Past Data Research Articles

Patterns of visiting nurse services for older adults in Japanese residential facilities: A latent class analysis.

To examine the patterns of the visiting nurse services provided to older adults in a type of residential facility in Japan and to identify associated factors. This secondary analysis used past survey data from visiting nurse service agencies providing services to older adults in residential facilities with few nurses, known as non-specified facilities in Japan. Approximately 515 cases were used to determine the patterns of visiting nurse services using latent class analysis. Multinomial logistic regression analysis examined the relationships between identified classes, residents' characteristics, facilities, and services provided by visiting nurses. Three service patterns were identified as follows: Class 1, observational and follow-up care (37.1%); Class 2, chronic disease care (35.7%); and Class 3, end-of-life care (27.2%). Class 1 comprised fewer nursing services than the other classes and primarily involved the observation of medical conditions, whereas Classes 2 and 3 involved higher care-need levels and provided diverse nursing care support. Specifically, Class 3 was associated with the inclusion of family (odds ratio: 2.42) and a visiting nurse at the attached facility (odds ratio: 4.88). The three identified classes describe the older residents' healthcare needs. Additionally, the factors associated with the end-of-life care class suggest that older residents with these factors may have difficulty accessing end-of-life care by visiting nurses. Geriatr Gerontol Int 2023; 23: 326-333.

Transfer Learning for Fault Detection with Application to Wind Turbine SCADA Data

The installed wind power capacity is growing worldwide. Remote condition monitoring of wind turbines is employed to achieve higher uptimes and lower maintenance costs. Machine learning models can detect developing damages in wind turbines. Therefore, this paper demonstrates that cross–turbine transfer learning can drastically improve the accuracy of fault detection models in turbines with scarce SCADA data. In particular, it shows that combining the knowledge from turbines with scarce and turbines with plentiful data enables earlier detection of faults than prior art methods. Training fault detection models require large amounts of past and present SCADA data but these data are often unavailable or not representative of the current operation behavior. Newly commissioned wind farms lack SCADA data from the previous operation. Due to control software updates or hardware replacements, older turbines may also lack representative SCADA data. After such events, a turbine’s operation behavior can change significantly so its SCADA data no longer represent its current behavior. Therefore, the work highlights how to reuse and transfer knowledge across wind turbines to overcome this lack of data and enable the earlier detection of faults in wind turbines.

Nowcasting of Weather Parameters Impacting Solar PV Output Using Grey System Model

Objectives: To nowcast the weather parameters having a direct impact on the power output of the solar PV installations, with high prediction accuracy and a limited quantity of past data. Methods: In this study, the GM (1,1) model with Fourier series of error residuals has been proposed and used for forecasting the weather parameters namely Ambient temperature, Solar Photo Voltaic Module temperature, and Solar Irradiation. Real-time data has been used for showing the suitability of the proposed model to nowcast the weather parameters. The existing models like Autoregression and Double Exponential Smoothing are applied to the same data to prove the superiority of the GM (1,1) model with Fourier series of error residuals. Findings: It is found that the GM (1,1) model with Fourier series of error residuals is an apt model for nowcasting the weather parameters. The accuracy of the predicted result of this model on the real-time data ascertains the appropriateness of this model for nowcasting. The precision of the prediction accuracy of GM (1,1) with the Fourier series of error residuals model is verified by comparing it with other time series prediction models such as Autoregression and Double Exponential Smoothing algorithms. Novelty: Using GM (1,1) with the Fourier series of error residuals model for nowcasting the weather data is novel when compared with the existing algorithms because for nowcasting the weather parameters with accuracy, many of the existing algorithms require a huge volume of past data and involve complex computation. On the other hand, GM (1,1) with the Fourier series of error residuals model requires only a limited measure of past data and involves simple computation. Moreover, the accuracy of prediction is significantly higher than the other models. Keywords: Grey Theory; Renewable Energy; PV Installation; Auto Regression; Double Exponential Smoothing

Opioid death projections with AI-based forecasts using social media language

Targeting of location-specific aid for the U.S. opioid epidemic is difficult due to our inability to accurately predict changes in opioid mortality across heterogeneous communities. AI-based language analyses, having recently shown promise in cross-sectional (between-community) well-being assessments, may offer a way to more accurately longitudinally predict community-level overdose mortality. Here, we develop and evaluate, TrOP (Transformer for Opiod Prediction), a model for community-specific trend projection that uses community-specific social media language along with past opioid-related mortality data to predict future changes in opioid-related deaths. TOP builds on recent advances in sequence modeling, namely transformer networks, to use changes in yearly language on Twitter and past mortality to project the following year’s mortality rates by county. Trained over five years and evaluated over the next two years TrOP demonstrated state-of-the-art accuracy in predicting future county-specific opioid trends. A model built using linear auto-regression and traditional socioeconomic data gave 7% error (MAPE) or within 2.93 deaths per 100,000 people on average; our proposed architecture was able to forecast yearly death rates with less than half that error: 3% MAPE and within 1.15 per 100,000 people.

RF-BiLSTM Neural Network Incorporating Attention Mechanism for Online Ride-Hailing Demand Forecasting

Accurately predicting online ride-hailing demand can help operators allocate vehicle resources on demand, avoid idle time, and improve traffic conditions. However, due to the randomness and complexity of online ride-hailing demand data, which are affected by many factors and mostly time-series in nature, it is difficult to forecast accurately and effectively based on traditional forecasting models. Therefore, this study proposes an online ride-hailing demand forecasting model based on the attention mechanism of a random forest (RF) combined with a symmetric bidirectional long short-term memory (BiLSTM) neural network (Att-RF-BiLSTM). The model optimizes the inputs and can use past and future data to forecast, improving the forecasting precision of online ride-hailing demand. The model utilizes a random forest to filter and optimize the input variables to reduce the neural network complexity, and then an attention mechanism was incorporated into the BiLSTM neural network to construct a demand forecasting model and validate it using actual Uber pickup data from New York City. Compared with other forecasting models (Att-XGBoost-BiLSTM, Att-BiLSTM, and pure LSTM), the results show that the proposed symmetrical Att-RF-BiLSTM online ride-hailing demand forecasting model has a higher forecasting precision and fitting degree, which indicates that the proposed model can be satisfactorily applied to the area of online ride-hailing demand.

Comparative Analysis of Support Vector Machine And Perceptron Algorithms In Classification Of The Best Work Programs In P2KBP3A

With the rapid growth of government agencies that are required to carry out an activity in every aspect that publishes and carries out obligations every year, it is required to be held accountable and also implemented for every device that receives such as fostered villages by utilizing the available APBD funds to maximize the work program that has been designed. so that it can be implemented as much as possible. That way, to get the best from all aspects of every work program implementation, there must be an important point from the annual work program design that is made without exception. Data mining itself can help P2KBP3A in analyzing each work program that is designed before being implemented in the future for the annual work program by looking at various aspects of past work program data and grouping work programs in the form of classification. In designing the work program, this research builds a classification model by adding a sigmoid activation function that uses SVM and perceptron to compare the accuracy results of the algorithm used to get the best work program design. From the various classifications used, the best value for classifying the dataset of the best P2KBP3A work programs can be seen from the average accuracy value of 87.5%, F1 value of 82.2%, the precision value of 80.2%, and recall value of 87.5%.

State of Health Estimation Methods for Lithium-Ion Batteries

Contemporary lithium-ion batteries (LIBs) are one of the main components of energy storage systems that need effective management to extend service life and increase reliability and safety. Their characteristics depend highly on internal and external conditions (ageing, temperature, and chemistry). Currently, the state of batteries is determined using two parameters: the state of charge (SOC) and the state of health (SOH). Applying these two parameters makes it possible to calculate the expected battery life and a battery’s performance. There are many methods for estimating the SOH of batteries, including experimental, model-based, and machine learning methods. By comparing model-based estimations with experimental techniques, it can be concluded that the use of experimental methods is not applicable for commercial cases. The electrochemical model-based SOH estimation method clearly explains processes in the battery with the help of multidifferential equations. The machine learning method is based on creating a program trained to predict the battery’s state of health with the help of past ageing data. In this review paper, we analyze the research available in the literature in this direction. It is found that all methods used to assess the SOH of an LIB play an essential role, and each method has its pros and cons.

Road Risk-Index Analysis Using Satellite Products.

This paper proposes a service called intelligent routing using satellite products (IRUS) that can be used in order to analyze risks to the road infrastructure during bad weather conditions, such as heavy rainfall, storms, or floods. By diminishing movement risk, rescuers can arrive safely at their destination. To analyze these routes, the application uses both data provided by Sentinel satellites from the Copernicus program and meteorological data from local weather stations. Moreover, the application uses algorithms to determine the night driving time. From this analysis we obtain a risk index for each road provided by Google Maps API and then we present the path alongside the risk index in a friendly graphic interface. In order to obtain an accurate risk index, the application analyzes both recent and past data (up to 12 months).

Russian housing price forecast based on BP neural Network

Contemporarily, Russia's real estate market has huge potential for investors. The purpose of this paper is to investigated the impact of various factors on the real estate market price based on various statistical models. We collected past housing price data from the Internet. Subsequently, PCA model was adopted to reduce the dimension of the data to simplify the calculation. Finally, the BP model is implemented to predict the price changes in the real estate market. After calculation, our forecast data and the reality of housing prices are slightly different, but the forecast of the real estate market is consistent with the reality. In order to further find the influence of various factors on housing price, we also calculated the characteristic importance of different factors and found the most important influencing factor, i.e., the region. These results can provide investors with an idea to analyze the real estate market prices, and help policy makers find the real estate market prices in line with the market rules to avoid the occurrence of bubbles.

Investment in the Covid-19 Period

From 2019 to 2022, the covid-19 pandemic has lasted for three years. Its impact to society activity and economy is prominent. As a result, investors must change their investment strategy to suit the new environment. In this research, I used Markowitz model and Index Model to analyze the past data I collected and handled from past years in some specific conditions and the current information, like the impact of covid-19 to supply chain and Ukraine war, to correct the simulation coming from these two models. By the Markowitz model and Index model, people can find the weights of stocks in portfolio that can help they get maximum expected return over risk rate and minimum risk. Of course, people can some other suggestions from these weights, like which sector’s stocks is more stable or has higher interest rate. Additionally, I concluded some investment suggestions about the investment direction and strategies from some articles. They include the relationship between government debt and financial debt, impact of Covid-19 to different sectors, spread of collapse, natural resource volatility, investment diversification, value of bitcoin for investment, and others.

Domain-Incremental Cardiac Image Segmentation With Style-Oriented Replay and Domain-Sensitive Feature Whitening.

Contemporary methods have shown promising results on cardiac image segmentation, but merely in static learning, i.e., optimizing the network once for all, ignoring potential needs for model updating. In real-world scenarios, new data continues to be gathered from multiple institutions over time and new demands keep growing to pursue more satisfying performance. The desired model should incrementally learn from each incoming dataset and progressively update with improved functionality as time goes by. As the datasets sequentially delivered from multiple sites are normally heterogenous with domain discrepancy, each updated model should not catastrophically forget previously learned domains while well generalizing to currently arrived domains or even unseen domains. In medical scenarios, this is particularly challenging as accessing or storing past data is commonly not allowed due to data privacy. To this end, we propose a novel domain-incremental learning framework to recover past domain inputs first and then regularly replay them during model optimization. Particularly, we first present a style-oriented replay module to enable structure-realistic and memory-efficient reproduction of past data, and then incorporate the replayed past data to jointly optimize the model with current data to alleviate catastrophic forgetting. During optimization, we additionally perform domain-sensitive feature whitening to suppress model's dependency on features that are sensitive to domain changes (e.g., domain-distinctive style features) to assist domain-invariant feature exploration and gradually improve the generalization performance of the network. We have extensively evaluated our approach with the M&Ms Dataset in single-domain and compound-domain incremental learning settings. Our approach outperforms other comparison methods with less forgetting on past domains and better generalization on current domains and unseen domains.

A Data Transfer and Relevant Metrics Matching Based Approach for Heterogeneous Defect Prediction

Heterogeneous defect prediction (HDP) is a promising research area in the software defect prediction domain to handle the unavailability of the past homogeneous data. In HDP, the prediction is performed using source dataset in which the independent features (metrics) are entirely different than the independent features of target dataset. One important assumption in machine learning is that independent features of the source and target datasets should be relevant to each other for better prediction accuracy. However, these assumptions do not generally hold in HDP. Further in HDP, the selected source dataset for a given target dataset may be of small size causing insufficient training. To resolve these issues, we have proposed a novel heterogeneous data preprocessing method, namely, Transfer of Data from Target dataset to Source dataset selected using Relevance score (TDTSR), for heterogeneous defect prediction. In the proposed approach, we have used chi-square test to select the relevant metrics between source and target datasets and have performed experiments using proposed approach with various machine learning algorithms. Our proposed method shows an improvement of at least 14% in terms of AUC score in the HDP scenario compared to the existing state of the art models.

Frequency Response Function-Based Learning Control: Analysis and Design for Finite-Time Convergence

Iterative learning control (ILC) enables substantial performance improvement by using past operational data in combination with approximate plant models. The aim of this article is to develop an ILC framework based on nonparametric frequency response function (FRF) models that requires very limited modeling effort. These FRF models describe the behavior of a system in periodic steady state, yet are employed for the control of arbitrary finite-length tasks. A detailed analysis and design framework is developed to construct noncausal learning filters directly from uncertain FRF models, that achieve ILC convergence for arbitrary tasks. The resulting framework provides a unification between ILC and iterative inversion-based control, where the latter is a learning method specifically developed for periodic tasks.

THE USE OF NEURAL NETWORKS TO FORECAST THE NUMBER OF ROAD ACCIDENTS IN POLAND

Every year, a large number of traffic accidents occur on Polish roads. However, the pandemic of recent years has reduced the number of these accidents, although the number is still very high. For this reason, all measures should be taken to reduce this number. This article aims to forecast the number of road accidents in Poland. Thus, using Statistica software, the annual data on the number of road accidents in Poland were analyzed. Based on actual past data, a forecast was made for the future, for the period 2022-2040. Forecasting the number of accidents in Poland was conducted using selected neural network models. The results show that a reduction in the number of traffic accidents is likely. The choice of the number of random samples (learning, testing and validation) affects the results obtained.

Latent-Variable-Modeling-Based Fault Diagnosis for Unknown Nonlinear Distributed Parameter Systems with Application to Snap Curing Oven

A novel fault detection method based on time/space separation and latent variable model is proposed for unknown nonlinear distributed parameter systems in sensor-constrained environments. By performing time/space separation, the augmented matrix formed by the spatio-temporal distribution data of the DPSs can be split into a spatial basis function (SBFs) and a time series model, and the dimensionality reduction capability of the SBF is further utilized to obtain a low-order temporal model. Then, the temporal model is further extracted by using a dynamic latent variable modeling method to obtain the dominant time components and establish the corresponding monitoring statistics. Utilizing the appropriate kernel density function, confidence bounds are selected for the monitoring statistics when the system is normal. As a data-based fault diagnosis method, it requires only the past data records of the system and no reliance on complex mathematical models. Two sets of experiments performed on a snap curing oven verified the effectiveness of the proposed method.

Progress Of the Study on The Impact of Investor Sentiment on Stock Returns

Traditional pricing models do not reckon investor sentiment as a systematic factor of stock returns. However, past market events around the world have shown considerable return anomalies that are unexplainable by well-known market factors alone. It has inspired researchers to develop behavioral finance, that is, the study of investor irrationality and how it influences the stock market in a less efficient setting. This essay reviews three popular methods to research the impact of investor sentiment on stock returns in stock markets: conventional regression models, the GARCH model family, and the DSSW model, proposed by De Long et al. It is found that conventional regressions links investor sentiment with a significant effect on stock returns, with different patterns that reflect past data and predict future prices. The GARCH family explores the sentiment-return relationship in more detail, considering factors such as volatility, while the DSSW model gives a more behavioral account and captures the complexity of such patterns of effect. In the research of this paper, investor sentiment does influence stock returns as well as volatility in both markets, which implies a deeper understanding of the stock markets and calls for attention of retail investors, institutional investors, and financial regulators.

A mathematical model of demand with unpredictable salt supply: a case study in Madura

The proposed mathematical model aims to optimize the amount of demand based on the uncertainty of salt supply to meet the needs of salt in society. This study uses an econometric approach to model the supply and demand for salt based on the uncertainty of the amount of supply. The research results are the salt productivity model used to build a salt availability model. Meanwhile, the salt demand model is built based on past data about the number of requests for each period. The demand and supply functions for salt use a two-variable linear equation system. The salt field productivity models follow a quadratic polynomial pattern according to the difference in seasons at each time. The demand and supply of salt model follows a linear regression. The supply of salt in Indonesia is uncertain at any time. But, the demand for salt increased at the time. It is necessary to model the demand and supply of salt to meet the demand for salt

Long-term trajectory of the macroinvertebrate community in the downstream channel of a hydropower dam after the operation of a sediment bypass tunnel

A sediment bypass tunnel (SBT) is a dam facility used to mitigate reservoir sedimentation and is expected to contribute to river ecosystem restoration by delivering sediment to the degraded downstream channel. We analyzed past monitoring data of benthic invertebrates sampled downstream of the Asahi Dam in Nara Prefecture Japan to understand the long-term changes in the invertebrate community after the installation of the SBT. Before the SBT installation, the reaches upstream of the dam (one Up site) and downstream of the dam (two Down sites) showed different invertebrate communities, dominated by invertebrates that preferred clean cobble beds, sand and gravel layers, or run habitats in the former and invertebrates that preferred stable or porous beds or pool habitats in the latter. A significant increase in the Up–Down site community similarity within a few years after the SBT installation indicated a rapid recovery of the invertebrate community at the Down sites. During the 16 years of the after-SBT period, the total density and taxon richness at the Down sites increased in the early half, while they decreased in the latter half due to bed changes associated with the continuous accumulation of sediments. The change in dominant invertebrates from taxa of the large sizes to those of the small sizes and the increase in the interhabitat community similarity suggest that the bed of the Down sites became finer and flatter by sand and gravel that filled interstitial spaces of cobbles in the late stage. While the invertebrate community at the Down sites recovered dramatically by the sediment supply through the SBT, additional actions (e.g., cobble augmentation, installation of flow guiding structures) to maintain a balanced sediment size distribution and developed pool-riffle structures would be necessary to maintain the potential habitat heterogeneity and biodiversity in the long term. Our study also demonstrated that invertebrates responded to the sediment supply according to their habitat preferences, which suggests that the classifications of invertebrates by habitat-related traits (e.g., bed-residence types and pool-riffle types in this study) would help to properly understand the sediment and bed status of river reaches under channel degradation or aggradation.

Emotions and transcripts after a while: An interview with a ‘parachute kid’

AbstractThis paper discussions the implications of returning to ‘old’ interview data many years later and asks what can be learned from the different emotions that such revisiting can invoke in the researcher? It considers the significance of emotions in the process of analysing interview transcripts and how a researcher changes over time. It calls on researchers to ‘revisit’ past data for another look.

A critique of the use of modified and moving average‐based EWMA control charts

AbstractWe show that the so‐called modified exponentially weighted moving average (EWMA) control chart is very closely related to a moving average‐based EWMA chart. These charts have ill‐advised weighting functions that assign more weight to some past data values than to the most recent data values. We show that these weighting functions adversely affect their run length properties and that one can obtain better overall performance with an ordinary EWMA chart.