Input Dimension Research Articles

Machine Learning Prediction of Global Photovoltaic Energy in Spain

The growing presence of solar energy in the electrical systems of many countries has made its accurate forecasting an important issue. In this work we will explore the application of Support Vector Regression (SVR), an advanced Machine Learning modelling tool, to forecast the daily photovoltaic generation of Spain. Given the very large geographical spread of photovoltaic installations, we will use as input features NWP forecasts of relevant meteorological variables for the entire Iberian Peninsula. The input dimension is thus very large but, while further work is needed, our results show SVR to be an effective tool to deal with the problem's underlying dimension, yield useful forecasts and further provide some insights on the relationship between NWP and actual solar energy production.

The Impact of Fintech Innovation on Bank’s Performance: Evidence from the Kingdom of Bahrain

This study investigates the impact of bank-level Financial Technology (Fintech) innovations on banks’ performance in the Kingdom of Bahrain from 2012-2021. Annual data of banks listed in Bahrain’s Bourse has been utilized to achieve this objective. In addition, bank-level FinTech indices have been constructed by a textual analysis method that assesses input dimensions through 32 keywords under four categories, including artificial intelligence, blockchain, cloud computing, and big data; FinTech output dimension is evaluated through payment and settlement, resource allocation, risk management, and channel construction technology. Using different panel data estimators, such as pooled OLS, fixed effects, random effects, and panel-corrected standard errors linear regression, results show that FinTech innovations increase banks’ performance. In addition, the findings demonstrate that banks' capital adequacy ratio, earnings ability, total assets, and annual GDP growth rate also significantly positively impact bank performance. Years in business have a significant adverse effect on bank performance. Interestingly, conventional and state-owned banks have a higher positive impact on returns on assets (ROA) than Islamic and privately owned banks. Policymakers and investors should pay close attention to facilitating ongoing FinTech innovations in the Kingdom of Bahrain to create opportunities and build a more inclusive and efficient financial sector.

Towards federated transfer learning in electrocardiogram signal analysis

Modern methods in artificial intelligence perform very well on many healthcare datasets, at times outperforming trained doctors. However, many assumptions made in model training are not justifiable in clinical settings. In this work, we propose a method to train classifiers for electrocardiograms, able to deal with data of disparate input dimensions, distributed across different institutions, and able to protect patient privacy. In addition, we propose a simple method for creating federated datasets from any centralized dataset. We use autoencoders in conjunction with federated learning to model a highly heterogeneous modeling problem using the Massachusetts Institute of Technology Beth Israel Hospital Arrhythmia dataset, the Computing in Cardiology 2017 challenge dataset, and the PTB-XL dataset. For an encoding dimension of 1000, our federated classifier achieves an accuracy, precision, recall, and F1 score of 73.0%, 66.6%, 73.0%, and 69.7%, respectively. Our results suggest that dropping commonly made assumptions significantly complicate training and that as a result, estimates of performance of many machine learning models may overestimate performance when adopted for clinical settings.

Target setting using an enhanced inverse DEA approach: application to Spanish universities

This paper proposes a non-oriented, non-radial Inverse Data Envelopment Analysis (InvDEA) approach for target setting. The changes of any input and output dimensions may be fixed. In addition to the corresponding multiobjective optimisation model, the approach includes a prior step that checks the feasibility of the input and output changes that are fixed and revises them if necessary. This Pre-InvDEA model uses Goal Programming (GP) and guarantees the feasibility of those changes. The approach also includes a Post-InvDEA multiobjective optimisation model to optimise the input and output changes given in order to guarantee that the computed target is efficient. The overall target-setting approach is conceived as an interactive process so that the three steps (Pre-InvDEA, InvDEA, and Post-InvDEA) may be run more than once until a satisfactory target is determined. The application of the proposed approach to computing efficient targets for Spanish universities is presented.

Elitist-opposition-based artificial electric field algorithm for higher-order neural network optimization and financial time series forecasting

This study attempts to accelerate the learning ability of an artificial electric field algorithm (AEFA) by attributing it with two mechanisms: elitism and opposition-based learning. Elitism advances the convergence of the AEFA towards global optima by retaining the fine-tuned solutions obtained thus far, and opposition-based learning helps enhance its exploration ability. The new version of the AEFA, called elitist opposition leaning-based AEFA (EOAEFA), retains the properties of the basic AEFA while taking advantage of both elitism and opposition-based learning. Hence, the improved version attempts to reach optimum solutions by enabling the diversification of solutions with guaranteed convergence. Higher-order neural networks (HONNs) have single-layer adjustable parameters, fast learning, a robust fault tolerance, and good approximation ability compared with multilayer neural networks. They consider a higher order of input signals, increased the dimensionality of inputs through functional expansion and could thus discriminate between them. However, determining the number of expansion units in HONNs along with their associated parameters (i.e., weight and threshold) is a bottleneck in the design of such networks. Here, we used EOAEFA to design two HONNs, namely, a pi-sigma neural network and a functional link artificial neural network, called EOAEFA-PSNN and EOAEFA-FLN, respectively, in a fully automated manner. The proposed models were evaluated on financial time-series datasets, focusing on predicting four closing prices, four exchange rates, and three energy prices. Experiments, comparative studies, and statistical tests were conducted to establish the efficacy of the proposed approach.

Automatic stroke classification: Domain knowledge injection augmented transfer learning approach

Aim: To build an artificial intelligence model to classify stroke into ischemic or hemorrhagic classes using the labeled stroke CT slices that were shared in the 2021 Teknofest artificial intelligence in health competition
 Methods: We developed a set of methods which can inject domain knowledge to the models to provide a more refined search space for the model for better performance. We used pretrained MobileNet and EfficientNet architectures and fine-tuned them for our 2-class output model. We discarded impossible pixel values and pixel spatial locations to provide a space which was conditioned into only possible spatial locations and signal values using our knowledge about brain anatomy, stroke pathology and imaging.
 Results: With the dataset which we just used [0-1] normalization and adjusted the input dimension into 224*224, accuracy values of 0.74 with adapted MobileNetV2 and 0.72 with adapted EfficentNetB0 were obtained in the group without further pre-processing. In the data transformation group where bone structures were removed and pixel values were restricted by eliminating impossible values, an accuracy level of 0.91 with MobileNetV2 and 0.88 with EfficientNetB0 at test time were achieved.
 Conclusion: In conclusion, CT based slice wise prediction of mechanism of stroke as ischemic or hemorrhagic was achieved with high accuracy by integrating human knowledge to the pretrained off the shelf models which was promising to shorten the time of the triage of stroke patients which can potentially improve stroke patient outcomes

Universal approximation properties for an ODENet and a ResNet: Mathematical analysis and numerical experiments

We prove a universal approximation property (UAP) for a class of ODENet and a class of ResNet, which are simplified mathematical models for deep learning systems with skip connections. The UAP can be stated as follows. Let $ n $ and $ m $ be the dimension of input and output data, and assume $ m\leq n $. Then we show that ODENet of width $ n+m $ with any non-polynomial continuous activation function can approximate any continuous function on a compact subset on $ \mathbb{R}^n $. We also show that ResNet has the same property as the depth tends to infinity. Furthermore, we derive the gradient of a loss function explicitly with respect to a certain tuning variable. We use this to construct a learning algorithm for ODENet. To demonstrate the usefulness of this algorithm, we apply it to a regression problem, a binary classification, and a multinomial classification in MNIST.

An Improved Particle Swarm Optimization Method for Nonlinear Optimization

Nonlinear optimization is becoming more challenging in information sciences and various industrial applications, but nonlinear problems solved by the classical particle swarm‐based methods are usually characterized by low efficiency, accuracy, and convergence speed in specific issues. To solve these problems and enhance the nonlinear optimization performance, an improved metaheuristic particle swarm optimization (PSO) model is proposed here. First, the optimization principles and model of the new method are introduced, and algorithms of the improved PSO are presented by updating the displacement and velocity of the moving particle according to Euler–Maruyama (EM) principle rather than traditional standard normal distribution. Then, the influence of the model parameters, input dimensions, and different nonlinear problems on the PSO optimization characterizations are studied by Pareto set solving and optimization performance comparison. The analysis regarding diverse nonlinear problems and optimization methods manifests that the improved method is capable of solving various nonlinear problems especially for multiobjective models, while the robustness and reliability can always keep consistent regardless of the change of model parameters. Finally, the performance evaluation is exhibited by the case study of nonlinear parameter optimization, 3 groups of CEC benchmark problems, and rank‐sum test for 6 comparable optimization algorithms, which all verify its effectiveness and reliability, as well as the significance and great application promise. The results show that the new proposed PSO method has the fastest convergence speed and least iteration numbers in searching for the global best solution of 9 nonlinear problems among 8 different optimization models indicated by the p values smaller than 0.05. Additionally, the main conclusions showing the calculation efficiency, stability, robustness, and great application promise of the proposed method are summarized, and future work is discussed.

Detailed analysis of a deep learning energy forecast model considering different input units and dimensions

The use of smart buildings technologies in residential houses, commonly referred to as smart homes, can benefit the operation and management of the buildings as well and its users. The application of energy related models in smart buildings can improve building sustainability and efficiency, but, usually, requires the existence of energy forecasting models capable of providing accurate and precise information. In this paper, it was studied how different input configurations of an energy forecasting model impact its performance by considering multiple error metrics. The conducted study allowed a detailed analysis of each input configuration to see which one is better suitable to be integrated with the smart grid management platform developed by the authors, which is also presented in this paper. The results demonstrate that the inputs’ units and dimensions have impact on the forecast error, indicating that accumulative Wh values are potentially better for the energy forecast model considered.

An effective U-net model for diagnosing Covid-19 infection

Coronavirus disease 2019 (COVID-19) has become a pandemic all over the world and has spread rapidly. To distinguish between infected and non-infected areas, there is a critical need for segmentation methods that can identify infected areas from Chest Computed Tomography (CT) scans. In recent years, deep learning has become the most widely used approach for medical image segmentation, including the identification of infected areas in Chest CT scans. We propose an encoder-decoder based on the U-NET architecture for segmenting the MedSeg dataset, which contains lung CT scans. To study the effect of input dimensions on the model's output results, we gave CT images with dimensions of 224 × 224, 256 × 256, and 512 × 512 as inputs to the model. The results showed that 224 × 224 achieved higher results compared to 256 × 256 and 512 × 512, with a dicecoef of 81.36, accuracy of 87.65, sensitivity of 84.71, and specificity of 88.35. Additionally, the 224 × 224 input based on the proposed model achieved the highest number of correct answers compared to previous U-net methods. The proposed model can be applied as an effective screening tool to help primary service staff better refer suspected patients to specialists.

A Study on Learning Strategies and Styles Segmentation of Learner Groups in Online Education Platforms

Abstract Online education is becoming a popular form of education day by day. In order to provide corresponding learning programs according to the types of learning styles and learning strategies of different learners, this paper proposes a learning style recognition algorithm based on the Felder-Silverman model. The Bayesian discriminant analysis algorithm is used to calculate the probability estimation of samples belonging to one category, and then the clustering center is calculated by k-means clustering analysis so as to complete the classification of learning styles. According to the results of learning style identification, the perceptual learning group has the highest level of cooperative communication strategies in the information perception style dimension, with a mean score of 3.575, the linguistic group is the least adept at all strategies in the information input dimension, with a mean score of only 3.55, the sequential group excels in a variety of strategies for information comprehension, and the active group is the most adept at performing information processing. The personalized learning experiment was conducted according to the characteristics of different style groups. It was found that students who received personalized teaching based on their learning styles improved their performance more significantly, with their average scores in each subject increasing from 73.174 to 82.247 in the pre-test. Their scores in the post-test were 5.729 points ahead of those of the control class. The application effects are significant when learning strategies and styles are classified, and teaching programs are improved accordingly.

Accurate Multi-segment Probability Density Estimation Through Moment Matching

The impact of process variations continues to grow as transistor feature size shrinks. Such variations in transistor parameters lead to variations and unpredictability in circuit output and may ultimately cause them to violate specifications leading to circuit failure. In fact, timely failures in critical circuits may lead to catastrophic failures in the entire chip. As such, statistical modeling of circuit behavior is becoming increasingly important. However, existing statistical circuit simulation approaches fail to accurately and efficiently analyze the high sigma behavior of probabilistic circuit output. To this end, we propose PDM (Piecewise Distribution Model) - a piecewise distribution modeling approach via moment matching using maximum entropy to model the high sigma behavior of analog/mixed-signal (AMS) circuit probability distributions. PDM is independent of the number of input dimensions and matches region specific probabilistic moments which allows for significantly greater accuracy compared to other moment matching approaches. PDM also utilizes Spearman’s rank correlation coefficient to select the optimal approximation for the tail of the distribution. Experiments on a known mathematical distribution and various circuits obtain accurate results up to 4.8 sigma with 2-3 orders of speedup relative to Monte Carlo. PDM also demonstrates better accuracy while compared against other state-of-the-art statistical modeling approaches, such as maximum entropy, importance sampling, and subset simulation.

Hamlet: Hierarchical Multi-Objective Reinforcement Learning Method for Charging Power Control of Electric Vehicles With Dynamic Quantity

Electric vehicle (EV) is emerging as an effective choice to reduce carbon emissions in the modern transportation system. However, the large access of EVs brings tremendous passive influence on the stability and performance of power grid operation. Thus, it is valuable to investigate the charging power management method to improve charging experience of EV drivers while maintaining normal bus voltage of the power system. This paper builds the charging power control problem of multiple EVs as a hierarchical Markov decision process (MDP) model and propose a hierarchical multi-objective reinforcement learning method (Hamlet) to obtain the real-time charging power control decisions. Specifically, by treating EVs with the same remaining charging time as the same agent, this paper addresses the dynamic state and action spaces issue and significantly reduce the input dimension of both state and action spaces. Besides, this paper quantifies drivers’ anxiety about the battery volume and distributes the comfort reward over the entire charging interval, which overcomes the delayed comfort reward issue. In order to alleviate the influence of multiple objectives on training stability, this paper develops a multi-objective learning method to dynamically adjust the optimization direction. Simulation results on IEEE 33 bus and 69 bus test feeders prove the validity of the proposed method in mitigating voltage fluctuation, minimizing bill payment and maximizing battery comfort.

A Parallel-Cascaded Ensemble of Machine Learning Models for Crop Type Classification in Google Earth Engine Using Multi-Temporal Sentinel-1/2 and Landsat-8/9 Remote Sensing Data

The accurate mapping of crop types is crucial for ensuring food security. Remote Sensing (RS) satellite data have emerged as a promising tool in this field, offering broad spatial coverage and high temporal frequency. However, there is still a growing need for accurate crop type classification methods using RS data due to the high intra- and inter-class variability of crops. In this vein, the current study proposed a novel Parallel-Cascaded ensemble structure (Pa-PCA-Ca) with seven target classes in Google Earth Engine (GEE). The Pa section consisted of five parallel branches, each generating Probability Maps (PMs) for different target classes using multi-temporal Sentinel-1/2 and Landsat-8/9 satellite images, along with Machine Learning (ML) models. The PMs exhibited high correlation within each target class, necessitating the use of the most relevant information to reduce the input dimensionality in the Ca part. Thereby, Principal Component Analysis (PCA) was employed to extract the top uncorrelated components. These components were then utilized in the Ca structure, and the final classification was performed using another ML model referred to as the Meta-model. The Pa-PCA-Ca model was evaluated using in-situ data collected from extensive field surveys in the northwest part of Iran. The results demonstrated the superior performance of the proposed structure, achieving an Overall Accuracy (OA) of 96.25% and a Kappa coefficient of 0.955. The incorporation of PCA led to an OA improvement of over 6%. Furthermore, the proposed model significantly outperformed conventional classification approaches, which simply stack RS data sources and feed them to a single ML model, resulting in a 10% increase in OA.

Evaluation of the Local Content Learning Program for Local Language in Jayapura District Using the CIPP Evaluation Model

Local governments in Indonesia through schools are starting to be moved to teach local languages as local content lessons to protect local local languages. Jayapura district also implements local content of the local language. This study aims to evaluate the Local Language Learning Program in Jayapura District. This research was evaluative research with CIPP model and using a qualitative approach. The data analysis technique used was Miles and Huberman. The formulation of the problem in this study was that the Regional Language Local Content Learning program in Jayapura Regency has been running since 2022 and, in fact, faces various obstacles; for this reason, it is necessary to conduct a thorough evaluation research on this program. Data collection techniques in this study were carried out by means of interviews, documentation studies, and observations. The results obtained in this study include in the context dimension, it is found that the local language local content program in Jayapura Regency was implemented quite well. In the input dimension, this program is also in the sufficient category. The process dimension in this study is also in the sufficient category. Then the product evaluation, has less results.

Improved tropical cyclone wind speed estimation for microwave altimeter using machine learning

Satellite altimeters can provide excellent global wind speed at 10 m above the sea surface (U10), however, the U10 becomes inaccurate and difficult to measure in tropical cyclone conditions. The violent wind, rough waves and torrential rain manifest an exceptionally complex ocean-atmospheric environment for wind estimation. Although the backscatter signal is measured equally well in normal condition, the interpretation is not straightforward in tropical cyclones that requires complex associations with ocean-atmospheric geophysical variables. Typical U10 regression model developed in normal atmospheric conditions would inevitably reduce the estimation quality and encounter high modelling uncertainties from high dimensional input data that provide ill-posed solutions in extreme U10 estimation. However, other secondary parameters simultaneously measured by the altimeter have properties that convey additional atmospheric information to enhance U10 estimation accuracy in storm condition. Therefore, the present study proposes machine learning (ML) approaches based on artificial neural network (ANN), support vector machine (SVM), and Gaussian process regression (GPR) to integrate the multi-dimensional parameters and provide accurate U10 estimates for different parameter combination. Results suggest that the GPR method, considering atmospheric and sea surface related parameters, can provide the highest accuracy of U10 up to 35 ms−1 with quality perseverance against rain contamination. This study highlights the quality improvement of altimeter U10 estimation and the assessment method of ML in dealing with complex ocean-atmospheric interactions.

Feature selection of ground motion intensity measures for data‐driven surrogate modeling of structures

AbstractIn the probabilistic seismic performance assessment of structures, intensity measures (IMs) represent seismic characteristics and variations. Traditional fragility analysis method based on the assumption of linear regression requires selecting an optimal IM as input variable. By introducing machine learning (ML) techniques, nonparametric fragility analysis theoretically allows for considering all potential IMs as inputs. Nevertheless, to reduce input dimensionality and improve training efficiency, the feature selection of IMs remains imperative. This paper proposes a method to select optimal ground motion IMs for data‐driven surrogate modeling of structures. Specifically, the elastic net algorithm is employed to select the optimal multiple IMs based on the coefficient of determination and regression coefficient, differing from the efficiency and practicality emphasized in the traditional method. Using the optimal multiple IMs as input variables, several ML techniques are employed to construct surrogate models for seismic damage assessment of structures, thereby developing fragility functions, that is, the conditional probability of exceeding a damaged state given seismic intensity. A 3‐span, 6‐storey, reinforced concrete frame is utilized to illustrate the proposed methodology. The predictive performance of all ML models with the optimal multiple IMs outperforms that of the models with the commonly used IM (e.g., peak ground acceleration, PGA) as sole input and all candidate IMs as inputs. Additionally, the surrogate models with the optimal multiple IMs enable a more comprehensive seismic fragility modeling of structures under two or more IMs simultaneously, such as the fragility surface under spectral acceleration at 1.0s (Sa‐1.0s) and velocity spectrum intensity (VSI).

Predicting indoor 3D airflow distribution using artificial neural networks with two different architectures

To meet the thermal comfort requirements of room occupants, a fast and accurate method for predicting indoor high-resolution 3D airflow distribution is necessary, which can be combined with heating, ventilation, and air conditioning (HVAC) systems to adjust the indoor environment. Artificial neural networks (ANN) can establish complex mappings between variables with nonlinear relationships. The aim of this study was to verify the feasibility of an ANN for the fast and accurate prediction of indoor 3D airflow distribution. Considering two prediction strategies, ANN A and B were constructed. The inputs of both ANNs were the air supply velocity, air supply direction, and section location. ANN A output the airflow distribution of the entire room directly, whereas ANN B output the airflow distribution of the characteristic sections first and then the airflow distribution of other sections according to the contribution of the characteristic sections. The two ANNs’ performances on training and testing datasets were compared. Both ANNs performed well in the training and interpolation cases. However, their performances decreased rapidly in the extrapolation cases. In the case of a large gap between the input and output dimensions of the ANNs, the two-step prediction strategy improved the prediction accuracy, accelerated convergence, and alleviated data imbalance problems. The prediction time of ANN A and B was 0.89 s and 12.38 s, respectively. The ANN can predict indoor high-resolution 3D airflow distribution quickly and accurately, and can add functions that include identifying internal heat sources and air supply inlet position changes. ANN combined with HVAC could allow real-time control to regulate indoor airflow.

On parametric optimal execution and machine learning surrogates

We investigate optimal order execution problems in discrete time with instantaneous price impact and stochastic resilience. First, in the setting of linear transient price impact we derive a closed-form recursion for the optimal strategy, extending the deterministic results from Obizhaeva and Wang [Optimal trading strategy and supply/demand dynamics. J. Financial Mark., 2013, 16(1), 1–32]. Second, we develop a numerical algorithm based on dynamic programming and deep learning for the case of nonlinear transient price impact as proposed by Bouchaud et al. [Fluctuations and response in financial markets: The subtle nature of ‘random’ price changes. Quant. Finance, 2004, 4(2), 176–190]. Specifically, we utilize an actor-critic framework that constructs two neural-network (NN) surrogates for the value function and the feedback control. The flexible scalability of NN functional approximators enables parametric learning, i.e. incorporating several model or market parameters as part of the input space. Precise calibration of price impact, resilience, etc., is known to be extremely challenging and hence it is critical to understand the sensitivity of the execution policy to these parameters. Our NN learner organically scales across multiple input dimensions and is shown to accurately approximate optimal strategies across a wide range of parameter configurations. We provide a fully reproducible Jupyter Notebook with our NN implementation, which is of independent pedagogical interest, demonstrating the ease of use of NN surrogates in (parametric) stochastic control problems.

Disparities in innovation capital of cities in Poland

ObjectivesThe aim of the paper is to measure innovation capital of Polish cities and verify whether its dispersion changed.Material and methodsData was retrieved from the Local Data Bank provided by Statistics Poland. The sample consists of 18 Polish cities. The research period covers years 2014-2021. The TOPSIS method was applied to build the innovation capital index of cities and sigma convergence of innovation capital was verified.ResultsThe findings indicate relatively high cross-sectional disparities in the vast majority of the investigated indicators of cities’ innovation capital. The largest disparities were identified for the number of higher education graduates and the new registered enterprises in high-tech services sector. The constructed composite indicator - innovation capital index took the highest average values in the analysed period for Wrocław, Poznań, Kraków, Warszawa, and Lublin, whereas Gorzów Wielkopolski, Zielona Góra, Bydgoszcz, and Białystok achieved the lowest values. The sigma convergence analysis revealed that the differentiation of innovation capital in Polish cities remained at a similar level and there was no convergence in this area.ConclusionsInnovation capital plays a key role in the development of urban economy. Given the complexity of innovation capital, its measurement should comprise a set of indicators reflecting both output and input dimensions. Identification of spatial patterns of innovation capital distribution should enable to adjust smart specialization strategy to territorial characteristics of a given city.