Short-Term Aggregated Residential Load Forecasting using BiLSTM and CNN-BiLSTM

Sharia stocks are one of the investment instruments in the Islamic capital market. In the capital market, it is known that stock prices are very volatile. This makes investors need to carry out a strategy for making the right decision in investing, one of which can be done by predicting stock prices. In this study, predictions were made using historical data on the closing price of Islamic shares of PT. Telkom Indonesia Tbk with the Bidirectional Long Short Term Memory (BiLSTM) method. In building the best prediction model, it is necessary to choose the right parameters and one way to do this is to use the grid search algorithm. Based on the results of the test analysis, it was found that the smallest Mean Absolute Percentage Error (MAPE) value was found in the BiLSTM model in the distribution of data with a percentage of 90% training data and 10% testing data and parameter values obtained based on parameter tuning using grid search, including the number of neurons 25, 100 epochs, 4 batches, and 0.2 dropouts. The MAPE obtained in this study was 10.83% and based on the scale on the MAPE value criteria, this shows that the resulting prediction model is accurate. As for the test results from the comparisons made on the BiLSTM and LSTM models using grid search as a tuning parameter and models without using a grid search or it can be called a trial and error approach as a tuning parameter, it is found that the model with better predictive performance is found in BiLSTM using a grid search. compared to other models.

Consumer opinions on purchased products significantly influence the number of purchases on ecommerce platforms. Sentiment analysis of these opinions assists both consumers and companies in making informed decisions. Currently, consumers judiciously evaluate different aspects of a product to guarantee it meets their requirements. Consequently, performing sentiment analysis on various product aspects can aid consumers select the precise product and support companies in meeting on areas that prerequisite upgrading. This study recommends a systematic and expandable method to aspect-based sentiment analysis of user reviews for electrical devices like computers, mobile phones, and headphones. Through an amalgamation of advanced preprocessing, sentence-level segmentation by means of transformer-based models, and similarity-based aspect detection, the system is competent of extracting fine-grained sentiment for each product aspect. A hybrid sentiment classification strategy employing both BERT and VADER certifies robustness and accurateness in sentiment detection, while the structured transformation of data into a multi-label format permits for efficient model training. The proposed Bi-LSTM architecture with an integrated attention mechanism boosts sentiment prediction by concentrating on the utmost pertinent parts of each sentence, directing to better-quality performance across evaluation metrics, including accuracy, precision, recall, and F1-score. The exploration offers a comparative study of the performance of aspect-based sentiment analysis using a standard Bi-LSTM model versus a Bi-LSTM model with an attention mechanism, applied to analyse the reviews of three diverse products. The outcomes explains that the Bidirectional Long Short-Term Memory (Bi-LSTM) model with an attention mechanism outpaces the Bidirectional Long Short-Term Memory (Bi-LSTM) in terms of accuracy, precision, recall, and F1-score. The investigation validates the efficiency of merging NLP techniques and deep learning to deliver understandings from product reviews, presenting beneficial assistance to consumers by aiding to make informed decisions and focus on companies directing to improve their offerings based on targeted feedback.

Solar energy has become a promising renewable energy source, offering significant opportunities for photovoltaic (PV) systems. Accurate and reliable PV generation forecasts are crucial for efficient grid integration and optimized system planning. However, the complexity of environmental factors, including seasonal and daily patterns, as well as social behaviors and user habits, presents significant challenges. Traditional prediction models often struggle with capturing the complex nonlinear dynamics in multivariate time series, leading to low prediction accuracy. To address this issue, this paper proposes a new PV power prediction method that considers factors such as light, air pressure, wind direction, and social behavior, assigning different weights to them to accurately extract nonlinear feature relationships. The framework integrates long short-term memory (LSTM) and gated recurrent units (GRU) to capture local time features, while bidirectional LSTM (BiLSTM) and an attention mechanism extract global spatiotemporal relationships, effectively capturing key features related to historical output. This improves the accuracy of multi-step predictions. To verify the feasibility of the method for multivariate time series, we conducted experiments using PV power prediction as a scenario and compared the results with LSTM, CNN, BiLSTM, CNN-LSTM and GRU models. The experimental results show that the proposed method outperforms these models, with a mean absolute error (MAE) of 12.133, root mean square error (RMSE) of 14.234, mean absolute percentage error (MAPE) of 2.1%, and a coefficient of determination (R2) of 0.895. These results indicate the effectiveness and potential of the method in PV prediction tasks.

In cybersecurity, the rise of fileless malware poses a significant challenge to endpoint security. Traditional detection methods often fail against these sophisticated attacks, necessitating advanced techniques like deep learning models. This study highlights the limitations of Bi-Directional Long Short-Term Memory (BLSTM) models in dynamic malware analysis and proposes enhancements through Convolutional Long Short-Term Memory (ConvLSTM) architecture. BLSTM models process input sequences in forward and backward directions, combining the results into one output. While this dual-layer approach improves analysis, it is time-consuming, potentially increasing the risk of fileless malware attacks. A key limitation of BLSTM is the lack of parameter sharing between forward and backward directions. This reduces its ability to capture spatial and temporal features simultaneously, hindering effectiveness in detecting fileless malware. To address this, the ConvLSTM model consolidates feature extraction within a single LSTM cell layer. ConvLSTM breaks down samples into subsequence and uses timesteps for additional feature extraction, enabling spatial-temporal data analysis and improving malware prediction accuracy. The model was tested using a dynamic malware dataset. Unlike traditional LSTM, ConvLSTM integrates convolutional layers, allowing parameter sharing across both spatial and temporal dimensions. This reduces computational complexity and improves model performance in handling multidimensional data. The research re-simulated prior work with BLSTM using the same malware dataset. The Spyder app ran the event simulator and the ConvLSTM model's results replaced BLSTM's using identical parameters. Time, accuracy and loss were the main performance metrics. ConvLSTM outperformed BLSTM, achieving 98% detection accuracy compared to BLSTM's 90%. It also significantly reduced processing time, averaging 10 seconds, while BLSTM took 22 seconds. ConvLSTM experienced lower losses, averaging 10% per epoch versus BLSTM's 20%. In conclusion, ConvLSTM offers superior performance over BLSTM in fileless malware detection. Its enhanced computational efficiency and ability to quickly mitigate threats make it a robust solution for fortifying endpoint security against evolving cyber threats. ConvLSTM holds potential in strengthening defence mechanisms against sophisticated malware attacks, providing a proactive approach to safeguarding networks and data.

Recommender systems have become an essential tool for enhancing user experiences by providing personalized recommendations. In this study, we present a novel approach to constructing a recommender system specifically tailored for Malayalam travel reviews. Our objective was to extract relevant features from these reviews and employ a bidirectional Long Short-Term Memory (BiLSTM) architecture to construct a robust and accurate recommendation model. We focused on four key features extracted from the travel reviews: travel mode, travel type, location climate, and location type. The travel mode feature encompassed the mode of transport opted for the travel such as bus, car, train, etc., while the travel type captured the nature of the trip, including family, friends, or solo travel. Additionally, we considered the climate of the location, including rainy, snowy, hot, and dry, among others, and the location type, such as beach, hilly, or forest destinations. To construct our recommender system, we implemented a BiLSTM architecture, a powerful deep-learning model known for effectively capturing temporal dependencies in sequential data. This architecture allowed us to process the extracted features and learn the underlying patterns within the Malayalam travel reviews. Our experiments were conducted on a comprehensive dataset of Malayalam travel reviews, carefully curated for this study. The dataset encompassed a diverse range of travel experiences, enabling our model to learn from a wide variety of user preferences and recommendations. The performance evaluation of our recommender system yielded promising results. With an accuracy of 83.65 percent, our model showcased its ability to accurately predict and recommend travel options based on the extracted features from the reviews. The high accuracy achieved by our model underscores the effectiveness of the BiLSTM architecture in capturing the nuances of the Malayalam language and understanding the subtle preferences expressed in travel reviews. The practical implications of our work are significant, as it offers a valuable tool for travelers seeking personalized recommendations based on their travel preferences. The use of the Malayalam language in this context expands the reach of recommender systems to a wider audience, catering specifically to individuals who prefer to consume content and make decisions in their native language.

Assessing forecast performance of daily reference evapotranspiration: A comparison of equations, machine and deep learning using weather forecasts

To enhance the intelligence level of switch machine fault diagnosis, this paper proposes an end-to-end diagnostic model integrating Variational Mode Decomposition (VMD), 1D Convolutional Neural Network (1DCNN), and Bidirectional Long Short-Term Memory (BiLSTM). First, VMD is employed to adaptively decompose raw traction force signals into multiple Intrinsic Mode Functions (IMFs) with specific frequency-band characteristics. This multi-scale decomposition enhances the representation of critical fault-related features while effectively suppressing noise interference. Subsequently, 1DCNN extracts localized time-frequency features from each modal component. Finally, a BiLSTM network is introduced to comprehensively capture bidirectional dependencies within temporal features, thereby strengthening the model's ability to perceive fault pattern evolution and enabling precise classification of switch machine faults. Experimental results demonstrate that the proposed VMD-1DCNN-BiLSTM model outperforms conventional 1DCNN, BiLSTM, and CNN-LSTM models in metrics including precision, recall, and F1-score, achieving an overall diagnostic accuracy exceeding 97 %. The model exhibits superior feature extraction and temporal modeling capabilities. Compared to existing methods predominantly reliant on electrical signals such as current or power, this study adopts traction force curves - a direct indicator of switch machine operational states - as the core diagnostic signal. This approach enables earlier detection and enhanced sensitivity to latent fault precursors, providing a more efficient and reliable solution for condition assessment of switch machines and intelligent maintenance of turnout systems.

The integration of renewable energy resources into smart grids has become increasingly important to address the challenges of managing and forecasting energy production in the fourth energy revolution. To this end, artificial intelligence (AI) has emerged as a powerful tool for improving energy production control and management. This study investigates the application of machine learning techniques, specifically ARIMA (auto-regressive integrated moving average) and Bi-LSTM (bidirectional long short-term memory) models, for predicting solar power production for the next year. Using one year of real-time solar power production data, this study trains and tests these models on performance measures such as mean absolute error (MAE) and root mean squared error (RMSE). The results demonstrate that the Bi-LSTM (bidirectional long short-term memory) model outperforms the ARIMA (auto-regressive integrated moving average) model in terms of accuracy and is able to successfully identify intricate patterns and long-term relationships in the real-time-series data. The findings suggest that machine learning techniques can optimize the integration of renewable energy resources into smart grids, leading to more efficient and sustainable power systems.

According to the World Health Organization (WHO), 17.9 million people die yearly from cardiovascular Diseases (CVDs), including heart attacks. Cardiovascular diseases, including heart attack, kill 32% of people globally. Current approaches struggle with electrocardiogram (ECG) signal variability, causing diagnosing errors. The adoption of automated and accurate models for heart disease detection is lacking since conventional methods rely on human analysis, which is time-consuming and error-prone. This work covers the crucial topic of heart disease diagnosis, especially ECG data analysis for cardiovascular disease detection. The integration of the Deep-Convolutional Neural Network (Deep-CNN) and Bidirectional Long Short-Term Memory (Bi-LSTM) model with an Attention Mechanism enhances the accuracy and reliability of heart disease categorisation. The Deep-CNN component efficiently extracts features from capture spatial linkages, while the Bi-LSTM layers handle temporal dependencies to identify patient health patterns over time. The model is evaluated on 303 patient records with 14 clinical characteristics from the University of California, Irvine (UCI) Cleveland Heart Disease dataset. The suggested technique has 97.23% accuracy, 97.72% recall, precision, and 96.90% F1 score. These findings show that the proposed architecture improves diagnostic performance more than boosting ensemble approaches and hybrid models.

In this study, the Bidirectional Long Short-Term Memory (Bi-LSTM) model is used to optimize the unbalanced allocation mechanism in the electricity market, and an empirical analysis is made based on the Nord Pool dataset. It is found that the Bi-LSTM model is significantly better than the traditional method in reducing the total deviation power and unbalanced cost. Specifically, the proposed model reduces the total deviation power by 16.5% and the unbalanced cost by 12.3%. The incentive compatibility index shows that the incentive effect of the proposed model reaches 27.0%. It is higher than that of the F1 mechanism (23.5%) and the F2 mechanism (25.5%). In terms of cost allocation efficiency, the mean square error of the proposed model is 0.018. This shows higher cost allocation accuracy compared with 0.028 in the F1 mechanism and 0.022 in the F2 mechanism. These results show that the proposed model has obvious advantages in improving market operation efficiency, encouraging market participants to reduce deviation electricity, and realizing fair cost allocation. The research conclusion provides a new technical path for the design of the allocation mechanism in the power market and provides strong support for future market management and optimization.

People post their emotional expressions about specific events through social media platforms, such as praise and complaints about city bus travel services. In this paper, we propose a social media-based method to analyze bus service satisfaction. Firstly, we collect microblogs related to bus service from Sina Weibo, and pre-process the collected microblogs. Then the Bidirectional Encoder Representations from Transformers (BERT) pre-trained model is combined with Bi-directional Long Short-Term Memory (BiLSTM) model to form BERT- BiLSTM model, using the pre-processed data to train the BERT-BiLSTM model to realize the classification of text data into positive and negative sentiments, and finally using the method of this paper to analyze and verify the satisfaction of Guangzhou bus services system in 2010. Comparing the method of this paper with traditional machine learning and deep learning methods, the experiments show the effectiveness of the BERT-BiLSTM model. KeywordsSocial mediaBus serviceSentiment analysisBERT-BiLSTMBinary classification

The accurate prediction of Municipal Solid Waste (MSW) electricity generation is very important for the fine management of a city. This paper selects Shanghai as the research object, through the construction of a Bidirectional Long Short-Term Memory (BiLSTM) model, and chooses six influencing factors of MSW generation as the input indicators, to realize the effective prediction of MSW generation. Then, this study obtains the MSW electricity generation capacity in Shanghai by using the aforementioned prediction results and the calculation formula of theMSW electricity generation. The experimental results show that, firstly, the mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE) values of the BiLSTM model are 42.31, 7.390, and 63.32. Second, it is estimated that by 2025, the maximum and minimum production of MSW in Shanghai will be 17.35 million tons and 8.82 million tons under the three scenarios. Third, it is predicted that in 2025, the maximum and minimum electricity generation of Shanghai MSW under the three scenarios will be 512.752 GWh/y and 260.668 GWh/y. Finally, this paper can be used as a scientific information source for environmental sustainability decision-making for domestic MSW electricity generation technology.

The increasing penetration of renewable energy sources and the variability of demand in microgrids have intensified the need for accurate short-term load forecasting (STLF). This study develops a multivariable STLF framework using the Bayesian Regularization Algorithm (BRA) and validates its performance against a Bidirectional Long Short-Term Memory (Bi-LSTM) network. Temperature, humidity, and feeder current were employed as input variables, while electrical load (MW) served as the output. Both models were implemented in MATLAB R2018a with a 120-minute forecasting horizon. Performance evaluation was conducted using regression analysis and statistical metrics, including Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and the coefficient of determination (R²). The results show that the BRA model achieved superior regression accuracy (R² ≈ 0.997), demonstrating robustness and generalization under noisy conditions. By contrast, Bi-LSTM attained R² ≈ 0.991 with lower error magnitudes (MAE = 0.0293, RMSE = 0.0412), reflecting its strength in capturing sequential dependencies. A comparative review with existing literature, including confirms the competitiveness of Bi-LSTM while highlighting BRA’s resilience against noisy fluctuations typical of Nigerian microgrid environments. This study contributes to microgrid energy management by providing practical insights into selecting forecasting models, emphasizing that BRA is well-suited for noisy, small datasets, while Bi-LSTM excels in sequence-sensitive forecasting. The findings also suggest the potential for hybrid BRA–Bi-LSTM architectures to enhance forecasting accuracy in real-world microgrid applications.

The proliferation of misinformation, as insidious and pervasive as water, presents an unprecedented challenge to public discourse and comprehension. Often propagated to further specific ideologies or political objectives, misinformation not only misleads the populace but also fuels online advertising revenue generation. As such, the urgent need to pinpoint and eliminate misinformation from digital platforms has never been more critical. In response to this dilemma, this paper proposes a solution built on the backbone of massive data generation in today’s digital landscape. By leveraging advanced technologies, such as AI-driven systems with deep learning models and natural language processing capabilities, we can monitor and analyze an extensive scope of social media data. This, in turn, facilitates the identification of misinformation across multiple platforms and alerts users to potential propaganda. Central to our study is the development of misinformation classifiers based on a deep bi-directional long short-term memory (Bi-LSTM) model. This model is further enhanced by employing a genetic algorithm (GA), which automates the search for an optimal neural architecture, thereby significantly impacting the training behavior of the deep learning algorithm and the performance of the model being trained. To validate our approach, we compared the efficacy of our proposed model with nine traditional machine learning algorithms and a deep learning model rooted in long short-term memory (LSTM). The results affirmed the superiority of our GA-tuned Bi-LSTM model, which outperformed all other models in detecting misinformation with remarkable accuracy. Our intention with this paper is not to present our model as a comprehensive solution to misinformation but rather as a technological tool that can aid in the process, supplementing and bolstering the existing methodologies in the field of misinformation detection.

Phishing attacks continue to pose a serious threat to cybersecurity, underscoring the need for effective and scalable detection methods. This study evaluates the performance of Recurrent Neural Network (RNN) architectures—specifically Long Short-Term Memory (LSTM) and Bidirectional LSTM (BiLSTM)—for detecting phishing websites based on the sequential patterns in URL structures and webpage content. The LSTM model achieved an Area Under the Curve (AUC) of 0.92, with an overall accuracy of 98.4%, precision of 98.9%, and recall of 97.1%. These results indicate a strong ability to identify phishing URLs with a low false positive rate, although performance declined when detecting sophisticated or zero-day phishing attempts. The BiLSTM model, which incorporates bidirectional context, achieved a higher AUC of 0.95 and improved precision of 91% at a recall of 89%. However, it exhibited a slightly lower overall accuracy of 97.9% and a higher false negative rate. Both models effectively differentiated phishing from legitimate URLs, with BiLSTM offering improved context awareness but at the cost of reduced recall. The results suggest that while BiLSTM enhances contextual understanding, the LSTM model offers better generalization and computational efficiency for real-time deployment. This work highlights the potential of RNN-based models in phishing detection and the importance of balancing sensitivity and specificity in cybersecurity applications.