed treatment without treatment breaks. No patients received treatment for active IBD during XRT, but most (83%) were on loperamide at baseline. No patients experienced any acute 3 or greater toxicity; most common toxicities reported during treatment were grade 1-2 fatigue (100%), grade 1 diarrhea (50%), grade 1 nausea (66%), and grade 1-2 dermatitis (66%). One patient experienced grade 2 rectal proctitis 5 years out from treatment; no other patients experienced late GI toxicity at last recorded follow-up. Two patients (33%) developed distant disease within 6 months of completing XRT, one of whom died of disease. No patient developed locoregional recurrence and 66% of patients remain disease free at the time of last follow-up. 83% were able to avoid upfront radical surgery, including 3 patients who would have required total pelvic exenteration.In our single-institution series, pelvic XRT +/-chemotherapy is well-tolerated in patients with gynecologic malignancies and IBD. The use of pelvic XRT in this population led to excellent loco-regional control outcomes that allowed patients to avoid radical upfront surgeries and preserve organ function. Our small series represents the largest cohort of IBD patients with gynecologic cancers treated with pelvic XRT or chemo-XRT and suggests that strong consideration should be given to offering pelvic XRT to IBD patients diagnosed with gynecologic cancers.S.R. Amarnath: None.

Forecast with forecasts: Diversity matters

Forecast combinations have been widely applied in the last few decades to improve forecasting. Estimating optimal weights that can outperform simple averages is not always an easy task. In recent years, the idea of using time series features for forecast combinations has flourished. Although this idea has been proved to be beneficial in several forecasting competitions, it may not be practical in many situations. For example, the task of selecting appropriate features to build forecasting models is often challenging. Even if there was an acceptable way to define the features, existing features are estimated based on the historical patterns, which are likely to change in the future. Other times, the estimation of the features is infeasible due to limited historical data. In this work, we suggest a change of focus from the historical data to the produced forecasts to extract features. We use out-of-sample forecasts to obtain weights for forecast combinations by amplifying the diversity of the pool of methods being combined. A rich set of time series is used to evaluate the performance of the proposed method. Experimental results show that our diversity-based forecast combination framework not only simplifies the modeling process but also achieves superior forecasting performance in terms of both point forecasts and prediction intervals. The value of our proposition lies on its simplicity, transparency, and computational efficiency, elements that are important from both an optimization and a decision analysis perspective.

A Review of Risk Analysis Research for the Operations of Autonomous Underwater Vehicles

Risk analysis for autonomous underwater vehicles (AUVs) is essential to assist decision making for safer operations. This study aims to provide a systematic review of risk analysis research to enhance the safety performance of AUVs. Forty-two domain articles were retrieved and analyzed. Critical risk factors and causal relationships of AUV operations were identified. A comparative analysis of evolving methods and models was performed by categorizing them as qualitative, semi-quantitative, and quantitative. Future trends of research in this field were also outlined. The study observes that as AUV technologies gradually mature, environmental factors, human factors, and their interactive impacts are gathering more attention. Quantitative risk analysis methods have recently played a key role in improving the accuracy and handling the uncertainties of risk estimation. The study recommends devoting efforts to dynamic risk analysis, addressing limited historical data, intelligent risk analysis, and multi-vehicles risk analysis for future works. This study is expected to help AUV stakeholders gain comprehensive insights into fundamental concepts and evolving methods for risk analysis of AUVs. At the same time, it is expected to highlight future directions to bridge existing gaps.

A data-driven approach for tool wear recognition and quantitative prediction based on radar map feature fusion

Tool wear monitoring during the cutting process is crucial for ensuring part quality and productivity. A data-driven monitoring approach based on radar map feature fusion is proposed for tool wear recognition and quantitative prediction, aiming at tracking the evolution of tool wear comprehensively. Specifically, the sensitive features from multi-source signals are fused by a radar map, and health indicators capable of characterizing the tool wear evolution are obtained. For the recognition of tool wear state and the quantitative prediction of tool wear values, the Adaboost Decision Tree (Adaboost-DT) ensemble learning model and stacked bi-directional long short-term memory (SBiLSTM) deep learning network are established, respectively. Experimental results demonstrated that the proposed approach could recognize the current wear state quickly and accurately whilst predicting wear values based on limited historical data available. Combining tool wear recognition and prediction results contributes to making a more flexible tool replacement decision in intelligent manufacturing processes.

Using generative adversarial networks to evaluate robustness of reinforcement learning agents against uncertainties

This paper describes the process of creating uncertainty-infused synthetic profiles of building performance. The synthetic profiles are utilized as a resource for evaluating the response of trained machine learning models to unseen events. Applications of the introduced method can benefit researchers and practitioners who train data-driven building models on limited historical data and is particularly useful when a physics-based model of the building is unavailable. As an original contribution, we propose a conditional deep convolutional Generative Adversarial Network (GAN) for projecting multi-dimensional time-series profiles of building performance. The proposed GAN reflects climate and operation variations into the synthetic building performance profiles, while preserving the internal consistency within the generated data. To ensure high quality synthetic profiles, this study validates the plausibility of generated data through qualitative (visualization) and quantitative (Pearson correlation, Wasserstein distance) assessments. Synthetic profiles are fed to a trained reinforcement learning model and a rule-based controller to compare their performances in the presence of uncertainty. Results show that with limited training data, a reinforcement learning model's response can be fairly sensitive to uncertainties and disturbances, insofar, some advantages over rule-based controllers may be overestimated. To ensure the reproducibility of the presented results, this study is conducted on open data and models are shared as open source.

A Hybrid Energy System Workflow for Energy Portfolio Optimization

This manuscript develops a workflow, driven by data analytics algorithms, to support the optimization of the economic performance of an Integrated Energy System. The goal is to determine the optimum mix of capacities from a set of different energy producers (e.g., nuclear, gas, wind and solar). A stochastic-based optimizer is employed, based on Gaussian Process Modeling, which requires numerous samples for its training. Each sample represents a time series describing the demand, load, or other operational and economic profiles for various types of energy producers. These samples are synthetically generated using a reduced order modeling algorithm that reads a limited set of historical data, such as demand and load data from past years. Numerous data analysis methods are employed to construct the reduced order models, including, for example, the Auto Regressive Moving Average, Fourier series decomposition, and the peak detection algorithm. All these algorithms are designed to detrend the data and extract features that can be employed to generate synthetic time histories that preserve the statistical properties of the original limited historical data. The optimization cost function is based on an economic model that assesses the effective cost of energy based on two figures of merit: the specific cash flow stream for each energy producer and the total Net Present Value. An initial guess for the optimal capacities is obtained using the screening curve method. The results of the Gaussian Process model-based optimization are assessed using an exhaustive Monte Carlo search, with the results indicating reasonable optimization results. The workflow has been implemented inside the Idaho National Laboratory’s Risk Analysis and Virtual Environment (RAVEN) framework. The main contribution of this study addresses several challenges in the current optimization methods of the energy portfolios in IES: First, the feasibility of generating the synthetic time series of the periodic peak data; Second, the computational burden of the conventional stochastic optimization of the energy portfolio, associated with the need for repeated executions of system models; Third, the inadequacies of previous studies in terms of the comparisons of the impact of the economic parameters. The proposed workflow can provide a scientifically defendable strategy to support decision-making in the electricity market and to help energy distributors develop a better understanding of the performance of integrated energy systems.

Short-Term Load Forecasting Using Convolutional Neural Networks in COVID-19 Context: The Romanian Case Study

Short-term load forecasting (STLF) is fundamental for the proper operation of power systems, as it finds its use in various basic processes. Therefore, advanced calculation techniques are needed to obtain accurate results of the consumption prediction, taking into account the numerous exogenous factors that influence the results’ precision. The purpose of this study is to integrate, additionally to the conventional factors (weather, holidays, etc.), the current aspects regarding the global COVID-19 pandemic in solving the STLF problem, using a convolutional neural network (CNN)-based model. To evaluate and validate the impact of the new variables considered in the model, the simulations are conducted using publicly available data from the Romanian power system. A comparison study is further carried out to assess the performance of the proposed model, using the multiple linear regression method and load forecasting results provided by the Romanian Transmission System Operator (TSO). In this regard, the Mean Squared Error (MSE), the Mean Absolute Error (MAE), the Mean Absolute Percentage Error (MAPE), and the Root Mean Square Error (RMSE) are used as evaluation indexes. The proposed methodology shows great potential, as the results reveal better error values compared to the TSO results, despite the limited historical data.

Early-stage cost estimation model for power generation project with limited historical data

PurposeReliable conceptual cost estimation of large-scale construction projects is critical for successful project planning and execution. For addressing the limited data availability in conceptual cost estimation, this study proposes an enhanced ANN-based cost estimating model that incorporates artificial neural networks, ensemble modeling and a factor analysis approach.Design/methodology/approachIn the ANN-based conceptual cost estimating model, the ensemble modeling component enhances training, and thus, improves its predictive accuracy and stability when project data quantity is low; and the factor analysis component finds the optimal input for an estimating model, rendering explanations of project data more descriptive.FindingsOn the basis of the results of experiments, it can be concluded that ensemble modeling and FAMD (Factor Analysis of Mixed Data) are both conjointly capable of improving the accuracy of conceptual cost estimates. The ANN model version combining bootstrap aggregation and FAMD improved estimation accuracy and reliability despite these very low project sample sizes.Research limitations/implicationsThe generalizability of the findings is hard to justify since it is difficult to collect cost data of construction projects comprehensively. But this difficulty means that our proposed approaches and findings can provide more accurate and stable conceptual cost forecasting in the early stages of project development.Originality/valueFrom the perspective of this research, previous uses of past-project data can be deemed to have underutilized that information, and this study has highlighted that — even when limited in quantity — past-project data can and should be utilized effectively in the generation of conceptual cost estimates.

Comparative Study of COVID-19 Pandemic Progressions in 175 Regions in Australia, Canada, Italy, Japan, Spain, U.K. and USA Using a Novel Model That Considers Testing Capacity and Deficiency in Confirming Infected Cases.

Not identified as being exposed or infected, the group of asymptomatic and presymptomatic patients has become the key source of infectious hosts for the COVID-19 pandemic, triggering the re-emergence of outbreaks. Acknowledging the impacts of movement of unidentified patients and the limited testing capacity on understanding the spread of the virus, an augmented Susceptible-Exposed-Infectious-Confirmed-Recovered (SEICR) model integrating intercity migration data and testing capacity is developed to probe into the number of unidentified COVID-19 infected patients. This model allows evaluation of the effectiveness of active interventions, and more accurate prediction of the pandemic progression in a country, region or city. A pseudo-coevolutionary algorithm is adopted in the model fitting to provide an effective estimation of high-dimensional unknown parameter sets using a limited amount of historical data. The model is applied to 175 regions in Australia, Canada, Italy, Japan, Spain, the UK and USA to estimate the number of unconfirmed cases using limited historical data. Results showed that the actual number of infected cases could be 4.309 times as many as the official confirmed number. By implementing mass COVID-19 testing, the number of infected cases could be reduced by about 50%.

Particle Swarm Optimization for Least Square Support Vector Machine in MediumTerm Electricity Price Prediction

Predicting electricity price has now become an important task for planning and maintenance of power system. In medium-term forecast, electricity price can be predicted for several weeks ahead, up to a few months or a year ahead. It is useful for resources reallocation where the market players have to manage the price risk on the expected market scenario. However, the research on medium-term price forecast have also exhibited low forecast accuracy due to the limited historical data for training and testing purposes. Therefore, an optimization technique using Particle Swarm Optimization (PSO) for Least Square Support Vector Machine (LSSVM) was developed in this study to provide an accurate electricity price forecast with optimized LSSVM parameters. After thorough database mining in English language, no literature has been found on parameter optimization using LSSVM-PSO for medium-term price prediction. The model was examined on the Ontario power market which was reported as among the most volatile market worldwide. Monthly average of Hourly Ontario Electricity Price (HOEP) for the past 12 months and month index were selected as the input. The developed LSSVM-PSO showed higher forecast accuracy with lower complexity than the existing models.

Reliability Modelling and Analysis of the Power Take-Off System of an Oscillating Wave Surge Converter

Wave power is a potential technology for generating sustainable renewable energy. Several types of wave energy converters (WECs) have been proposed for this purpose. WECs operate in a harsh maritime environment that sets strict limitations on how and when the device can be economically and safely reached for maintenance. Thus, to ensure profitable energy generation over the system life cycle, system reliability is a key aspect to be considered in WEC development. In this article, we describe a reliability analysis approach for WEC development, based on the use of reliability block diagram (RBD) modelling. We apply the approach in a case study involving a submerged oscillating wave surge converter device concept that utilizes hydraulics in its power take-off system. In addition to describing the modelling approach, we discuss the data sources that were used for gathering reliability data for the components used in a novel system concept with very limited historical or experimental data available. This includes considerations of the data quality from various sources. As a result, we present examples of applying the RBD modelling approach in the context of WEC development and discuss the applicability of the approach in supporting the development of new technologies.

Personalized Multimodal Travel Service Design for sustainable intercity transport

Efficient multimodal public transport has been recognized as an effective solution for realizing sustainable intercity transport. However, it has not been well accepted by the general public due to the poor coordination among transport modes, inconvenience caused by transfers, difficulties on route plannings, etc. To address the problem and facilitate the popularity of multimodal intercity travel, this research proposed the design of personalized multimodal travel service based on SPSS (Smart Product Service System). It not only could provide multimodal route recommendations based on individual preferences, but also could improve its performance along usage. Meanwhile, the influencing factors of multimodal intercity travel is investigated, based on which the travel choice model is developed. Besides, a few-short learning method was proposed, which could mine the individual preference even from limited historical travel data. Finally, an experimental case study was conducted that demonstrates the effectiveness of the proposed solution.

Integrating data-driven and simulation models to predict traffic state affected by road incidents

ABSTRACT Predicting the traffic conditions in urban networks is a priority for traffic management centres. This becomes very challenging, especially when the network is affected by traffic incidents that vary in both time and space. Although data-driven modelling can be considered an ideal tool for short-term traffic predictions, its performance is severely degraded if little historical traffic information is available under incident conditions. This paper addresses this challenge by integrating data-driven and traffic simulation modelling approaches. Instead of directly predicting the traffic states using limited historical data, we employ a traffic simulation reinforced by data-driven models. The traffic simulation uses newly reported incident information and the estimated origin-destination (OD) demand flows to capture the complex interaction between drivers and road network, and predicts traffic states under extreme conditions. We showcase the capability of the proposed data-driven enforced traffic simulation platform for incident impact analysis in a real-life sub-network in Sydney, Australia.