Short-term Memory Algorithm Research Articles

Long Short-Term Memory Algorithm for Personalized Tacrolimus Dosing: A Simple and Effective Time Series Forecasting Approach Post-Lung Transplantation.

Management of tacrolimus trough levels influences morbidity and mortality after lung transplantation. Several studies have explored pharmacokinetic and artificial intelligence models to monitor tacrolimus levels. However, many models depend on a wide range of variables, some of which, like genetic polymorphisms, are not commonly tested for in regular clinical practice. This study aimed to verify the efficacy of a novel approach simply utilizing time series data of tacrolimus dosing, with the objective of accurately predicting trough levels in the variety of clinical settings. Data encompassing 36 clinical variables for each patient were gathered, and a multivariate long short-term memory algorithm was applied to forecast subsequent tacrolimus trough levels based on the selected clinical variables. The tool was developed using a dataset of 87,112 data points from 117 patients and its efficacy was confirmed using six additional cases. Shapley Additive exPlanations revealed a significant correlation between trough levels and prior dose-concentration data. By using simple trend learning of dose, administration route, and previous trough levels of tacrolimus, we could predict values within 30% of the actual values for 88.5% of time points, which facilitated the creation of a tool for simulating tacrolimus trough levels in response to dosage adjustments. The tool exhibited the potential for rectifying clinical misjudgments in a simulation cohort. Utilizing our time series forecasting tool, precise prediction of trough levels is attainable independently of other clinical variables, through the analysis of historical tacrolimus dose-concentration trends alone.

Pore size classification and prediction based on distribution of reservoir fluid volumes utilizing well logs and deep learning algorithm in a complex lithology

Pore size analysis plays a pivotal role in unraveling reservoir behavior and its intricate relationship with confined fluids. Traditional methods for predicting pore size distribution (PSD), relying on drilling cores or thin sections, face limitations associated with depth specificity. In this study, we introduce an innovative framework that leverages nuclear magnetic resonance (NMR) log data, encompassing clay-bound water (CBW), bound volume irreducible (BVI), and free fluid volume (FFV), to determine three PSDs (micropores, mesopores, and macropores). Moreover, we establish a robust pore size classification (PSC) system utilizing ternary plots, derived from the PSDs.Within the three studied wells, NMR log data is exclusive to one well (well-A), while conventional well logs are accessible for all three wells (well-A, well-B, and well-C). This distinction enables PSD predictions for the remaining two wells (B and C). To prognosticate NMR outputs (CBW, BVI, FFV) for these wells, a two-step deep learning (DL) algorithm is implemented. Initially, three feature selection algorithms (f-classif, f-regression, and mutual-info-regression) identify the conventional well logs most correlated to NMR outputs in well-A. The three feature selection algorithms utilize statistical computations. These algorithms are utilized to systematically identify and optimize pertinent input features, thereby augmenting model interpretability and predictive efficacy within intricate data-driven endeavors. So, all three feature selection algorithms introduced the number of 4 logs as the most optimal number of inputs to the DL algorithm with different combinations of logs for each of the three desired outputs. Subsequently, the CUDA Deep Neural Network Long Short-Term Memory algorithm(CUDNNLSTM), belonging to the category of DL algorithms and harnessing the computational power of GPUs, is employed for the prediction of CBW, BVI, and FFV logs. This prediction leverages the optimal logs identified in the preceding step. Estimation of NMR outputs was done first in well-A (80% of data as training and 20% as testing). The correlation coefficient (CC) between the actual and estimated data for the three outputs CBW, BVI and FFV are 95%, 94%, and 97%, respectively, as well as root mean square error (RMSE) was obtained 0.0081, 0.098, and 0.0089, respectively. To assess the effectiveness of the proposed algorithm, we compared it with two traditional methods for log estimation: multiple regression and multi-resolution graph-based clustering methods. The results demonstrate the superior accuracy of our algorithm in comparison to these conventional approaches. This DL-driven approach facilitates PSD prediction grounded in fluid saturation for wells B and C.Ternary plots are then employed for PSCs. Seven distinct PSCs within well-A employing actual NMR logs (CBW, BVI, FFV), in conjunction with an equivalent count within wells B and C utilizing three predicted logs, are harmoniously categorized leading to the identification of seven distinct pore size classification facies (PSCF). this research introduces an advanced approach to pore size classification and prediction, fusing NMR logs with deep learning techniques and extending their application to nearby wells without NMR log. The resulting PSCFs offer valuable insights into generating precise and detailed reservoir 3D models.

Safety management system of new energy vehicle power battery based on improved LSTM

With the development of sustainable economy, new energy materials are widely used in various industries, and many cars also adopt new energy power batteries as power sources. However, it is currently not possible to accurately diagnose faults in power batteries, which results in the safety of power batteries not being guaranteed. To address this issue, this study utilizes the Whale Optimization Algorithm to improve the Long Short-Term Memory algorithm and constructs a fault diagnosis model based on the improved algorithm. The purpose of using this model for fault diagnosis of power batteries is to strengthen the safety management of batteries. This study first conducted experiments on the improved algorithm and obtained an accuracy of 95.3%. The simulation results of the fault diagnosis model showed that the diagnosis time was only 1.2s. The analysis of the power battery showed that after using this model, the safety performance has been improved by 90.1%, while the maintenance cost has been reduced to 20.3% of the original. The above results verify that the fault diagnosis model based on the improved algorithm can accurately diagnose faults in power batteries, thereby improving the safety of power batteries.

Data-driven energy utilization for plug-in hybrid electric bus with driving patten application and battery health considerations

The efficient utilization of energy is a crucial consideration for plug-in hybrid electric buses (PHEB). However, achieving the optimal energy management strategy (EMS) for PHEB necessitates the harmonious optimization of both driving conditions and battery status. Consequently, an innovative EMS that integrates the data-driven prediction of the driving cycles and battery health status. Explicitly, the bi-directional long short-term memory algorithm is employed to reconstruct and train the health status data of power batteries, enabling the accurate estimation of their State of Health (SOH) values. Moreover, a database of historical driving behavior is meticulously collected through a real-world PHEB platform, and the driving cycles for a certain while are predicted with the clustering algorithm. Then, this invaluable information on the battery status and driving cycles is seamlessly introduced into an adaptive-network-based fuzzy inference system (ANFIS) to regulate of the energy conversion factor in the following equivalent consumption minimization strategy (ECMS). Besides, the offline global optimization of equivalent fuel consumption with the Pontryagin's minimum principle (PMP) is preliminarily conducted to generate the pre-established energy conversion factors, facilitating convenient access to the engine and motor's reasonable power distribution range. Finally, simulations and experiments have validated the effectiveness of the proposed EMS, demonstrating a 23.57 % reduction in battery SOC depletion and a 10.53 % decrease in fuel consumption compared to traditional ECMS methods. More importantly, the proposed method has been validated for its executability in embedded devices through real-world bench experiments. This research holds significant reference value for energy utilization and practical implementation in the PHEB, thereby contributing to the advancement of knowledge in this field.

Optimizing the long short-term memory algorithm to improve the accuracy of infectious diseases prediction

<p><span lang="EN-US">This study discusses the implementation of the proposed optimized</span><span lang="EN-US">long short-term memory (LSTM) to predict the number of infectious disease cases that spread in Central Java, Indonesia. The proposed model is developed by optimizing the output layer, which affects the output value of the cell state. This study used cases of four infectious diseases in Indonesia's Central Java Province, namely COVID-19, dengue, diarrhea, and hepatitis A. This model was compared to basic LSTM and MinMax schaler LSTM improvement to see the difference in the accuracy of each disease. The results showed a significant difference in the average prediction results with real cases between the three models. The main objectives of this study were: </span><span lang="EN-US">modifying the LSTM algorithm to predict the number of infectious disease cases to get a smaller residual value</span><span lang="EN-US">, <span>comparing the results of the optimization accuracy of the LSTM algorithm with the LSTM algorithm in previous studies, and evaluating the use of spatial variables in applying infectious disease prediction models using the LSTM algorithm. The results found that the performance difference between the proposed optimization algorithm and the model in the previous study was obtained. The proposed LSTM optimization algorithm had an accuracy improvement of about 2% over the previous model.</span></span></p>

Designing Human-Robot Communication in the Indonesian Language Using the Deep Bidirectional Long Short-Term Memory Algorithm

Designing Human-Robot Communication in the Indonesian Language Using the Deep Bidirectional Long Short-Term Memory Algorithm

Air pollutant removal performance using a BiLSTM-based demand-controlled ventilation method after tunnel blasting

Efficient tunnel ventilation is essential for ensuring construction safety and protecting personnel health during tunnel construction. This study proposes a demand-controlled ventilation (DCV) method on the basis of deep learning algorithm to both improve pollutant removal efficiency and reduce energy consumption. The DCV method utilizes a two-layer bidirectional long short-term memory algorithm (BiLSTM) to predict pollutant concentrations. The air volume is dynamically adjusted based on the gaseous pollutant removal requirements. The coefficient of ventilation performance (COVP) is proposed to evaluate the performance of two ventilation methods (DCV and constant air-volume ventilation (CAV)) through computational fluid dynamics (CFD) simulations. The results show that the DCV results in a lower maximum average CO concentration and higher removal efficiency in the heading area (372.3 mg/m3) than the CAV does (404.1 mg/m3). The fan's energy consumption of DCV is 64.6% lower than that of CAV during a 1000 s ventilation period. The COVPs for both methods exhibit temporal variation and achieves their maximums (2.25 for DCV and 0.741 for CAV) after reaching the constraint conditions (air volume threshold). The DCV method expedites pollutant elimination, reduces construction waiting period, and minimizes energy consumption, providing a novel application of a deep learning algorithm in construction engineering.

Prediction of Dissolved Oxygen Factor at Oncheon Stream Watershed Using Long Short-Term Memory Algorithm

Prediction of Dissolved Oxygen Factor at Oncheon Stream Watershed Using Long Short-Term Memory Algorithm

Airborne particulate matter measurement and prediction with machine learning techniques

Air quality is a fundamental component of a healthy environment for human beings. Monitoring networks for air pollution have been established in numerous industrial zones. The data collected by the pervasive monitoring devices can be utilized not only for determining the current environmental condition, but also for forecasting it in the near future. This paper considers the applications of different machine learning methods for the prediction of the two most widely used quantities. Particulate matter (PM) with a diameter of 2.5 and 10 µm, respectively. The data are collected via a proprietary monitoring station, designated as the Ecolumn. The Ecolumn monitors a number of key parameters, including temperature, pressure, humidity, PM 1.0, PM 2.5, and PM 10, in a timely manner. The data were employed in the development of multiple models based on selected machine learning methods. The decision tree, random forest, recurrent neural network, and long short-term memory models were employed. Experiments were conducted with varying hyperparameters and network architectures. Different time scales (10 min, 1 h, and 24 h) were examined. The most optimal results were observed for the Long Short-Term Memory algorithm when utilizing the shortest available time spans (shortest averaging times). The decision tree and random forest algorithms demonstrated unexpectedly high performance for long averaging times, exhibiting only a slight decline in accuracy compared to neural networks for shorter averaging times. Recommendations for the potential applicability of the tested methods were formulated.

Machine learning-based short-term solar power forecasting: a comparison between regression and classification approaches using extensive Australian dataset

Solar energy production is an intermittent process that is affected by weather and climate conditions. This can lead to unstable and fluctuating electricity generation, which can cause financial losses and damage to the power grid. To better control power production, it is important to predict solar energy production. Big data and machine learning algorithms have yielded excellent results in this regard. This study compares the performance of two different machine learning approaches to solar energy production prediction: regression and classification. The regression approach predicts the actual power output, while the classification approach predicts whether the power output will be above or below a certain threshold. The study found that the random forest regressor algorithm performed the best in terms of accuracy, with mean absolute errors and root mean square errors of 0.046 and 0.11, respectively. However, it did not predict peak power values effectively, which can lead to higher errors. The long short-term memory algorithm performed better in classifying peak power values. The study concluded that classification models may be better at generalizing than regression models. This proposed approach is valuable for interpreting model performance and improving prediction accuracy.

Predicting Stock Prices Using LSTM Networks: A Comprehensive Approach

Abstract: Prediction of stock prices calls for strong algorithmic foundations for predictions of greater magnitude in share prices because the stock market epitomizes volatility. There exist several models used for the prediction of stock prices. The Long ShortTerm Memory algorithm is one model that seems well-suited for such time series problems. The key objective is to best predict current trends in the market and stock prices, which can be done through point prediction, scenario prediction, anomaly prediction, interval prediction, and volatility prediction. The objective of the study is to provide insight to investors and analysts to understand and predict the behavior of the stock market

A data-driven optimization model for renewable electricity supply chain design

The rising electricity demand, driven by technological progress and population growth, has made the design of the electricity supply chain a major challenge. The transition to renewable energy sources has led to a substantial reduction in the costs and greenhouse gases emissions of electricity production. This study proposes a hybrid approach consisting of a recurrent neural network and an optimization model for designing a resilient electricity supply chain based on wind and solar energies. First, the electricity demand is forecast using the long short-term memory algorithm. Second, based on the forecast demand, a two-stage stochastic programming model is developed for designing an electricity supply chain under disruption risks. Partial and complete disruptions in power lines and facilities, including power plants, distribution and subtransmission substations, and low-power solar panel farms are considered. The objective of the proposed model is to minimize the total cost of the supply chain by making optimal decisions on facility location; technology selection for power plants; setting up transmission lines; and the quantity of power generation, power transmission, and power shortage in low- and high-voltage energy consumption systems. Three strategies are adopted to enhance the supply chain's resilience: multiple renewable energy sources, multiple power transmission lines, and lateral power lines. Lastly, the Sistan and Baluchestan Electric Power Distribution Company in Iran is used as a case study for this work to demonstrate the applicability of the proposed approach, analyze the findings, and derive relevant managerial insights.

Predicting hourly heating load in district heating system based on the hybrid Bi-directional long short-term memory and temporal convolutional network model

Accurate heating load prediction is essential to supply-demand collaboration in district heating systems and energy sustainability. With the development of the Internet of Things technology, massive monitoring data is collected that provides abundant data sources for the prediction models. However, the missing data from equipment anomalies, network transmissions and the complexity of heat load changing cause difficulty in realizing accurate heat load prediction. Therefore, the hybrid model based on bidirectional long short-term memory network and temporal convolutional network proposed in this paper to improve the accuracy of the heat load prediction from the aspects of reconstruction of missing values and extraction of complex features. The bidirectional long short-term memory algorithm imputes missing values based on bi-directional features of existing data and provides solid data foundation for subsequent training of predictive models. The temporal convolutional network achieves high-level feature extraction and parallel computation, which contributes to the effective modeling of complex changes in heat load and the realization of accurate heat load prediction. Comparative study was conducted with four heat exchange stations of actual environment to evaluate the performance of the prediction model. The results showed the mean absolute percentage error of the hybrid model decreased to 2.47% which is lower than the State-of-the-Art models, such as long short-term memory, temporal convolutional network, and support vector regression etc. The superior results further validated the feasibility of this hybrid model in heat load prediction.

Link prediction method for social networks based on a hierarchical and progressive user interaction matrix

Link prediction within social networks represents a notable research area. Leveraging semantic text and multidimensional user interaction features, this paper introduces a hierarchical and progressive user interaction matrix-based link prediction method in social networks. First, considering the dynamic nature of a user’s semantic text features, the word2vec algorithm was employed to extract text features. Given the long short-term memory algorithm’s proficiency in learning long-term sequence dependencies, we proposed a method to represent time-series text called “TT2vec” in order to capture the evolution pattern of a user’s semantic text features. Second, to address the multidimensionality of user interactions, we constructed a hierarchical and progressive interaction network adjacency matrix, thereby deriving advanced user interaction features using an enhanced graph convolutional network. Then, using TT2vec to further capture the interaction network’s evolution, we proposed HPIN2vec, which is a method for representing hierarchical and progressive user interaction networks. Finally, for the users’ semantic text features and multidimensional interaction features, we proposed a link prediction method based on graph attention networks (GATs) and users’ dynamic features, considering the advantage of GAT aggregating neighbor node features. Experimental findings indicated that the proposed method not only identifies immediate dynamics in user features but also enhances link prediction accuracy through complex hierarchical and progressive interaction patterns.

LSTM-Autoencoder Deep Learning Model for Anomaly Detection in Electric Motor

Anomaly detection is the process of detecting unusual or unforeseen patterns or events in data. Many factors, such as malfunctioning hardware, malevolent activities, or modifications to the data’s underlying distribution, might cause anomalies. One of the key factors in anomaly detection is balancing the trade-off between sensitivity and specificity. Balancing these trade-offs requires careful tuning of the anomaly detection algorithm and consideration of the specific domain and application. Deep learning techniques’ applications, such as LSTMs (long short-term memory algorithms), which are autoencoders for detecting an anomaly, have garnered increasing attention in recent years. The main goal of this work was to develop an anomaly detection solution for an electrical machine using an LSTM-autoencoder deep learning model. The work focused on detecting anomalies in an electrical motor’s variation vibrations in three axes: axial (X), radial (Y), and tangential (Z), which are indicative of potential faults or failures. The presented model is a combination of the two architectures; LSTM layers were added to the autoencoder in order to leverage the LSTM capacity for handling large amounts of temporal data. To prove the LSTM efficiency, we will create a regular autoencoder model using the Python programming language and the TensorFlow machine learning framework, and compare its performance with our main LSTM-based autoencoder model. The two models will be trained on the same database, and evaluated on three primary points: training time, loss function, and MSE anomalies. Based on the obtained results, it is clear that the LSTM-autoencoder shows significantly smaller loss values and MSE anomalies compared to the regular autoencoder. On the other hand, the regular autoencoder performs better than the LSTM, comparing the training time. It appears then, that the LSTM-autoencoder presents a superior performance although it was slower than the standard autoencoder due to the complexity of the added LSTM layers.

Machine learning insights into PM2.5 changes during COVID-19 lockdown: LSTM and RF analysis in Mashhad.

This study seeks to investigate the impact of COVID-19 lockdown measures on air quality in the city of Mashhad employing two strategies. We initiated our research using basic statistical methods such as paired sample t-tests to compare hourly PM2.5 data in two scenarios: before and during quarantine, and pre- and post-lockdown. This initial analysis provided a broad understanding of potential changes in air quality. Notably, a low reduction of 2.40% in PM2.5 was recorded when compared to air quality prior to the lockdown period. This finding highlights the wide range of factors that impact the levels of particulate matter in urban settings, with the transportation sector often being widely recognized as one of the principal causes of this issue. Nevertheless, throughout the period after the quarantine, a remarkable decrease in air quality was observed characterized by distinct seasonal patterns, in contrast to previous years. This finding demonstrates a significant correlation between changes in human mobility patterns and their influence on the air quality of urban areas. It also emphasizes the need to use air pollution modeling as a fundamental tool to evaluate and understand these linkages to support long-term plans for reducing air pollution. To obtain a more quantitative understanding, we then employed cutting-edge machine learning methods, such as random forest and long short-term memory algorithms, to accurately determine the effect of the lockdown on PM2.5 levels. Our models' results demonstrated remarkable efficacy in assessing the pollutant concentration in Mashhad during lockdown measures. The test set yielded an R-squared value of 0.82 for the long short-term memory network model, whereas the random forest model showed a calculated cross-validation R-squared of 0.78. The required computational cost for training the LSTM and the RF models across all data was 25min and 3s, respectively. In summary, through the integration of statistical methods and machine learning, this research attempts to provide a comprehensive understanding of the impact of human interventions on air quality dynamics.

Klasifikasi Sentimen Terhadap Kualitas Aplikasi Bahan Ajar Digital Akademik Universitas Terbuka di Google Play

Terbuka University is a leading institution that implements the optimization of digital transformation, especially in distance learning systems. To improve the quality of service to students and stakeholders, Terbuka University has developed the Terbuka University Digital Learning Materials application. This application offers several learning modules that can be accessed through the Google Play Store. This research aims to classify data using different labels related to reviews of the Terbuka University Digital Learning Materials application using the Long Short-Term Memory classification algorithm. Evaluation is conducted to find accuracy, f1-score, precision, and recall values. The research results show that classification with Long Short-Term Memory achieves an accuracy of 76.72% with the Vader label, and the accuracy with the TextBlob label reaches 74.21%. Confusion matrix evaluation shows precision results of 0.91 and recall of 0.78, with an f1-score of 0.84 for the Vader label. For the TextBlob label, the precision is 0.96, recall is 0.45, and the f1-score is 0.61. This research contributes positively to understanding the evaluation and classification of reviews of the Terbuka University Digital Learning application. Implementing the Long Short-Term Memory algorithm with the Vader label can be an effective choice to improve service and learning quality through the application.

Implementation of Neural Machine Translation in Translating from Indonesian to Sasak Language

Language translation is part of Natural Language Processing, also known as Machine Translation, which helps the process of learning foreign and regional languages using translation technology in sentence form. In Lombok, there are still people who are not very fluent in Indonesian because Indonesian is generally only used at formal events. This research aimed to develop a translation model from Indonesian to Sasak. The method used was the Neural Machine Translation with the Recurrent Neural Network - Long Short Term Memory architecture and the Word2Vec Embedding with a sentence translation system. The dataset used was a parallel corpus from the Tatoeba Project and other open sources, divided into 80% training and 20% validation data. The result of this research was the application of Neural Machine Translation with the Recurrent Neural Network - Long Short Term Memory algorithm, which could produce a model with an accuracy of 99.6% in training data and 71.9% in test data. The highest ROUGE evaluation metric result obtained on the model was 88%. This research contributed to providing a translation model from Indonesian to Sasak for the local community to facilitate communication and preserve regional language culture.

Algoritma Gated Recurrent Unit untuk Prediksi Harga Indeks Penutupan Saham LQ45

The Indonesia Stock Exchange (IDX) states that stocks, including LQ45 stocks, which constitute the stock market index for the IDX, have become one of the preferred investment options for the public. Investors need to have accurate analysis and information to gain significant profits as stock prices fluctuate due to company performance, industry factors, changes in interest rates, liquidity, global market conditions, market sentiment, and investor psychology. The Gated Recurrent Unit algorithm is suitable for application on historical stock data sets because they are time series data, can be computed and compared on a numerical scale. This algorithm is a variant of the Long Short-Term Memory algorithm or other types of processing modules for Recurrent Neural Networks. The data set used consists of closing price data or close features, comprising a training data set of 4,406 data and a test data set of 1,889 data that have undergone data preparation using various techniques, including data cleansing, data scrubbing, data splitting, data normalization, and data reshaping. The results showed that the Gated Recurrent Unit algorithm is the right strategy because it obtains a good evaluation of model performance, namely MSE of 0.0009; RMSE of 0.17325 and MAE of 0.0207.

Construction and improvement of English vocabulary learning model integrating spiking neural network and convolutional long short-term memory algorithm.

To help non-native English speakers quickly master English vocabulary, and improve reading, writing, listening and speaking skills, and communication skills, this study designs, constructs, and improves an English vocabulary learning model that integrates Spiking Neural Network (SNN) and Convolutional Long Short-Term Memory (Conv LSTM) algorithms. The fusion of SNN and Conv LSTM algorithm can fully utilize the advantages of SNN in processing temporal information and Conv LSTM in sequence data modeling, and implement a fusion model that performs well in English vocabulary learning. By adding information transfer and interaction modules, the feature learning and the timing information processing are optimized to improve the vocabulary learning ability of the model in different text contents. The training set used in this study is an open data set from the WordNet and Oxford English Corpus data corpora. The model is presented as a computer program and applied to an English learning application program, an online vocabulary learning platform, or a language education software. The experiment will use the open data set to generate a test set with text volume ranging from 100 to 4000. The performance indicators of the proposed fusion model are compared with those of five traditional models and applied to the latest vocabulary exercises. From the perspective of learners, 10 kinds of model accuracy, loss, polysemy processing accuracy, training time, syntactic structure capturing accuracy, vocabulary coverage, F1-score, context understanding accuracy, word sense disambiguation accuracy, and word order relation processing accuracy are considered. The experimental results reveal that the performance of the fusion model is better under different text sizes. In the range of 100-400 text volume, the accuracy is 0.75-0.77, the loss is less than 0.45, the F1-score is greater than 0.75, the training time is within 300s, and the other performance indicators are more than 65%; In the range of 500-1000 text volume, the accuracy is 0.81-0.83, the loss is not more than 0.40, the F1-score is not less than 0.78, the training time is within 400s, and the other performance indicators are above 70%; In the range of 1500-3000 text volume, the accuracy is 0.82-0.84, the loss is less than 0.28, the F1-score is not less than 0.78, the training time is within 600s, and the remaining performance indicators are higher than 70%. The fusion model can adapt to various types of questions in practical application. After the evaluation of professional teachers, the average scores of the choice, filling-in-the-blank, spelling, matching, exercises, and synonyms are 85.72, 89.45, 80.31, 92.15, 87.62, and 78.94, which are much higher than other traditional models. This shows that as text volume increases, the performance of the fusion model is gradually improved, indicating higher accuracy and lower loss. At the same time, in practical application, the fusion model proposed in this study has a good effect on English learning tasks and offers greater benefits for people unfamiliar with English vocabulary structure, grammar, and question types. This study aims to provide efficient and accurate natural language processing tools to help non-native English speakers understand and apply language more easily, and improve English vocabulary learning and comprehension.