Long Short-term Memory Recurrent Neural Network Research Articles

In-Situ early anomaly detection and remaining useful lifetime prediction for high-power white LEDs with distance and entropy-based long short-term memory recurrent neural networks

High-power white light-emitting diodes (LEDs) have demonstrated superior efficiency and reliability compared to traditional white light sources. However, ensuring maximum performance for a prolonged lifetime use presents a significant challenge for manufacturers and end users, especially in safety–critical applications. Thus, identifying functional anomalies and predicting the remaining useful lifetime (RUL) is of enormous importance in the operational longevity of the device. To address such challenges, this study proposes a combination of distance-based Mahalanobis distance (MD), entropy generation rate (EGR), and deep learning models for improved anomaly detection and RUL prediction accuracy. Unlike conventional health indicators based on luminous flux data that are challenging to monitor relevant optical performance, the MD and EGR methods are employed to extract in-situ monitored thermal and electrical data as new health indicators. Long short-term memory recurrent neural networks (LSTM-RNN) and convolutional neural networks (CNN) are established to detect anomalies and predict the RUL. The accelerated degradation tests of 3 W high-power white LED have been conducted, and the online and offline collected experimental data are deployed for model development and performance evaluation. The performance of the proposed methods is compared against the Illuminating Engineering Society of North America (IESNA) TM-21 method. The results indicate that LSTM-RNN, when combined with either MD or EGR, can detect anomalies with significantly fewer data (70 %) than is typically required. Furthermore, a significant improvement in prediction accuracy in RUL prediction based on MD and EGR-constructed time series health indicators and employed with the LSTM-RNN model demonstrates the effectiveness of the proposed methods.

An LSTM RNN proposal for surrogate modeling the dynamic response of buried structures to earthquake plane waves in soil half-spaces

This paper presents a variant of the long short-term memory (LSTM) recurrent neural network for surrogate modeling and predicting the seismic response of buried structures to in-plane inclined shear waves in heterogeneous soil profiles. The proposed model is designed to accept scalar and time-dependent random variables as input features to facilitate Monte Carlo (MC) based uncertainty quantification tasks where both the earthquake wave field and system properties are sources of uncertainty. To this end, it employs short-duration synthetic time histories for training instead of actual earthquake records to ensure efficiency. We discuss several configurations to evaluate the performance of the proposed methodology in predicting the tunnel response at different locations. We use two application examples concerning forward uncertainty quantification and sensitivity analysis to assess further the efficacy and accuracy of the proposed surrogate modeling approach.

Taxi Demand and Fare Prediction with Hybrid Models: Enhancing Efficiency and User Experience in City Transportation

An essential part of a city’s transportation infrastructure, taxis allow for regular encounters between drivers and customers. Nevertheless, there are issues with efficiency since there is an imbalance in the supply and demand for taxis. This study describes the creation of a platform that serves both customers and taxi drivers by offering immediate forecasts of demand and fare. Root mean squared error (RMSE) of 3.31 and a negative log-likelihood of −3.84, the long short-term memory recurrent neural network (LSTM-RNN) with the mixture density network (MDN) is employed to forecast taxi demand. The best RMSE of 3.24 is obtained for fare prediction via an ensemble learning model that integrates linear regression (LR), ridge regression (RR), and multilayer perceptron (MLP). To ensure peak performance, the models are systematically created, implemented, trained, and improved. By integrating these models into a web application interface, the taxi service system offers a better overall user experience, which improves urban mobility.

Forest Sound Event Detection with Convolutional Recurrent Neural Network-Long Short-Term Memory

Sound event detection tackles an audio environment's complex sound analysis and recognition problem. The process involves localizing and classifying sounds mainly to estimate the start point and end points of the separate sounds and describe each sound. Sound event detection capability relies on the type of sound. Although detecting sequences of distinct temporal sounds is straightforward, the situation becomes complex when the sound is multiple overlapping of much single audio. This situation usually occurs in the forest environment. Therefore, this aim of the paper is to propose a Convolution Recurrent Neural Network-Long Short-Term Memory algorithm to detect an audio signature of intruders in the forest environment. The audio is extracted in the Mel-frequency cepstrum coefficient and fed into the algorithm as an input. Six sound categories are chainsaw, machete, car, hatchet, ambiance, and bike. They were tested using several epochs, batch size, and filter of the layer in the model. The proposed model can achieve an accuracy of 98.52 percent in detecting the audio signature with a suitable parameter selection. In the future, additional types of audio signatures of intruders and further aspects of evaluation can be added to make the algorithm better at detecting intruders in the forest environment.

Prediksi Saham Menggunakan Recurrent Neural Network (RNN-LSTM) dengan Optimasi Adaptive Moment Estimation

Predicting stock price movements is a complex challenge in the financial market due to unpredictable price fluctuations and high sensitivity levels. Noise in historical stock price data and temporal dependencies between previous and current prices make recognizing price movement patterns difficult. In a dynamic market environment, the model's ability to generate accurate predictions holds significant implications for more informed investment decision-making. The Recurrent Neural Network - Long Short-Term Memory (RNN-LSTM) model holds great potential for stock price prediction. It captures temporal dependencies, identifies non-linear relationships, and deciphers complex trends in stock price data. This study employs deep learning techniques with the RNN-LSTM model optimized using Adaptive Moment Estimation (Adam) to enhance stock price prediction accuracy by leveraging historical stock price data and technical factors. Data preprocessing, including handling missing values and data normalization, aids the model in navigating the dataset's intricacies. Test results utilizing the Mean Squared Error (MSE) metric reveal the model's ability to produce predictions that closely resemble actual stock prices, with a low loss value of 0109012. The model also exhibits good predictive accuracy, as evidenced by a favorable Mean Percentage Error (MPE) score of 1.74% between predicted and actual values. These findings hold valuable implications for assisting investors and financial practitioners in managing complexity and uncertainty within the stock market

Textual emotion analysis using improved metaheuristics with deep learning model for intelligent systems

AbstractComplex, intelligent systems, namely social network services (SNS), internet‐of‐things (IoT), augmented reality (AR), virtual reality (VR), and so on have become a crucial part of our day‐to‐day life. The dynamics of this system alter human behaviors in various forms. Therefore, it is important to understand the fundamental dynamics of the system and what it involves for user sentiment. Emotion analysis is a commonly used and very powerful analysis methodology in data mining. It gives an outstanding option to evaluate, determine, understand, and monitor the sentiments of consumers with respect to a service or a product. This article focuses on the design of textual emotion analysis using the improved metaheuristics with deep learning (TEA‐IMWDL) model for intellectual systems. The goal of the TEA‐IMWDL technique lies in the identification and classification of emotions that exist in social media content. To attain this, the presented TEA‐IMWDL technique undergoes preprocessing to make the input data compatible with the latter stages of processing. For the word embedding process, the Fast text approach is utilized in this study. Next, the long short term memory‐recurrent neural network (LSTM‐RNN) method can be enforced for the emotion classification procedure. To boost the classification efficiency of the LSTM‐RNN model, an improved gravitational search optimization algorithm (IGSA) based hyperparameter tuning process is involved in this study. A brief set of simulation analyses of the TEA‐IMWDL technique is carried out and the results are tested on a benchmark dataset. The experimental outcomes demonstrate the enhanced performance of the TEA‐IMWDL technique over other recent approaches.

Overcoming Nonlinear Dynamics in Diabetic Retinopathy Classification: A Robust AI-Based Model with Chaotic Swarm Intelligence Optimization and Recurrent Long Short-Term Memory

Diabetic retinopathy (DR), which is seen in approximately one-third of diabetes patients worldwide, leads to irreversible vision loss and even blindness if not diagnosed and treated in time. It is vital to limit the progression of DR disease in order to prevent the loss of vision in diabetic patients. It is therefore essential that DR disease is diagnosed at an early phase. Thanks to retinal screening at least twice a year, DR disease can be diagnosed in its early phases. However, due to the variations and complexity of DR, it is really difficult to determine the phase of DR disease in current clinical diagnoses. This paper presents a robust artificial intelligence (AI)-based model that can overcome nonlinear dynamics with low computational complexity and high classification accuracy using fundus images to determine the phase of DR disease. The proposed model consists of four stages, excluding the preprocessing stage. In the preprocessing stage, fractal analysis is performed to reveal the presence of chaos in the dataset consisting of 12,500 color fundus images. In the first stage, two-dimensional stationary wavelet transform (2D-SWT) is applied to the dataset consisting of color fundus images in order to prevent information loss in the images and to reveal their characteristic features. In the second stage, 96 features are extracted by applying statistical- and entropy-based feature functions to approximate, horizontal, vertical, and diagonal matrices of 2D-SWT. In the third stage, the features that keep the classifier performance high are selected by a chaotic-based wrapper approach consisting of the k-nearest neighbor (kNN) and chaotic particle swarm optimization algorithms (CPSO) to cope with both chaoticity and computational complexity in the fundus images. At the last stage, an AI-based classification model is created with the recurrent neural network-long short-term memory (RNN-LSTM) architecture by selecting the lowest number of feature sets that can keep the classification performance high. The performance of the DR disease classification model was tested on 2500 color fundus image data, which included five classes: no DR, mild non-proliferative DR (NPDR), moderate NPDR, severe NPDR, and proliferative DR (PDR). The robustness of the DR disease classification model was confirmed by the 10-fold cross-validation. In addition, the classification performance of the proposed model is compared with the support vector machine (SVM), which is one of the machine learning techniques. The results obtained show that the proposed model can overcome nonlinear dynamics in color fundus images with low computational complexity and is very effective and successful in precisely diagnosing all phases of DR disease.

Password strength verification based on machine learning algorithms and LSTM recurrent neural networks

Objectives. One of the most commonly used authentication methods in computer systems, password authentication is susceptible to various attacks including brute-force and dictionary attacks. This susceptibility requires not only the strict protection of user credentials, but also the definition of criteria for increasing a password’s strength to minimize the possibility of its exploitation by an attacker. Thus, an important task is the development of a verifier for checking passwords for strength and prohibiting the user from setting passwords that are susceptible to cracking. The use of machine learning methods to construct a verifier involves algorithms for formulating requirements for password complexity based on lists of known passwords available for each strength category.Methods. The proposed supervised machine learning algorithms comprise support vector machines, random forest, boosting, and long short-term memory (LSTM) recurrent neural network types. Embedding and term frequency–inverse document frequency (TF-IDF) methods are used for data preprocessing, while cross-validation is used for selecting hyperparameters.Results. Password strength recommendations and requirements from international and Russian standards are described. The existing methods of password strength verification in various operating systems are analyzed. The experimental results based on existing datasets comprising passwords having an associated level of strength are presented.Conclusions. A LSTM recurrent neural network is highlighted as one of the most promising areas for building a password strength verifier.

Lithium battery life prediction based on deep learning

Li-ion batteries have the advantages of high efficiency, high energy density and long life, having developed rapidly in recent years. However, it is tough to accurately predict the decline trend of Li-ion battery capacity, which limits the further improvement of their service life and safety. In this study, a method of using wavelet denoising to preprocess the data and predicting the life of Li-ion battery based on whale optimization algorithm combined with long short-term memory network (WOA-LSTM) is proposed. In this paper, two sets of data sets B0005 and B0006 of NASA 's public data set are used. The original battery capacity data is subjected to wavelet transform and noise reduction to remove noise and redundant information. The calculation results of SNR and RMSE are 48.1119 and 0.006225, respectively. Then LSTM (Long short-term memory), RNN (Recurrent Neural Network), GRU (Gated Recurrent Unit) and WOA-LSTM are used to predict the remaining useful life of the data set RUL (Remaining Useful Lifetime, RUL), and the data error is compared, showing that in the results of MAE, RMSE and MAPE three prediction error indicators. The prediction results of WOA-LSTM in B0005 and B0006 show the minimum prediction errors, which are 0.0563,0.0710,0.0415 in B0005 data set and 0.0583,0.0831,0.0454 in B0006 data set. Compared with the standard LSTM model, RNN model and GRU model, the error indexes of the model are decreased, which are 7 %, 4 % and 3 % respectively, which has great advantages. This method can provide a reliable predictive analysis method for battery design and fault diagnosis.

Comprehensive early warning method of microseismic, acoustic emission, and electromagnetic radiation signals of rock burst based on deep learning

Microseismic, acoustic emission, and electromagnetic radiation monitoring methods are often used to monitor rock burst disasters in coal mines. In the process of coal mining, the time series characteristics and amplitude characteristics of microseismic, acoustic emission, and electromagnetic radiation data are mainly used to identify rockburst risk, but the results of risk identification through the three monitoring methods are quite different. Consequently, the accurate and comprehensive early warning of rock burst risk is still an urgent problem to be solved. The development of deep learning provides a new means for intelligent early warning of rock burst risk. In this paper, a comprehensive early warning method of microseismic, acoustic emission, and electromagnetic radiation (MS-AE-EMR) signals of rock bursts was proposed based on a deep learning algorithm. This method uses long short-term memory recurrent neural networks (LSTM-RNNs) to intelligently identify the MS-AE-EMR precursor signal of rock burst risk, predicts the MS-AE-EMR signal by a convolution neural network (CNN), analyses the MS-AE-EMR precursor signal of rock burst risk through the data analysis method and obtains the risk coefficient of rock burst. Moreover, by using the MS-AE-EMR original signal and risk coefficient, it trains the multi-input CNN and inputs the predicted signal into the trained multi-input CNN to obtain the predicted risk coefficient of rock burst. Analysing the risk coefficient completes the comprehensive early warning of the MS-AE-EMR signal of rock burst. After field verification, the RNN-based comprehensive early warning method of the MS-AE-EMR signal can respond positively to rock burst risk and capture the information in advance. Therefore, this method is of great significance for accurate monitoring and early warning of rock burst in coal mines.

Imagined Speech Classification Using EEG and Deep Learning.

In this paper, we propose an imagined speech-based brain wave pattern recognition using deep learning. Multiple features were extracted concurrently from eight-channel electroencephalography (EEG) signals. To obtain classifiable EEG data with fewer sensors, we placed the EEG sensors on carefully selected spots on the scalp. To decrease the dimensions and complexity of the EEG dataset and to avoid overfitting during the deep learning algorithm, we utilized the wavelet scattering transformation. A low-cost 8-channel EEG headset was used with MATLAB 2023a to acquire the EEG data. The long-short term memory recurrent neural network (LSTM-RNN) was used to decode the identified EEG signals into four audio commands: up, down, left, and right. Wavelet scattering transformation was applied to extract the most stable features by passing the EEG dataset through a series of filtration processes. Filtration was implemented for each individual command in the EEG datasets. The proposed imagined speech-based brain wave pattern recognition approach achieved a 92.50% overall classification accuracy. This accuracy is promising for designing a trustworthy imagined speech-based brain-computer interface (BCI) future real-time systems. For better evaluation of the classification performance, other metrics were considered, and we obtained 92.74%, 92.50%, and 92.62% for precision, recall, and F1-score, respectively.

Projection of morbidity and mortality due to breast cancer between 2020 and 2050 across 42 low- and middle-income countries

ObjectiveAim of the study is to predict trends in morbidity and mortality due to breast cancer in 42 LMICs between 2020 and 2050. Designand Setting: National level cross-sectional breast cancer related data between 1990 and 2019 were used. Recurrent Neural Network, Long-Short-Term Memory (RNN-LSTM) model was employed to forecast the trend in breast cancer burden. Main outcomes and measuresAge standardized breast cancer incidence, mortality, and disability adjusted life years (DALYs) rates. ResultsBy 2050, the age standardized breast cancer incidence rate is expected to increase in 38 LMICs with highest incidence rate in Namibia; 127.0 (78.0–176.0) followed by Nigeria 71.1 (53.9–88.3) and Papua New Guinea 70.6 (88.7–74.6). Similarly, the age standardized breast cancer mortality and DALYs rates in 2050 are expected to increase in 33 and 35 LMICs respectively. The highest mortality and DALYs rates in 2050 are expected to be 64.7 (42.6–86.7) in Namibia and 1543.6 (1463.1–1624.1) in Pakistan. The estimated annual percentage change (EAPC) is expected to increase uniformly in all the countries during the same period. Due to considerable variation in exposure risk, such as high plasma glucose level, high body mass index (BMI) and socio-demographic index (SDI), high regional disparity in burden of breast cancer is expected among the countries. Conclusionand Relevance: Breast cancer burden is expected to increase in most of the LMICs with high regional disparity by 2050. Our study's finding focuses on LMICs with high breast cancer burden that require tailored strategies and effective action plans to ensure prevention from catastrophic consequences in the future. Minimizing the exposure to behavioral and metabolic risk factors such as high plasma glucose, high BMI, along with tackling the issue of low fertility rate would be important in managing breast cancer burden in LMICs.

Sentiment analysis of amazon user reviews using a hybrid approach

Consumer reviews of services and products are a critical performance indicator for businesses looking to improve their contributions. They're also important for future customers to understand previous customers' experiences. Sentimental Analysis (SA) identifies patterns of sentiment in reviews by analyzing and extracting opinions from websites, papers, blogs, and other sources. The customer finds it difficult to locate a review for a specific aspect of a product they intend to purchase. Since it is difficult to go through a huge number of reviews, feature selection is used to reduce the data's dimensionality. Because choosing an appropriate feature can identify the product attributes that are disliked or discussed by consumers, selecting a feature is a significant step in sentiment analysis. In this paper, a Pearson correlation coefficient-based Harris Hawks Optimization – based Recurrent Neural Network-Long Short-Term Memory (PCCHHO-RNN-LSTM) algorithm is proposed to select features from reviews given by users for classifying their sentiments according to their appropriate polarity. In this proposed PCCHHO-RNNLSTM, to achieve greater accuracy, the correlation coefficients of features are used to conduct an initial dimensionality reduction. The HHO algorithm is then used to choose a small group of non-redundant features and RNN-LSTM is used to classify sentiments to their suitable polarity. The proposed method has been evaluated using the data gathered from Amazon. com's online product reviews. MATLAB software is used to implement the proposed work. From experimental results, the proposed PCCHH-RNNLSTM demonstrates considerable improvements of 95.8% accuracy, 95.4% precision, 95.6% recall, and 95.2% F-measure, respectively.

Predict the priority of end-users’ maintenance requests and the required technical staff through LSTM and Bi-LSTM recurrent neural networks

PurposeThis paper aims to develop predictive methods, based on recurrent neural networks, useful to support facility managers in building maintenance tasks, by collecting information coming from a computerized maintenance management system (CMMS).Design/methodology/approachThis study applies data-driven and text-mining approaches to a CMMS data set comprising more than 14,500 end-users’ requests for corrective maintenance actions, collected over 14 months. Unidirectional long short-term memory (LSTM) and bidirectional LSTM (Bi-LSTM) recurrent neural networks are trained to predict the priority of each maintenance request and the related technical staff assignment. The data set is also used to depict an overview of corrective maintenance needs and related performances and to verify the most relevant elements in the building and how the current facility management (FM) relates to the requests.FindingsThe study shows that LSTM and Bi-LSTM recurrent neural networks can properly recognize the words contained in the requests, thus correctly and automatically assigning the priority and predicting the technical staff to assign for each end-user’s maintenance request. The obtained global accuracy is very high, reaching 93.3% for priority identification and 96.7% for technical staff assignment. Results also show the main critical building elements for maintenance requests and the related intervention timings.Research limitations/implicationsThis work shows that LSTM and Bi-LSTM recurrent neural networks can automate the assignment process of end-users’ maintenance requests if trained with historical CMMS data. Results are promising; however, the trained LSTM and Bi-LSTM RNN can be applied only to different hospitals adopting similar categorization.Practical implicationsThe data-driven and text-mining approaches can be integrated into the CMMS to support corrective maintenance management by facilities management contractors, i.e. to properly and timely identify the actions to be carried out and the technical staff to assign.Social implicationsThe improvement of the maintenance of the health-care system is a key component of improving health service delivery. This work shows how to reduce health-care service interruptions due to maintenance needs through machine learning methods.Originality/valueThis study develops original methods and tools easily integrable into IT workflow systems (i.e. CMMS) in the FM field.

Forecasting atmospheric turbulence conditions from prior environmental parameters using artificial neural networks.

Atmospheric turbulence (C n2) modeling has been proposed by physics-based models, but they are unable to capture the many cases. Recently, machine learning surrogate models have been used to learn the relationship between local meteorological conditions and turbulence strength. These models predict C n2 at time t from weather at time t. This work expands modeling capabilities by proposing a technique to forecast 3h of future turbulence conditions at 30min intervals from prior environmental parameters using artificial neural networks. First, local weather and turbulence measurements are formatted to pairs of the input sequence and output forecast. Next, a grid search is used to find the best combination of model architecture, input variables, and training parameters. The architectures investigated are the multilayer perceptron and three variants of the recurrent neural network (RNN): the simple RNN, the long short-term memory RNN (LSTM-RNN), and the gated recurrent unit RNN (GRU-RNN). A GRU-RNN architecture that uses 12h of prior inputs is found to have the best performance. Finally, this model is applied to the test dataset and analyzed. It is shown that the model has generally learned the relationship between prior environmental and future turbulence conditions.

Unmanned aerial vehicles assisted rice seedling detection using shark smell optimization with deep learning model

Rice seedling classification using an unmanned aerial vehicle (UAV) images remains a challenging problem that needs to be addressed. It is still a difficult task because it is prone to low temporal and spatial resolution images. Recently, machine learning (ML) and deep learning (DL) models can be employed for several image preprocessing tasks such as classification, object detection, and segmentation. Therefore, this study focuses on the design of shark smell optimization with deep learning based rice seedling detection (SSODL-RSD) on UAV imagery. The presented SSODL-RSD technique recognizes the UAV images into arable land and rice seedlings. To achieve this, the SSODL-RSD technique employs the adaptive Wiener filtering (AWF) technique for the noise removal procedure. In addition, the SSODL-RSD technique exploits the NestNet feature extractor model. Moreover, the SSO algorithm is used for the hyperparameter tuning of the NestNet model. Finally, the long short term memory-recurrent neural network (LSTM-RNN) model is employed for the classification of rice seedlings. The extensive comparative study highlighted the improved outcomes of the SSODL-RSD technique over other existing models.

MFPD-LSTM: A steganalysis method based on multiple features of pitch delay using RNN-LSTM

In recent years, adaptive multi-rate (AMR) steganalysis based on pitch delay has become a hot research area for researchers. The main reason is that the steganalysis algorithm based on AMR can sensitively detect the change in the statistical characteristics of the pitch delay before and after the secret information is embedded, and achieve high accuracy. At present, many excellent AMR steganalysis algorithms based on pitch delay have been proposed. However, when these algorithms are aimed at short-term, low-embedding-rate speech samples, there is room for improvement in accuracy and performance. In this paper, a Recurrent Neural Network-Long-Short Term Memory (RNN-LSTM) is used to design a steganalysis method through multi-feature fusion based on pitch delay (MF-PD). This method extracts the pitch delay sequence in the speech information stream, mines four features that characterize the statistical feature of the intra-frame and the inter-frame, and establishes the feature matrix. Then construct the RNN-LSTM model, the dropout layer is introduced to avoid over-fitting and the model is pruned to improve the efficiency. Experimental results show that when the speech sample length is 0.1s, this method can achieve a detection accuracy of more than 87%, which is significantly higher than other steganalysis algorithms. For samples with low embedding rate, the method proposed in this paper can also achieve better performance, which satisfies the need for short time and low embedding rate sample detection.

Long Short-Term Memory Recurrent Neural Network Approach for Approximating Roots (Eigen Values) of Transcendental Equation of Cantilever Beam

In this study, the natural frequencies and roots (Eigenvalues) of the transcendental equation in a cantilever steel beam for transverse vibration with clamped free (CF) boundary conditions are estimated using a long short-term memory-recurrent neural network (LSTM-RNN) approach. The finite element method (FEM) package ANSYS is used for dynamic analysis and, with the aid of simulated results, the Euler–Bernoulli beam theory is adopted for the generation of sample datasets. Then, a deep neural network (DNN)-based LSTM-RNN technique is implemented to approximate the roots of the transcendental equation. Datasets are mainly based on the cantilever beam geometry characteristics used for training and testing the proposed LSTM-RNN network. Furthermore, an algorithm using MATLAB platform for numerical solutions is used to cross-validate the dataset results. The network performance is evaluated using the mean square error (MSE) and mean absolute error (MAE). Finally, the numerical and simulated results are compared using the LSTM-RNN methodology to demonstrate the network validity.

HH model based smart deep brain stimulator to detect, predict and control epilepsy using machine learning algorithm

BackgroundEpilepsy is the most common neurological disorder in the world. To control epilepsy, deep brain stimulation is one of the widely accepted treatment techniques. However, conventional deep brain stimulation technique provides continuous stimulation without optimizing the stimulation parameters, resulting in adverse side effects and unexpected death. Hence, understanding the dynamic behavior of brain neural networks at a cellular level is required for patient-specific epilepsy treatment. Considering the underlying mechanism of a single neuronal shift in the brain neural network, computational model-based techniques have a new face for healthcare, which aims to develop effective medical devices for preclinical investigations. New methodThis paper discusses the design of a Smart Deep Brain Stimulator (SDBS) using the Hodgkin-Huxley (HH) conductance-based cellular model of brain neurons to automatically detect, predict and regulate epilepsy against patient-specific conditions. Epileptic activity is simulated as a spike train of action potential due to sodium and potassium channel conductance variations in the single-neuron HH model. The proposed SDBS consists of three components:- i) seizure detection using bagging and boosting-based ensemble machine learning classifiers, ii) channel conductance prediction using Long Short Term Memory-Recurrent Neural Network (LSTM-RNN) based Deep Neural Network (DNN) for updating model parameters of brain neuron, and iii) model-based intelligent control of epileptic seizure with Nonlinear Autoregressive Moving Average-L2 (NARMA-L2) Controller and Nonlinear Model Predictive Controller (NMPC). ResultsFor effective treatment, improving the overall accuracy and efficiency of SDBS is essential. For epilepsy detection, the ensemble bagging machine learning algorithm provides better accuracy of 92.7% compared to the ensemble boosting algorithm. LSTM-RNN deep neural network model with four layers predicts the variations in channel conductance with Root Mean Square Error (RMSE) of 0.00568 and 0.009081 for sodium and potassium channel conductance, respectively. From the closed-loop performances of SDBS with an intelligent control scheme, it is observed that SDBS with NMPC provides efficient and accurate stimulation with minimum energy consumption. From a stability point of view, SDBS with NMPC provides better stability than SDBS with NARMA-L2 Controller. Comparison with existing methodThe proposed SDBS is designed to generate accurate stimulation pulses for epilepsy patients with specific conditions depending on the neuronal activity of a single neuron. Moreover, it will also adapt to the dynamic condition of epilepsy patients. The existing deep brain stimulator continuously provides stimulation pulses without adapting to the patient’s conditions. ConclusionThe proposed SDBS could provide patient-specific treatment based on sodium/potassium channel conductance variations of brain neurons. It will help increase the use of deep brain stimulation techniques and reduce sudden death. Furthermore, the proposed technique will be extended to neural network models with larger neuronal populations to improve the practical feasibility.

Development of a Soil Moisture Prediction Model Based on Recurrent Neural Network Long Short-Term Memory (RNN-LSTM) in Soybean Cultivation.

Due to climate change, soil moisture may increase, and outflows could become more frequent, which will have a considerable impact on crop growth. Crops are affected by soil moisture; thus, soil moisture prediction is necessary for irrigating at an appropriate time according to weather changes. Therefore, the aim of this study is to develop a future soil moisture (SM) prediction model to determine whether to conduct irrigation according to changes in soil moisture due to weather conditions. Sensors were used to measure soil moisture and soil temperature at a depth of 10 cm, 20 cm, and 30 cm from the topsoil. The combination of optimal variables was investigated using soil moisture and soil temperature at depths between 10 cm and 30 cm and weather data as input variables. The recurrent neural network long short-term memory (RNN-LSTM) models for predicting SM was developed using time series data. The loss and the coefficient of determination (R2) values were used as indicators for evaluating the model performance and two verification datasets were used to test various conditions. The best model performance for 10 cm depth was an R2 of 0.999, a loss of 0.022, and a validation loss of 0.105, and the best results for 20 cm and 30 cm depths were an R2 of 0.999, a loss of 0.016, and a validation loss of 0.098 and an R2 of 0.956, a loss of 0.057, and a validation loss of 2.883, respectively. The RNN-LSTM model was used to confirm the SM predictability in soybean arable land and could be applied to supply the appropriate moisture needed for crop growth. The results of this study show that a soil moisture prediction model based on time-series weather data can help determine the appropriate amount of irrigation required for crop cultivation.