Bidirectional Long Short-term Memory Research Articles

An Optimal Multi-Modal Approach for Stock Market Price Forecasting with Fused Sentiment Analysis for Real Time Data

Abstract: This study presents an innovative fusion-based methodology that integrates real-time stock market technical indicators with news sentiment analysis from financial news feeds to enhance stock selection decisions. The proposed framework employs a Bidirectional Long Short-Term Memory (Bi-LSTM) model for forecasting stock prices and a Deep Neural Network (DNN) used in conjunction with transformer-based model for sentiment classification, both optimized through the incorporation of real-time datasets. To further refine feature selection, Artificial Bee Colony (ABC) and Firefly algorithms are utilized, significantly improving the accuracy of price trend predictions and sentiment analysis. Technical indicators from stock market data are processed using the Bi-LSTM model to predict future stock prices. Concurrently, sentiment data from the Economic Times is pre-processed through Term Frequency-Inverse Document Frequency (TF-IDF) vectorization and pre-trained transformer models to extract key sentiment scores. The ABC algorithm enhances textual feature selection, while the Firefly algorithm reduces skewness in the error distribution of technical indicators, thereby improving the alignment between forecasted and actual stock prices. The fusion mechanism integrates predicted price movements with sentiment analysis to provide actionable insights for stock selection. The decision-making framework classifies stocks based on the consistency between price movements and sentiment: stocks predicted to rise with positive sentiment are labelled as "Sure Selection"; rising prices with negative sentiment are categorized as "Dicey Selection"; declining prices with positive sentiment result in a "Risky Selection"; and negative sentiment with falling prices trigger an "Avoid Selection". The proposed approach achieved a prediction accuracy of 67.00% for the Bi-LSTM model, with a precision of 0.714, recall of 0.695, and an F1 score of 0.677, demonstrating the model's robustness. The mean absolute error (MAE) for the Bi-LSTM model was 4.30, indicating strong predictive performance. The combined use of ABC and Firefly algorithms optimizes both technical and sentiment features, significantly contributing to the overall performance of the prediction model. By fusing multiple data sources and employing advanced optimization techniques, this methodology offers a unique and effective solution for stock market prediction and decision-making

Optimal design and control strategy for enhanced battery thermal management

The Battery Thermal Management System (BTMS) plays a crucial role in the safety and performance of new energy vehicles. This study proposes an innovative cooling structure design that ingeniously combines the advantages of air cooling, water cooling, and Phase Change Materials (PCMs) to enhance the cooling efficiency of the battery system. Through numerical simulations, the effectiveness of this cooling system was validated and further supported by experimental evidence confirming the accuracy of the simulation results. Subsequently, this research utilized the Response Surface Methodology (RSM) to optimize key parameters of the cooling structure, including the fin height, gap length, and PCM thickness, with their optimal values determined to be 5.20 mm, 17.69 mm, and 3.38 mm, respectively. This parameter optimization work provides precise guidance for the design of battery thermal management systems, contributing to enhanced heat dissipation performance. To further enhance the intelligence of battery thermal management, this study introduced a novel neural network model based on Temporal Convolutional Networks (TCN) and Bidirectional Long Short-Term Memory (BiLSTM) with Multi-Head Attention (MHA). This model is capable of predicting temperature changes during the battery discharge process in real-time with high accuracy. Based on the prediction results, this research designed an efficient control strategy that, compared to a scenario without a control strategy, demonstrated a reduction in energy consumption by 20.87 %, a decrease in maximum temperature by 2.52 K, and a reduction in the maximum temperature difference by 0.05 K. These results not only prove the effectiveness of the control strategy but also provide a new direction for the optimization of battery thermal management systems.

Hybrid Long Short-Term Memory Wavelet Transform Models for Short-Term Electricity Load Forecasting

To ensure the constant availability of electrical energy, power companies must consistently maintain a balance between supply and demand. However, electrical load is influenced by a variety of factors, necessitating the development of robust forecasting models. This study seeks to enhance electricity load forecasting by proposing a hybrid model that combines Sorted Coefficient Wavelet Decomposition with Long Short-Term Memory (LSTM) networks. This approach offers significant advantages in reducing algorithmic complexity and effectively processing patterns within the same class of data. Various models, including Stacked LSTM, Bidirectional Long Short-Term Memory (BiLSTM), Convolutional Neural Network—Long Short-Term Memory (CNN-LSTM), and Convolutional Long Short-Term Memory (ConvLSTM), were compared and optimized using grid search with cross-validation on consumption data from Lome, a city in Togo. The results indicate that the ConvLSTM model outperforms its counterparts based on Mean Absolute Percentage Error (MAPE), Root Mean Squared Error (RMSE), and correlation coefficient (R2) metrics. The ConvLSTM model was further refined using wavelet decomposition with coefficient sorting, resulting in the WT+ConvLSTM model. This proposed approach significantly narrows the gap between actual and predicted loads, reducing discrepancies from 10–50 MW to 0.5–3 MW. In comparison, the WT+ConvLSTM model surpasses Autoregressive Integrated Moving Average (ARIMA) models and Multilayer Perceptron (MLP) type artificial neural networks, achieving a MAPE of 0.485%, an RMSE of 0.61 MW, and an R2 of 0.99. This approach demonstrates substantial robustness in electricity load forecasting, aiding stakeholders in the energy sector to make more informed decisions.

Enhanced Foreign Exchange Volatility Forecasting Using Ceemdan with Optuna-Optimized Ensemble Deep Learning Model

Foreign Exchange (FX) is the largest financial market in the world, with a daily trading volume that significantly exceeds that of stock and futures markets. The prediction of FX volatility is a critical financial concern that has garnered significant attention from researchers and practitioners due to its far-reaching implications in the financial markets. This paper presents a novel hybrid ensemble forecasting model integrating a decomposition strategy and three deep learning (DL) models: Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), and Convolutional Neural Network (CNN). This combination addresses individual models' limitations and further improves the accuracy and stability of FX volatility forecasting. The proposed approach utilizes the CEEMDAN technique to decompose volatility into multiple distinct intrinsic mode functions (IMFs) and merges these IMFs with GARCH and EGARCH volatilities to form the input dataset for the DL models. In addition, we employed an attention mechanism to improve the effectiveness of the DL techniques. Furthermore, the hyperparameters for the DL models are optimized using the Optuna algorithm. Finally, a hybrid ensemble model for forecasting exchange rate volatility is developed by combining the predictions of three distinct DL models. The proposed approach is evaluated against various benchmark models using evaluation measures such as MSE, MAE, HMSE, HMAE, RMSE, Q-LIKE, and the model confidence set (MCS) approach. The results demonstrate that our proposed approach provides accurate and reliable forecasts of FX volatility under different forecasting regimes, making it a valuable tool for financial practitioners and researchers.

A novel bidirectional LSTM network model for very high cycle random fatigue performance of CFRP composite thin plates

This study proposes a novel Double-layer Bidirectional Long Short-Term Memory (BiLSTM) neural network model, shorted as TDA-BiLSTM, which integrates Transfer learning and Attention mechanisms. The model aims to predict the fatigue life of carbon fiber reinforced polymer (CFRP) thin plate structures subjected to very high cycle random vibration fatigue loads. Distinct from conventional servo-hydraulic and ultrasonic fatigue testing methods, this research pioneers the use of a vibration table for very high cycle fatigue (VHCF) testing to fill the gap in the field. By training and validating data across various life ranges, the TDA-BiLSTM model demonstrates significant advantages in training speed and predicting accuracy. Its bidirectional structure and attention mechanism effectively capture complex patterns and long-term dependencies in sequence data. Experimental results indicate that the TDA-BiLSTM model achieves significantly lower Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) compared to Long Short-Term Memory (LSTM) and Long Short-Term Memory with Transfer learning (Tr-LSTM) models across different life ranges on the ε-N curve, indicating higher accuracy and stability in strain life prediction tasks. Additionally, an analysis of typical damage areas using Scanning Electron Microscopy (SEM) reveals the failure mechanisms of CFRP plates under very high cycle vibration fatigue loads.

Recognition of Real-Time Video Activities Using Stacked Bi-GRU with Fusion-based Deep Architecture

Recognizing and understanding human activities in real-time videos is a challenging task due to the complex nature of video data and the need for efficient and accurate analysis. This research pioneers a breakthrough in video activity recognition by introducing a robust framework leveraging the power of a stacked Bidirectional Long Short-Term Memory (Bi-LSTM) and Gated Recurrent Unit (GRU) architecture, harmonized within a fusion-based deep model. The stacked Bi-LSTM-GRU model capitalizes on its dual recurrent architecture, capturing nuanced temporal dependencies within video sequences. The fusion-based deep architecture synergizes spatial and temporal features, enabling the model to discern intricate patterns in human activities. To further enhance the discriminative power of the model, we introduce a fusion module in the proposed deep architecture. The fusion module integrates multi-modal features extracted from different levels of the network hierarchy, allowing for a more comprehensive representation of video activities. We demonstrate the efficacy of our approach through rigorous experimentation on UCF50, UCF101, and HMDB51 datasets. In experiments on the UCF50 dataset, our model achieves an accuracy of 97.01% and 95.86% on training and validation sets respectively, showcasing its proficiency in discerning activities across a diverse range of scenarios. The evaluation extends to the UCF101 dataset, where the proposed approach achieves a competitive accuracy of 97.62% and 96.93% on training and validation sets, surpassing previous benchmarks by a margin of approx 1%. Further-more, on the challenging HMDB51 dataset, the model demonstrates a robust accuracy of 89.71%and 88.88% on training and validation sets, solidifying its efficacy in intricate action recognition tasks.

Comparison of Cobb-Douglas Production Function with Deep Learning Models to Estimate the Productivity of Organised Manufacturing Sector - Indian Perspective

The production function defines the output of a firm, industry, or economy based on various combinations of inputs. This study focuses on estimating the production function of India's manufacturing sector and determining the relationship between its inputs and outputs. To predict the output, measured as Gross Value Added (GVA), the study applies the Cobb Douglas Production Function (CDPF) along with several deep learning techniques, including Feedforward, Recurrent, Long Short-Term Memory (LSTM), and Bidirectional LSTM networks. Model performance is evaluated using metrics like Mean Square Error (MSE), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). The results indicate that the Cobb Douglas Production Function outperforms the deep learning models in predicting GVA for the organized manufacturing sector in India.

A multi-modal Parkinson’s disease diagnosis system from EEG signals and online handwritten tasks using grey wolf optimization based deep learning model

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by the gradual deterioration of motor function, affecting speech, writing, muscle control, and mobility. The existing studies have not utilized both the electroencephalography (EEG) signals and online handwritten tasks together to diagnose PD. The studies have also not explored the EEG signals collected from specific brain regions like the substansia niagra (SN) and ventral tegmental area (VTA), crucial for dopamine production linked to PD. This article proposes a multi-modal PD diagnosis system from EEG signals (collected from SN and VTA regions of the brain), collected during performing online handwritten tasks, using grey wolf optimization (GWO) algorithm. Mel-frequency cepstral coefficients (MFCC) features have been generated from the EEG signals and optimized by the GWO algorithm. The classification (diagnosis) experiments on the optimal number of feature values, obtained from GWO algorithm, have been carried out using bidirectional long short-term memory (BLSTM) variant of recurrent neural network (RNN). The classification experiments have also been conducted using support vector machine (SVM), bagged random forest (BRF), and long short-term memory (LSTM) variant of RNN classifier to have a performance comparison with the proposed method. The effectiveness of the introduced PD diagnosis system has been analyzed on a self-generated dataset named EEG signal based on online handwriting (ESOH). A maximum classification accuracy of 99.30% has been achieved from the proposed PD diagnosis system. The experimental outcomes illustrate that the introduced PD diagnosis system outperforms the state-of-the-art PD diagnosis systems relying on the EEG signals for diagnosing the PD.

IPerFEX-2023: Indonesian personal financial entity extraction using indoBERT-BiGRU-CRF model

There is minimal research focusing on applications of Indonesian named entity recognition (NER) in a specific domain. This study proposes an Indonesian personal financial entity extraction task that can be utilized in a financial assistant chatbot system to interpret user’s financial situation for personalization. Due to the simplicity that a chatbot has, it can promote financial management practices to youth as early as possible in their career. However, the challenge in financial NER is numerical entity extraction that relies heavily on contextual information and suffers the out-of-vocabulary (OOV) problem. Therefore, to extract 15 personal financial entities in daily Indonesian discussions (expense-type, expense-amount, income-type, income-amount, asset-type, asset-amount, saving-type, saving-amount, liability-type, liability-amount, family, time, financial-goal, age, and occupation), this research proposes a dataset, IPerFEX-2023, trained using the Bidirectional Gated Recurrent Unit BiGRU and Conditional Random Field (CRF) with Indonesian Bidirectional Encoder Representations from Transformers (IndoBERT) pre-trained model for feature embeddings (IndoBERT-BiGRU-CRF). It is compared with the corresponding Bidirectional Long Short-Term Memory (BiLSTM) (IndoBERT-BiLSTM-CRF) as baseline. Not only the IndoBERT-BiGRU-CRF model achieves the best performance with a 0.73 F1-score, but it is also 14% faster on average compared to the corresponding baseline model due to its simpler unit structure. This paper also discusses future directions covering model enhancement strategy based on the error analysis result and complementary tasks needed to complete personalization

Adaptive signal fusion for swashplate pump fault detection using bidirectional long short-term memory and wavelet scattering transform

In construction equipment, swashplate pumps are vital for uninterrupted operations. However, these pumps often experience multiple simultaneous faults. Current detection methods typically rely on a single signal, insufficient for accurately identifying multiple faults in this hydraulic system. To improve accuracy, we utilize multiple signals that provide comprehensive information, using an adaptive signal-level fusion technique. This method involves analyzing pressure, shaft torque, and swashplate torque signals to identify potential faults. The study employs a Bidirectional Long Short-Term Memory (BiLSTM) model to analyze each type of signal, assessing the model's ability to classify different faults. Once detection accuracy for each fault type is assessed, we propose using a weight matrix to adaptively fuse signals based on their assigned weights. Additionally, the Wavelet Scattering Transform (WST) is utilized for feature extraction from signals with low variance, and Bayesian optimization is applied to find the optimal settings for the BiLSTM model. The proposed approach is tested under various scenarios, including the impact of white Gaussian noise, to evaluate its effectiveness and stability. The results show that the adaptive signal-level fusion approach outperforms traditional methods and individual signal analyses in accurately and robustly detecting swashplate pump faults, highlighting its potential to significantly improve the reliability of construction equipment.

Post-Stroke Dysarthria Voice Recognition based on Fusion Feature MSA and 1D

Post-stroke Dysarthria (PSD) is one of the common sequelae of stroke. PSD can harm patients’ quality of life and, in severe cases, be life-threatening. Most of the existing methods use frequency domain features to recognize the pathological voice, which makes it hard to completely represent the characteristics of pathological voice. Although some results have been achieved, there is still a long way to go for practical applications. Therefore, an improved deep learning-based model is proposed to classify between the pathological voice and the normal voice, using a novel fusion feature (MSA) and an improved 1D ResNet network hybrid bi-directional LSTM with dilated convolution (named 1D DRN-biLSTM). The experimental results show that our fusion features bring greater improvement in pathological speech recognition than the method that only analyzes the MFCC features, and can better synthesize the hidden features that characterize pathological speech. In terms of model structure, the introduction of dilated convolution and LSTM can further improve the performance of the 1D Resnet network, compared to ordinary networks such as CNN and LSTM. The accuracy of this method reaches 82.41% and 100% at the syllable level and speaker level, respectively. Our scheme outperforms other existing methods in terms of feature learning capability and recognition rate, and will help to play an important role in the assessment and diagnosis of PSD in China.

Deep Learning Based Atrial Fibrillation Detection Using Combination of Dimensionality Reduction Techniques and RR Interval Features

Purpose: Accurate detection of Atrial Fibrillation (AF) has great significance in the field of medical science which can reduce the rate of mortality and morbidity. The present study focuses on Electrocardiography (ECG) signal classification using dimensionality reduction techniques combined with R wave to R wave interval (RR interval) features. Materials and Methods: In the first approach, Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Independent Component Analysis (ICA), and Probabilistic Principal Component Analysis (PPCA) are performed independently on denoised ECG signal using Discrete Wavelet Transform (DWT) for the classification of ECG signal. In the second approach, the dimensionality reduction techniques combined with RR interval features are used for the classification of ECG signal. Results: Machine Learning (ML) algorithms such as Decision Tree (DT), Support Vector Machine (SVM), and Deep Learning (DL) algorithms such as Long Short Term Memory (LSTM) and Bi-Directional LSTM (BiLSTM) are used for classification purposes. Conclusion: The proposed methodology provided an overall accuracy of 93.65% with PCA and LSTM classifier and an overall accuracy of 99.45% with PCA combined with RR interval features and LSTM classifier. The developed technology has potential applications in many practical solutions.

Real-time prediction of the week-ahead flood index using hybrid deep learning algorithms with synoptic climate mode indices

This paper aims to propose a hybrid deep learning (DL) model that combines a convolutional neural network (CNN) with a bi-directional long-short term memory (BiLSTM) for week-ahead prediction of daily flood index (IF) for Bangladesh. The neighbourhood component analysis (NCA) is assigned for significant feature selection with synoptic-scale climatic indicators. The results successfully reveal that the hybrid CNN-BiLSTM model outperforms the respective benchmark models based on forecasting capability, as supported by a minimal mean absolute error and high-efficiency metrics. With respect to IF prediction, the hybrid CNN-BiLSTM model shows over 98% of the prediction errors were less than 0.015, resulting in a low relative error and superiority performance against the benchmark models in this study. The adaptability and potential utility of the suggested model may be helpful in subsequent flood monitoring and may also be beneficial to policymakers at the federal and state levels.

Enhanced offshore wind resource assessment using hybrid data fusion and numerical models

Wind resource assessments are crucial for pre-construction planning of wind farms, especially offshore. This study proposes a novel hybrid model integrating Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Empirical Wavelet Transform (EWT) for enhanced wind speed forecasting. This secondary decomposition reduces forecasting complexity by processing high-frequency signals. A Bidirectional Long Short-Term Memory (BiLSTM) neural network optimized with the Grey Wolf Optimizer (GWO) is then employed for forecasting. The model’s accuracy is evaluated using simulated wind speeds along the coast of Denmark, combined with lidar measurements through data fusion. This approach demonstrates significant improvements in prediction accuracy, highlighting its potential for offshore wind resource assessment.

BiLSTM-InceptionV3-Transformer-fully-connected model for short-term wind power forecasting

Accurate short-term wind power prediction can help grid managers optimize the transmission and storage of electricity from wind farms, avoiding energy wastage and unnecessary operating costs, and ensuring the stability and reliability of power supply. However, stochastic nature and uncertainty data pose challenges for accurate wind power prediction models. This study combines complementary ensemble empirical mode decomposition (CEEMD), discrete wavelet transform (DWT), bidirectional long short-term memory (BiLSTM), InceptionV3, Transformer, and fully connected neural networks for wind power prediction. The proposed prediction model can fully capture the nonlinear and dynamic characteristics of the wind speed data that are decomposed into intrinsic modal functions by CEEMD. Simultaneously, the prediction model can capture the local characteristics and cyclic variations of the data at different time scales based on fifth-order DWT. A dataset spanning 378 days from Natal is employed and the proposed model is compared with 19 other networks (e.g., ResNet, GoogleNet, SqueezeNet) and 15 other traditional regression algorithms (e.g., Gaussian process regression, least-angle regression, and nearest neighbors regression). The experimental results show that the designed BiLSTM-InceptionV3-Transformer-fully-connected model (BITFM) can reduce the mean absolute error by at least 38.75 % and the root mean squared error by at least 33.33 % compared with 19 other networks. Compared with 15 traditional regression prediction models, BITFM can reduce the mean absolute error by at least 57.41 % and the root mean squared error by at least 53.95 %. To further validate the prediction performance of BITFM, this study conducts tests employing data from the Santa Vitoria do Palmar. The experimental results indicate that BITFM consistently achieved the best performance metrics.

A novel spatio-temporal attention-based bidirectional LSTM model for moisture content prediction in drying process

Accurate prediction of moisture content is significantly crucial for ensuring process stability and product quality in the cylinder drying process. However, the drying process exhibits complex spatio-temporal characteristics and strong interference, which make accurate prediction challenging for the deep learning approach. To address this issue, this article proposes a new spatio-temporal attention-based bidirectional long-short temporal memory network (STA-BiLSTM) model for accurate moisture content prediction. First, Maximum Relevance Minimum Redundancy (mRMR) is adopted to identify optimal features highly related to moisture content. Secondly, bidirectional long-short temporal memory (Bi-LSTM) network is utilized to extract temporal dependencies from the sequential data. Subsequently, spatio-temporal attention mechanisms are designed to adaptively focus on the most relevant features and timesteps, enhancing the model’s generalization ability. Finally, due to the harsh industrial environment, eXtreme Gradient Boosting (XGBoost) is adapted to improve generalizability and robustness. Extensive experiments on a real industrial dataset of the drying process demonstrate that the proposed STA-BiLSTM approach significantly outperforms alternative approaches for predicting moisture content, validating its effectiveness and superiority.

A novel method for rice identification: Coupling Raman spectroscopy with Fourier spectrum and analyzing with deep learning

The development of detection methods for food adulteration is of great significance for promoting innovation and development in food safety supervision technology. In this study, we proposed a novel model: Raman-Fourier-BiLSTM-CNN (RFBC), which combines Raman spectroscopy with deep learning to achieve precise identification of seven brands of japonica rice in the northeastern region of China. This model focuses on revealing the worth of the Fourier spectrum of Raman spectra. The raw Raman spectra and their Fourier spectra are cleverly coupled through a Bi-directional Long Short-Term Memory (BiLSTM) connection structure, and the resulting composite features are deeply analyzed by a Convolutional Neural Network (CNN). Compared to traditional algorithms, RFBC exhibits superior feature extraction capabilities. To further evaluate the recognition performance of RFBC, this study compared it with the RFBC model containing only the original Raman spectrum branch, as well as with machine learning models based on Principal Component Analysis (PCA), including Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), and eXtreme Gradient Boosting (XGBoost). The results show that, compared with the other five models, RFBC can accurately identify different brands of japonica rice, with significant advantages in accuracy, precision, recall, and other aspects. RFBC achieves a classification accuracy of 97.1 %, macro precision of 97.2 %, macro recall of 97.1 %, and macro F1-score of 97.2 %. The system proposed in this study can more accurately identify brands of japonica rice, providing strong technical support for combating counterfeit japonica rice products.

Comparative analysis of novel data‐driven techniques for remaining useful life estimation of wind turbine high‐speed shaft bearings

AbstractAs the global momentum for wind power generation accelerates, the industry faces substantial challenges due to premature failures in wind turbine components. These failures, particularly in critical elements like the high‐speed shaft bearing, lead to significant operational losses, including unplanned downtime and elevated maintenance costs. To mitigate these issues, it's crucial to have precise predictions of the remaining useful life (RUL) of these components, enabling timely interventions and more efficient maintenance schedules. This article proposes advanced, data‐driven approaches for estimating the RUL of wind turbine high‐speed shaft bearings, utilizing cutting‐edge techniques such as long short‐term memory (LSTM), bidirectional LSTM (BiLSTM), gated recurrent units (GRU), and random forest (RF) algorithms. Our analysis leverages vibration data from a 2 MW wind turbine equipped with a 20‐tooth pinion gear, providing a thorough validation and comparison of these methodologies against traditional models. Our results reveal that the LSTM and BiLSTM models excel in both accuracy and computational efficiency for predicting RUL and enhancing system prognosis, surpassing the performance of conventional RF and GRU methods. This research underscores the potential of our innovative data‐driven strategies to develop effective RUL estimation algorithms, significantly advancing wind turbine proactive operation and maintenance operations.

A two-task predictor for discovering phase separation proteins and their undergoing mechanism.

Liquid-liquid phase separation (LLPS) is one of the mechanisms mediating the compartmentalization of macromolecules (proteins and nucleic acids) in cells, forming biomolecular condensates or membraneless organelles. Consequently, the systematic identification of potential LLPS proteins is crucial for understanding the phase separation process and its biological mechanisms. A two-task predictor, Opt_PredLLPS, was developed to discover potential phase separation proteins and further evaluate their mechanism. The first task model of Opt_PredLLPS combines a convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) through a fully connected layer, where the CNN utilizes evolutionary information features as input, and BiLSTM utilizes multimodal features as input. If a protein is predicted to be an LLPS protein, it is input into the second task model to predict whether this protein needs to interact with its partners to undergo LLPS. The second task model employs the XGBoost classification algorithm and 37 physicochemical properties following a three-step feature selection. The effectiveness of the model was validated on multiple benchmark datasets, and in silico saturation mutagenesis was used to identify regions that play a key role in phase separation. These findings may assist future research on the LLPS mechanism and the discovery of potential phase separation proteins.

Retraction Note: Smart cataract detection system with bidirectional LSTM

Retraction Note: Smart cataract detection system with bidirectional LSTM