Input Time Series Research Articles

DRN-DSA: A Hybrid Deep Learning Network Model for Precipitation Nowcasting using Time Series Data

Precipitation nowcasting involves short-term weather forecasting, predicting rain or snow within the next two hours. By analyzing current atmospheric conditions, it aids meteorologists in identifying weather patterns and preparing for severe events such as flooding. These nowcasts are typically displayed on geographical maps by weather services. However, the rapidly changing climate conditions make precipitation nowcasting a formidable challenge, as accurate short-term forecasts are hindered by immediate weather fluctuations. Traditional nowcasting methods, like numerical models and radar extrapolation, have limitations in delivering highly detailed and timely precipitation nowcasts. To overcome this issue, an effective solution is framed for precipitation nowcasting using a hybrid network approach named Deep Residual Network-Deep Stacked Autoencoder (DRN-DSA). Initially, the input time series data is acquired from the dataset. Thereafter, the effective technical indicators are extracted at the feature extraction stage. Later on, precipitation-type nowcasting is carried out using the proposed hybrid DRN-DSA, which is developed by incorporating a Deep Stacked Autoencoder (DSA) and Deep Residual Network (DRN). Finally, Weather nowcasting is carried out using the same proposed hybrid DSA-DRN. Moreover, when compared to other traditional models, the proposed DRN-DSA has gained superior results with a Relative Absolute Error (RAE) of 0.295, Root Mean Square Error (RMSE) of 0.154, low Mean Square Error (MSE) of 0.0236, Mean Absolute Percentage Error (MAPE) of 0.295, and False Acceptance Rate (FAR) of 0.0118.

Wrapping input–output multipliers in confidence intervals

Input–output multipliers are typically calculated as point estimates of the Leontief quantity model. From the previous literature, they can also be estimated (and their confidence intervals) directly from establishments’/industries’ inputs and outputs data by running an appropriate econometric regression. However, previous contributions relied on analyses carried out for 1 single year, one country/region and a reduced number of degrees of freedom. This paper tests the robustness of the previous literature by using instead a time series of large-scale inter-country supply, use and input–output database with more than one country and more than 1 year, as well as panel data econometrics.

MODELING COOKING OIL PRODUCTION WITH CRUDE PALM OIL (CPO) PRICE USING TRANSFER FUNCTION

Indonesia is the largest palm oil producer in the world. However, Indonesia is currently experiencing a significant scarcity of cooking oil for the basic needs of the community. One of the causes of this problem is the high price of crude palm oil (CPO). The price of CPO affects the production of cooking oil produced by the cooking oil industry. CPO prices are released in the form of daily prices in the market which form a time series pattern that affects the production of cooking oil which is also produced daily in the industry. To forecast cooking oil production from CPO prices can be done with a transfer function model. The stages of analysis are forming an ARIMA model, processing input time series whitening, and output time series whitening. Checking white noise by plotting ACF αt, ACF βt. Make a CCF graph and determine the order (r,s,b). Estimate the first parameter of the disturbance series (nt) and identify the disturbance series (nt) in the ARIMA model. Next, perform a diagnostic check of the transfer function model by checking the cross-correlation between Xt & Yt and checking the autocorrelation of Xt and Yt for model feasibility testing. Then make an ACF graph of the transfer function model's residuals to test the model's suitability. The final stage is to select the best model based on the smallest AIC value, ensuring the accuracy of the selected model using MAPE value, and making predictions. As a result, the model for CPO price is an input series variable, and cooking oil production is an output series variable. However, the MAPE result was 100%, indicating that the model is not very accurate for this data. Nevertheless, by ignoring the result and continuing with the forecasting, this model shows that the forecasted values for cooking oil based on CPO price have increased from the form of the model obtained for the time data from 2013 to 2022.

A Bi‐level stacked LSTM‐DNN‐based decoder network for AGC dispatch under regulation market framework in presence of VPP and EV aggregators

AbstractThe consideration of mileage settlement in the frequency regulation market has encouraged fast‐acting units, such as converter‐interfaced generators (CIG) and electric vehicle stations, to actively participate in load‐generation balancing through automatic generation control (AGC). Conventional frequency regulation faces challenges in coping with the growing variability of CIGs and also lacks effective incentives for rapid‐responding units. In this context, a bi‐level AGC dispatch approach based on a stacked long short‐term memory (LSTM)‐deep neural network (DNN)‐based decoder framework is proposed for a power system comprising diverse CIGs forming a virtual power plant and electric vehicle aggregators. The proposed decoder network is comprised of stacked LSTM and DNN, wherein the cascaded LSTM layers are introduced to accurately capture temporal information from time series input. The inclusion of a dropout mechanism further enhances the model’s generalisability in unforeseen environments. The proposed dispatch framework uses mileage‐based compensation criteria to optimally allocate instructions among various participating units with differing regulation characteristics. The performance of the proposed method is analysed by considering packet loss, delay, unexpected generation failure, and denial of service attacks. The evaluation of the proposed approach reveals its superior performance compared to proportionality, particle swarm optimisation, decision tree, and DNN methods.

Relationship between the Human Development Index and public social spending: European experience for Ukraine

This study focuses on the relationship between the Human Development Index (HDI) and public social expenditures, analyzing socio-economic models using the examples of selected European countries and Ukraine. The study used the values of the HDI, GDP, and indicators of public expenditures for social purposes, namely, healthcare, education, leisure, culture and religion, and social protection for the period from 2010 to 2021. The analysis targeted 13 European countries using data sets from Eurostat, the Office for National Statistics of the United Kingdom, the State Statistics Service of Ukraine, and the Ministry of Finance of Ukraine. The input time series were checked for lagged values using the STATISTICA software.Empirical evidence suggests a relationship between HDI and public social spending. An increase in the share of public social expenditures in GDP leads to an increase in HDI and vice versa. European countries with a social-democratic model of development have the highest level of centralization of public expenditure in GDP (34.72%) and the highest HDI (0.930), while countries belonging to the Southern European model have the lowest share of socially oriented public expenditure (30.41%) and the lowest HDI (0.873). In addition, there is a time lag between the investment of public funds in healthcare, education, leisure, culture and religion, and social protection and their impact on HDI changes. Thus, ensuring a high level of HDI is achieved, among other things, through state financial support for the relevant components of the social sphere and social protection. AcknowledgmentThe study is funded by the EU NextGenerationEU through the Recovery and Resilience Plan for Slovakia under project No. 09I03-03-V01-00130.

Cryptocurrencies Price Prediction Using Deep Learning Models (Gated Recurrent Unit And Recurrent Neural Network))

Background: With their volatile prices, cryptocurrencies have become valuable assets in the financial market. Predicting cryptocurrency prices accurately is essential for making well-informed investment decisions. Time series prediction models, like Gated Recurrent Unit (GRU) and Recurrent Neural Networks (RNN), are popular tools for financial data forecasting because they can capture sequential dependencies in data. Aim: This study aims to predict the average monthly closing prices of five major cryptocurrencies—Bitcoin (BTC), Ethereum (ETH), Binance Coin (BNB), Litecoin (LTC), and Ripple (XRP)—using GRU and RNN models and evaluate their performance in forecasting these prices. Method: Time series input sequences were produced and historical price data for the chosen cryptocurrencies were preprocessed using Min-Max Scaling. This data was divided into training and test sets, and it was used to train both the GRU and RNN models. The Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) were used to assess the performance of the model. Results: For the majority of cryptocurrencies, the RNN model exhibited better predicted accuracy and consistently outperformed the GRU model. For instance, the RMSE for Ripple was 0.06 for the RNN model and 0.09 for GRU. In a similar vein, the RNN model outperformed the GRU model with a MAPE of 12.97% for Ethereum. These results imply that RNN models are more suitable for financial forecasting in this sector, as they yield more accurate predictions for cryptocurrency values.

Short-term load forecasting by GRU neural network and DDPG algorithm for adaptive optimization of hyperparameters

Short-term load forecasting (STLF) is critical to optimizing power system operation. Deep learning (DL) methods can provide extremely high accuracy for STLF. However, most models in existing research lack adaptive optimization capabilities in the prediction process and suffer from performance degradation. To resolve the above difficulties, we propose a hybrid model (DDPG-GRU) based on gated recurrent units and deep deterministic policy gradients for STLF. First, the GRU network has the advantage of processing multiple time series inputs and can simultaneously consider multi-dimensional load characteristics, thereby making the model more efficient. Since the GRU model structure is relatively complex, choosing a good set of hyperparameters is very difficult. Therefore, the purpose of using DDPG is to optimize the hyperparameters of the GRU model adaptively. The proposed model is a combination of DL methods and reinforcement learning. In order to prove the superiority of the proposed model, it is applied to the load data of Area 1 in China to perform single-step and multi-step load forecasting, respectively. The results show that DDPG-GRU has a better fitting effect than the baseline method. Taking the multi-step prediction results as an example, compared with the classic GRU network, the MAPE, MAE, and RMSE of the proposed model are reduced by 22.75 %, 14.44 %, and 14.02 %, respectively, while the R2 coefficient is increased by 13.23 %. At the same time, we use the China Area 2 data set to verify the universality of the proposed model. Furthermore, we compared the proposed method with state-of-the-art methods and achieved better accuracy.

Understanding the Dynamics of Ocean Wave-Current Interactions Through Multivariate Multi-Step Time Series Forecasting

ABSTRACT Understanding ocean wave-current interactions’ complex dynamics is crucial for coastal engineering, marine operations, and climate research applications. This study introduces a pioneering data-driven approach by employing advanced deep learning techniques, specifically Long Short-Term Memory (LSTM), and Bidirectional LSTM (BiLSTM) models, to forecast both wave and current parameters at varying depths. The models are designed to capture the complex temporal relationships inherent in ocean dynamics, considering wave speed and direction, current speed, and direction as multivariate time series inputs. Two comprehensive experiments are conducted, one utilizing historical values of all parameters and another focusing on using wave parameters to forecast current parameters. Model performance is rigorously evaluated across forecast horizons of 5, 12, and 24 hours ahead using metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). BiLSTM emerges as the superior model, demonstrating lower errors, particularly at higher depths, while nearshore predictions reveal challenges in shallower waters. Furthermore, the methodology incorporates hyperparameter optimization and cross-validation techniques to enhance the model’s robustness. Ultimately, this work represents a transformative leap toward smarter oceans, emphasizing the fusion of fluid dynamics and bathymetry to advance our understanding of coupled wave-current dynamics. The results showcase high accuracy and reliability across various forecast horizons and depths, signifying the method’s potential applications in oceanography, hydrodynamics, coastal engineering, and ocean renewable energy.

The power of mineral: Shock of the global supply chain from resource nationalism

State intervention has long affected the stability of the global mineral and raw material markets. Global energy transition has continued to boost demand for copper, lithium, cobalt, and nickel. It has exacerbated the global prevalence of resource nationalism. Research on resource nationalism has focused on the interaction of economic, political, cultural, and social elements within countries. The lack of understanding of the global economic effects of resource nationalism limits our flexibility in responding to upheavals in the global resource order. We combine time series counterfactual analysis and input–output modeling to advance our understanding of the worldwide supply chain effects of resource nationalism. We analyze the impacts of resource political behavior on global supply chains based on several representative episodes of resource nationalism. While major industrialized countries suffer more absolute welfare losses, formerly developed countries in the Global South face higher price effects. Moreover, nationalist resource shocks are not the only cause of welfare losses in the metals supply chain sector. Still, they are also harming various industries and groups, from food and energy to services. Such differences in price and welfare effects depend on the type of mineral and the resource power of the event-initiating country, on the one hand, and are influenced by the global industrial division of labor, on the other. More losses in developing countries come from the agricultural sector and primary manufacturing than from the severity of the losses suffered by the manufacturing and service sectors in industrial and rich countries. This is a reminder to pay more attention to the unequal distribution of economic pressures across countries and sectors in the global energy transition.

DCEnt‐PredictiveNet: A novel explainable hybrid model for time series forecasting

This work presents a novel hybrid framework called DCEnt-PredictiveNet (deep convolutional neural network (DCNN) + entropy + support vector regressor (SVR)) that concatenate both deep and handcrafted features for time series data analysis and forecasting. From the discrete wavelet transform coefficients of input time series data, computed four different handcrafted entropy features, which were then concatenated with deep features extracted using a modified DCNN. The concatenated deep and handcrafted feature vector was then fed to a SVR for prediction. The DCEnt-PredictiveNet framework was trained and tested on three time series datasets of real-world COVID-19, stock price and traffic information, and achieved mean absolute percentage errors of 0.03 %, 1.53 % and 11.41 % for daily cumulative COVID-19 positive cases, closing stock price, and hourly traffic (vehicle numbers) at one junction predictions, respectively. In addition, we incorporated local interpretable model-agnostic explanations and Shapley additive explanations methods into DCEnt-PredictiveNet to enable visualization of significant features that contributed to the model’s decision-making, thereby enhancing its explainability. Our DCEnt-PredictiveNet model yielded promising and interpretable forecasting results, which can facilitate advance resource planning in hospitals for incoming COVID-19 patients, stock market investment planning, and efficient traffic control management.

Hybrid Long Short-Term Memory Fused Convolution Neural Network for Weather Forecasting

Weather forecasting is an effort by meteorologists to predict whether states at a few prospective times and the conditions of weather that might be estimated. With new modern technology, present weather forecasting methods are highly precise. To achieve high accuracy, the methods developed for weather forecasting were much more complicated owing to many factors. Here, the usage of time series data for weather forecasting is done by the devised Long-Short Term Memory fused Convolutional neural network (LSTMFCNN). At first, the acquisition of input time series data from the specific dataset is done. In The feature extraction technical features are done by considering the input time series data. Then, the feature extraction is done utilizing the Rider Optimization Algorithm-Based Neural Network (RideNN) with the Soergel metric. RideNN is the integration of the Rider Optimization Algorithm (ROA) with the Neural Network (NN) classifier. Thus, the feature fusion step reduces the complexity and improves the accuracy. Thereafter, the oversampling technique is utilized for the data augmentation (DA) process. Finally, weather forecasting is done utilizing the newly designed LSTMFCNN and is obtained by the integration of Convolutional Neural Network (CNN) and Deep Long-Short Term Memory (DLSTM).

Charging Station Power System Data Prediction Model Based on Deep Learning

The increase growth and adoption of electric vehicles (EVs) are playing a crucial role in advanced transportation system, helping to minimizing the emissions of harmful greenhouse gases and enhancing environmental sustainability. The need for EVs to be charged immediately has gained significant attention due to the rise in EVs sales over the last years. As a result of this, need for electric car charging is important to reducing the impact of electric network and providing minimum charging fares. In order to calculate the demand for charging EVs, a novel Deep Learning (DL)-based Long-Short Term Memory (LSTM) recurrent neural network predictor model is attempted to be developed in this research study. The Modified Aquilla Optimizer Algorithm (MAOA) is used to optimizing the parameter of the new Deep LSTM (DLSTM) neural predictor models and Independent Component Analysis (ICA) is utilized to solve the input time series data while conserving its properties. In this research, a novel ICA—AOA—DLSTM neural predictor model was developed to addressed the challenges of vanishing and exploding gradient in basic recurrent neural learnings. The predictor model was tested on the EV charging dataset from Georgia Tech, Atlanta, USA, indicating its performance. During simulation, the predictor achieved a prediction accuracies of 96.24% with a mean absolute errors of 0.1083 and a RMSE errors of 3.0629 × 10^-5, outperforming previous techniques. Additionally, the mean absolute error was found to be 0.2083, with a mean square error of 3.25516 × 10^-10. These results highlight the effectiveness of the novel deep learning LSTM neural predictor for this dataset compared to existing techniques.

A novel method for aero-engine time-series forecasting based on multi-resolution transformer

Time-series forecasting is widely studied in the data-driven component-level modeling, fault diagnosis, and performance prediction of aero-engines. The operational data of aero-engines have the characteristics of complexity and nonlinearity. This paper proposes a multi-resolution transformer (MRT) model for the non-steady state process of aero-engines. This transformer-based model can learn temporal patterns of different scales by performing multi-resolution down-sampling and patch-based tokeniczation on the input time-series. On this basis, a two-stage multi-resolution transformer (TSMRT) framework is proposed for the entire processes time-series modeling of aero-engines. The TSMRT framework employs Augmented Dickey-Fuller (ADF) test to classify aero-engine operational data into steady state processes and non-steady state processes, and models the two processes separately using time–frequency feature neural network (TFNN) model and MRT model. The time-series forecasting performance of the MRT model is validated on three publicly available benchmark datasets and an ablation study is conducted on turbofan engine test bench datasets. The results indicate that the TSMRT framework demonstrates superior forecasting performance across steady state, non-steady state, and entire processes. This framework stands out as a novel method for component-level modeling of aero-engines.

U-HRNet+: a real-time precipitation estimation method based on the fusion of temporal and spatial domain features

ABSTRACT The precipitation estimation with high spatial and temporal resolution over Tibet is of great significance for monitoring the stability of the plateau ecosystem and early warning of natural disasters. In response to the problem of the unsatisfactory accuracy in traditional real-time precipitation inversion methods and quantitative accuracy estimation products in this region. This paper adopts deep-learning methods to explore the corresponding relationship between precipitation and multi-band observation data from geostationary satellite imagery. The U-High Resolution Network (U-HRNet) as backbone is fully exploited to generate multi-scale features and fused using attention gates to extract high-resolution spatial representations representing precipitation clouds, thereby achieving precipitation estimation of target clouds at different scales. In addition, in response to the characteristics of high precipitation intensity and drastic cloud cluster changes in summer convective weather within a short period of time, a multi-head attention mechanism is introduced to extract temporal information from time series inputs and fuse it with spatial features to improve the accuracy of real-time precipitation estimation. Compared with the operational precipitation products and baseline deep-learning models, the classification metrics such as Probability of detection (POD), False Alarm Ratio (FAR) and Critical success index (CSI) show that the proposed model (U-HRNet+) has better performance in test set and has high generalization ability for estimating short-term heavy rainfall over Tibet and its surrounding areas in summer. Therefore, the proposed model has the potential to serve as a more accurate and reliable satellite-based precipitation estimation product.

Leveraging advances in data-driven deep learning methods for hybrid epidemic modeling

Mathematical modeling of epidemic dynamics is crucial to understand its underlying mechanisms, quantify important parameters, and make predictions that facilitate more informed decision-making. There are three major types of models: mechanistic models including the SEIR-type paradigm, alternative data-driven (DD) approaches, and hybrid models that combine mechanistic models with DD approaches. In this paper, we summarize our work in the COVID-19 Scenario Modeling Hub (SMH) for more than 12 rounds since early 2021 for informed decision support. We emphasize the importance of deep learning techniques for epidemic modeling via a flexible DD framework that substantially complements the mechanistic paradigm to evaluate various future epidemic scenarios. We start with a traditional curve-fitting approach to model cumulative COVID-19 based on the underlying SEIR-type mechanisms. Hospitalizations and deaths are modeled as binomial processes of cases and hospitalization, respectively. We further formulate two types of deep learning models based on multivariate long short term memory (LSTM) to address the challenges of more traditional DD models. The first LSTM is structurally similar to the curve fitting approach and assumes that hospitalizations and deaths are binomial processes of cases. Instead of using a predefined exponential curve, LSTM relies on the underlying data to identify the most appropriate functions, and is capable of capturing both long-term and short-term epidemic behaviors. We then relax the assumption of dependent inputs among cases, hospitalizations, and death. Another type of LSTM that handles all input time series as parallel signals, the independent multivariate LSTM, is developed. Independent multivariate LSTM can incorporate a wide range of data sources beyond traditional case-based epidemiological surveillance. The DD framework unleashes its potential in big data era with previously neglected heterogeneous surveillance data sources, such as syndromic, environment, genomic, serologic, infoveillance, and mobility data. DD approaches, especially LSTM, complement and integrate with the mechanistic modeling paradigm, provide a feasible alternative approach to model today’s complex socio-epidemiological systems, and further leverage our ability to explore different scenarios for more informed decision-making during health emergencies.

An adaptive and interpretable SOH estimation method for lithium-ion batteries based-on relaxation voltage cross-scale features and multi-LSTM-RFR2

Accurate estimation of state of health (SOH) is crucial for ensuring reliable and safe operations of lithium-ion batteries across diverse engineering scenarios. However, current SOH estimation methods face challenges of reliance on specific working conditions, inadequate quality of degradation features, and limited precision and interpretability of estimation models. To address these issues, this paper presents an adaptive and interpretable SOH estimation method based on cross-scale features and a framework with multiple long short-term memory processes and two-level random forest regression (Multi-LSTM-RFR2). First, relaxation voltage, adapting to multiple charging-discharging conditions, is utilized as a source for degradation feature construction. In this aspect, multiple value features, variation features, statistic features, and parameter features are created from four scales: point data, local data, global data, and extended data. Then, the Multi-LSTM-RFR2 interpretable framework employs multi-LSTM processes to handle time-series inputs from multiple features, incorporating a two-level RFR mechanism to fuse and optimize pre-estimated SOH results. Finally, the importance of features is studied from both single and category feature perspectives for interpretability analysis. Degradation experiments with 24 commercial 18,650 battery cells under multiple charging-discharging conditions are used to validate effectiveness of the proposed method. The average values of four evaluation indicators in terms of mean absolute error, mean absolute percentage error, root mean squared error, and R-squared are 1.07 %, 1.33 %, 1.28 %, and 0.9654, respectively.

Robust explainer recommendation for time series classification

Time series classification is a task which deals with temporal sequences, a prevalent data type common in domains such as human activity recognition, sports analytics and general sensing. In this area, interest in explanability has been growing as explanation is key to understand the data and the model better. Recently, a great variety of techniques (e.g., LIME, SHAP, CAM) have been proposed and adapted for time series to provide explanation in the form of saliency maps, where the importance of each data point in the time series is quantified with a numerical value. However, the saliency maps can and often disagree, so it is unclear which one to use. This paper provides a novel framework to quantitatively evaluate and rank explanation methods for time series classification. We show how to robustly evaluate the informativeness of a given explanation method (i.e., relevance for the classification task), and how to compare explanations side-by-side. The goal is to recommend the best explainer for a given time series classification dataset. We propose AMEE, a Model-Agnostic Explanation Evaluation framework, for recommending saliency-based explanations for time series classification. In this approach, data perturbation is added to the input time series guided by each explanation. Our results show that perturbing discriminative parts of the time series leads to significant changes in classification accuracy, which can be used to evaluate each explanation. To be robust to different types of perturbations and different types of classifiers, we aggregate the accuracy loss across perturbations and classifiers. This novel approach allows us to recommend the best explainer among a set of different explainers, including random and oracle explainers. We provide a quantitative and qualitative analysis for synthetic datasets, a variety of time-series datasets, as well as a real-world case study with known expert ground truth.

Support vector machine-based stock market prediction using long short-term memory and convolutional neural network with aquila circle inspired optimization

ABSTRACT Predicting the stock market is one of the significant chores and has a successful prediction of stock rates, and it helps in making correct decisions. The prediction of the stock market is the main challenge due to blaring, chaotic data as well as non-stationary data. In this research, the support vector machine (SVM) is devised for performing an effective stock market prediction. At first, the input time series data is considered and the pre-processing of data is done by employing a standard scalar. Then, the time intrinsic features are extracted and the suitable features are selected in the feature selection stage by eliminating other features using recursive feature elimination. Afterwards, the Long Short-Term Memory (LSTM) based prediction is done, wherein LSTM is trained to employ Aquila circle-inspired optimization (ACIO) that is newly introduced by merging Aquila optimizer (AO) with circle-inspired optimization algorithm (CIOA). On the other hand, delay-based matrix formation is conducted by considering input time series data. After that, convolutional neural network (CNN)-based prediction is performed, where CNN is tuned by the same ACIO. Finally, stock market prediction is executed utilizing SVM by fusing the predicted outputs attained from LSTM-based prediction and CNN-based prediction. Furthermore, the SVM attains better outcomes of minimum mean absolute percentage error; (MAPE) and normalized root-mean-square error (RMSE) with values about 0.378 and 0.294.

Mapping drivers of tropical forest loss with satellite image time series and machine learning

The rates of tropical deforestation remain high, resulting in carbon emissions, biodiversity loss, and impacts on local communities. To design effective policies to tackle this, it is necessary to know what the drivers behind deforestation are. Since drivers vary in space and time, producing accurate spatially explicit maps with regular temporal updates is essential. Drivers can be recognized from satellite imagery but the scale of tropical deforestation makes it unfeasible to do so manually. Machine learning opens up possibilities for automating and scaling up this process. In this study, we developed and trained a deep learning model to classify the drivers of any forest loss—including deforestation—from satellite image time series. Our model architecture allows understanding of how the input time series is used to make a prediction, showing the model learns different patterns for recognizing each driver and highlighting the need for temporal data. We used our model to classify over 588′000 sites to produce a map detailing the drivers behind tropical forest loss. The results confirm that the majority of it is driven by agriculture, but also show significant regional differences. Such data is a crucial source of information to enable targeting specific drivers locally and can be updated in the future using free satellite data.

Transmission Line Fault Classification Using Conformer Convolution-Augmented Transformer Model

Ensuring a consistently reliable power supply is paramount in power systems. Researchers are engaged in the pursuit of categorizing transmission line failures to design countermeasures for mitigating the associated financial losses. Our study employs a machine learning-based methodology, specifically the Conformer Convolution-Augmented Transformer model, to classify transmission line fault types. This model processes time series input data directly, eliminating the need for expert feature extraction. The training and validation datasets are generated through simulations conducted on a two-terminal transmission line, while testing is conducted on historical data consisting of 108 events that occurred in the Taiwan power system. Due to the limited availability of historical data, they are utilized solely for inference purposes. Our simulations are meticulously designed to encompass potential faults based on an analysis of historical data. A significant aspect of our investigation focuses on the impact of the sampling rate on input data, establishing that a rate of four samples per cycle is sufficient. This suggests that, for our specific classification tasks, relying on lower frequency data might be adequate, thereby challenging the conventional emphasis on high-frequency analysis. Eventually, our methodology achieves a validation accuracy of 100%, although the testing accuracy is lower at 88.88%. The discrepancy in testing accuracy can be attributed to the limited information and the small number of historical events, which pose challenges in bridging the gap between simulated data and real-world measurements. Furthermore, we benchmarked our method against the ELM model proposed in 2023, demonstrating significant improvements in testing accuracy.