Short-term Memory Performance Research Articles

Enhancing Arabic Dialect Detection on Social Media: A Hybrid Model with an Attention Mechanism

Recently, the widespread use of social media and easy access to the Internet have brought about a significant transformation in the type of textual data available on the Web. This change is particularly evident in Arabic language usage, as the growing number of users from diverse domains has led to a considerable influx of Arabic text in various dialects, each characterized by differences in morphology, syntax, vocabulary, and pronunciation. Consequently, researchers in language recognition and natural language processing have become increasingly interested in identifying Arabic dialects. Numerous methods have been proposed to recognize this informal data, owing to its crucial implications for several applications, such as sentiment analysis, topic modeling, text summarization, and machine translation. However, Arabic dialect identification is a significant challenge due to the vast diversity of the Arabic language in its dialects. This study introduces a novel hybrid machine and deep learning model, incorporating an attention mechanism for detecting and classifying Arabic dialects. Several experiments were conducted using a novel dataset that collected information from user-generated comments from Twitter of Arabic dialects, namely, Egyptian, Gulf, Jordanian, and Yemeni, to evaluate the effectiveness of the proposed model. The dataset comprises 34,905 rows extracted from Twitter, representing an unbalanced data distribution. The data annotation was performed by native speakers proficient in each dialect. The results demonstrate that the proposed model outperforms the performance of long short-term memory, bidirectional long short-term memory, and logistic regression models in dialect classification using different word representations as follows: term frequency-inverse document frequency, Word2Vec, and global vector for word representation.

Traffic crash prediction model in Kano State, Nigeria: a multivariate LSTM approach

Accurate traffic crash prediction is crucial for implementing effective road safety measures. This study compares the performance of Long Short-Term Memory (LSTM) and Multivariate LSTM (MLSTM) models in forecasting total crash count data in Kano State, Nigeria. Human and vehicle factors, including speed violation, tire burst, brake failure, sign light violation, and phone use while driving, are incorporated as covariates in the MLSTM model. An ARIMAX model is employed to investigate the effects of the covariates. The MLSTM model outperforms both the basic LSTM model and individual covariate models, emphasizing the synergistic effect of considering a broad range of factors. The ARIMAX model results reveal that speed violation is significantly positively correlated with total crashes, while other covariates show positive correlations but do not reach the statistical significance. The findings underscore the importance of a multivariate approach in enhancing traffic crash prediction. The MLSTM model's superior performance highlights the value of considering a comprehensive range of factors that influence crash occurrence to achieve more accurate predictions. Practical applications of these models could involve leveraging them for proactive traffic safety measures, which include increased enforcement of traffic rules, targeted driver education and campaigns, and improvements to road infrastructure.

Toward better atmospheric polycyclic aromatic hydrocarbons pollution control in the Northern Hemisphere: Process analysis based on interpretable deep learning models

The accurate prediction of atmospheric polycyclic aromatic hydrocarbons (PAHs) concentrations in the Arctic is crucial in guiding the development of pollution control measures for atmospheric PAHs. However, the complex formation process and strong time-dependence of PAHs in the Arctic increase the difficulty of prediction. Herein, deep learning (DL) models were developed to predict monthly air concentrations of 16 priority PAHs at Alert monitoring station by incorporating the inherent periodicity of PAH concentrations and ancillary data including PAH emissions, meteorological information, fire emissions, and sea ice area. The prediction performance of long short-term memory (LSTM), gated recurrent unit (GRU) and bidirectional LSTM (Bi LSTM) algorithms were compared by multiple model evaluation metrics. The results showed that Bi LSTM models outperformed LSTM and GRU in predicting PAHs concentrations because it exhibited higher accuracy in terms of average R2ext (0.09) and errors (average MAPEext = 1.45). The DL interpretation based on SHAP values suggested that the main drivers of PAHs concentrations were PAHs emissions and sea ice area, with a more prominent contribution from meteorological conditions. The interpretable deep learning approach provides a potential shortcut for predicting time-delayed PAHs concentrations in the Arctic, promoting the management and control of global PAHs pollution.

When remembering less is more: Unfiltered items are associated with reduced memory fidelity in visual short-term memory.

Visual short-term memory (VSTM), the ability to store information no longer visible, is essential for human behavior. VSTM limits vary across the population and are correlated with overall cognitive ability. It has been proposed that low-memory individuals are unable to select only relevant items for storage and that these limitations are greatest when memory demands are high. However, it is unknown whether these effects simply reflect task difficulty and whether they impact the quality of memory representations. Here we varied the number of items presented, or set size, to investigate the effect of memory demands on the performance of visual short-term memory across low- and high-memory groups. Group differences emerged as set size exceeded memory limits, even when task difficulty was controlled. In a change-detection task, the low-memory group performed more poorly when set size exceeded their memory limits. We then predicted that low-memory individuals encoding items beyond measured memory limits would result in the degraded fidelity of memory representations. A continuous report task confirmed that low, but not high, memory individuals demonstrated decreased memory fidelity as set size exceeded measured memory limits. The current study demonstrates that items held in VSTM are stored distinctly across groups and task demands. These results link the ability to maintain high quality representations with overall cognitive ability.

Marrubium vulgare Extract Improves Spatial Working Memory and Oxidative Stress Damage in Scopolamine-Treated Rats.

The cholinergic neuronal loss in the basal forebrain and increasing brain oxidative stress are one of the main features of the brain suffering from Alzheimer's disease. Marrubium vulgare (M. vulgare), commonly known as 'white horehound,' possesses a variety of valuable properties, such as antioxidative, anti-inflammatory, and antidiabetic activities. Moreover, it possesses neuromodulatory properties that could potentially impact short-term memory functions. The present study was undertaken to investigate the preventive effects of water M. vulgare extract on working memory, cholinergic neurotransmission, and oxidative stress in rats with scopolamine (Sco)-induced dementia. Male Wistar rats (200-250 g) were divided into four experimental groups. The plant extract was administered orally for 21 days, and Sco (2 mg/kg) was administered intraperitoneally for 11 consecutive days. The behavioral performance of the animals was evaluated by the T-maze test. The effect of the extract on acetylcholinesterase (AChE) activity and antioxidant status in cortex and hippocampus were also monitored. Our experimental data revealed that treatment with M. vulgare significantly increased the percentage of correct choices of rats with Sco-induced dementia in the T maze test (by 38%, p < 0.05). Additionally, it reduced AChE activity in the hippocampus (by 20%, p < 0.05) and alleviated oxidative stress induced by Sco, particularly in the cortex. M. vulgare water extract demonstrated working memory preserving effect in rats with Sco-induced dementia, AChE inhibitory activity and in vivo antioxidant potential, and deserve further attention.

Attention and executive function impairments in obstructive sleep apnea are associated with decreased sleep spindles.

Sleep spindles play a key role in sleep-mediated cognitive processes. Cognitive functions are well-known to be affected in obstructive sleep apnea (OSA). Here, we analyzed attention and executive functions in patients with OSA and investigated the relationship between sleep spindles and cognitive abilities. Sixty patients with OSA (18-65years, 19 females and 41 males) and a control group (n = 41) including age-and sex-matched healthy individuals were consecutively and prospectively enrolled. All participants had a full-night polysomnography, and sleep spindles were analyzed using a semi-automated program. For the evaluation of short-term memory, attention and executive functions, Stroop test, forward and backward digit span tests were applied to all participants upon awakening following polysomnography. Scores of forward and backward digit span and Stroop tests were worse in OSA patients in compared to those in controls. Mean density of sleep spindles was decreased in OSA patients than those in controls (p = 0.044). A positive correlation was found between fast sleep spindle frequency and forward digit span (r = 2.222; p = 0.038) and backward digit span test scores (r = 2,157; p = 0.042) in OSA patients. In patients with moderate to severe OSA, sleep spindle density was positively correlated with forward (r = 2.323, p = 0.029) and backward (r = 2.500, p = 0.016) DSTs, and the duration of sleep spindles had positive correlation with backward DST (r = 2.452, p = 0.010). Our findings demonstrated that the disturbances in sleep spindle characteristics in OSA are associated with the cognitive impairments in attention, short-term memory, and executive functions, especially in patients with moderate to severe OSA.

IJCM_164A: Association between Sleep Quality and Short-Term memory among medical students in mangalore

Background: Sleep plays a crucial role in maintaining cognitive functions, including memory, which is essential for academic performance. However, there is limited research exploring the relationship between sleep quality and short-term memory among medical students. Methodology: Fourth-semester MBBS students from Kasturba Medical College, Mangalore, were selected using non- random convenience sampling. A total of 195 students participated, completing a standardized sleep quality assessment (Pittsburgh Sleep Quality Index) and a semi-structured self-administered questionnaire-based short-term memory test. Data were analyzed using SPSS version 16, with correlation tested using Karl Pearson’s Coefficient of Correlation Test. Results: The majority (56.4%) of participants exhibited good sleep quality, with 69.1% of these achieving high scores in short-term memory assessment tasks. A negative correlation (r = -0.450, p &lt; 0.001) was found between sleep quality and short-term memory scores. Students with good sleep quality were more likely to have higher short-term memory scores compared to those with poor sleep quality. Conclusion: The study suggests that good sleep quality is associated with better short-term memory performance among medical students. These findings underscore the importance of promoting good sleep habits to enhance cognitive function and academic performance in medical education.

An explainable long short-term memory network for surgical site infection identification

BackgroundCurrently, surgical site infection surveillance relies on labor-intensive manual chart review. Recently suggested solutions involve machine learning to identify surgical site infections directly from the medical record. Deep learning is a form of machine learning that has historically performed better than traditional methods while being harder to interpret. We propose a deep learning model, a long short-term memory network, for the identification of surgical site infection from the medical record with an attention layer for explainability. MethodsWe retrieved structured data and clinical notes from the University of Utah Health System’s electronic health care record for operative events randomly selected for manual chart review from January 2016 to June 2021. Surgical site infection occurring within 30 days of surgery was determined according to the National Surgical Quality Improvement Program definition. We trained the long short-term memory model along with traditional machine learning models for comparison. We calculated several performance metrics from a holdout test set and performed additional analyses to understand the performance of the long short-term memory, including an explainability analysis. ResultsSurgical site infection was present in 4.7% of the total 9,185 operative events. The area under the receiver operating characteristic curve and sensitivity of the long short-term memory was higher (area under the receiver operating characteristic curve: 0.954, sensitivity: 0.920) compared to the top traditional model (area under the receiver operating characteristic curve: 0.937, sensitivity: 0.736). The top 5 features of the long short-term memory included 2 procedure codes and 3 laboratory values. ConclusionSurgical site infection surveillance is vital for the reduction of surgical site infection rates. Our explainable long short-term memory achieved a comparable area under the receiver operating characteristic curve and greater sensitivity when compared to traditional machine learning methods. With explainable deep learning, automated surgical site infection surveillance could replace burdensome manual chart review processes.

Research on Fuzzy C- Clustering Recursive Genetic Algorithm based on Cloud Computing Bayes Function

Aiming at the problems of poor local search ability and precocious convergence of fuzzy C-cluster recursive genetic algorithm (FOLD++), a new fuzzy C-cluster recursive genetic algorithm based on Bayesian function adaptation search (TS) was proposed by incorporating the idea of Bayesian function adaptation search into fuzzy C-cluster recursive genetic algorithm. The new algorithm combines the advantages of FOLD++ and TS. In the early stage of optimization, fuzzy C-cluster recursive genetic algorithm is used to get a good initial value, and the individual extreme value pbest is put into Bayesian function adaptation table. In the late stage of optimization, when the searching ability of fuzzy C-cluster recursive genetic is weakened, the short term memory function of Bayesian function adaptation table in Bayesian function adaptation search algorithm is utilized. Make it jump out of the local optimal solution, and allow bad solutions to be accepted during the search. The improved algorithm is applied to function optimization, and the simulation results show that the calculation accuracy and stability of the algorithm are improved, and the effectiveness of the improved algorithm is verified.

Enhanced extended-input LSTM with an adaptive singular value decomposition UKF for LIB SOC estimation using full-cycle current rate and temperature data

Accurately estimating the state of charge (SOC) of lithium-ion batteries by the battery management system (BMS) is crucial for efficient energy management and power distribution control in electric vehicles (EVs). By far, data-driven methods have been extensively used in estimating the SOC of lithium-ion batteries to efficiently operate EVs with limited resources. However, the learning ability of these methods needs further enhancement. Therefore, this paper aims to analyze the SOC performance of an extended-input long short-term memory (ELSTM) model. The inputs of the ELSTM are improved with identified battery parameters based on an adaptive multi-timescale identification method with frequency difference decomposition. Second, an adaptive singular value decomposition-transformed unscented Kalman filter (ASVDUKF) is proposed to integrate all unknown variables and guarantee the positive definiteness of the error covariance matrix into a vector to form a multi-fusion model that robustly outputs the denoised and optimized SOCs based on the estimations of the ELSTM. The SOC results demonstrate that the ELSTM outperforms the LSTM, which uses conventional input data. Following that, the ELSTM-ASVDUKF model ensures high accuracy and stability with optimal mean absolute error, root-mean-square error, and mean absolute percentage error of 0.0939%, 0.1074%, and 0.1257%, respectively. The proposed model is validated using various full-cycle current rate and temperature data from two different batteries, establishing it as a promising solution for future development.

Role of GLR-1 in Age-Dependent Short-Term Memory Decline.

As the global elderly population grows, age-related cognitive decline is becoming an increasingly significant healthcare issue, often leading to various neuropsychiatric disorders. Among the many molecular players involved in memory, AMPA-type glutamate receptors are known to regulate learning and memory, but how their dynamics change with age and affect memory decline is not well understood. Here, we examined the in vivo properties of the AMPA-type glutamate receptor GLR-1 in the AVA interneuron of the Caenorhabditis elegans nervous system during physiological aging. We found that both total and membrane-bound GLR-1 receptor levels decrease with age in wild-type worms, regardless of their location along the axon. Using fluorescence recovery after photobleaching, we also demonstrated that a reduction in GLR-1 abundance correlates with decreased local, synaptic GLR-1 receptor dynamics. Importantly, we found that reduced GLR-1 levels strongly correlate with the age-related decline in short-term associative memory. Genetic manipulation of GLR-1 stability, by either deleting msi-1 or expressing a ubiquitination-defective GLR-1 (4KR) variant, prevented this age-related reduction in receptor abundance and improved the short-term memory performance in older animals, which reached performance levels similar to those of young animals. Overall, our data indicate that AMPA-type glutamate receptor abundance and dynamics are key factors in maintaining memory function and that changes in these parameters are linked to age-dependent short-term memory decline.

The impact of aerobic endurance training on cognitive performance in schizophrenic inpatients in a clinical routine setting.

The primary aim of this study was to investigate the impact of aerobic endurance training in schizophrenic inpatients on cognitive performance in a clinical routine setting. Of secondary interest was the influence on psychopathological symptoms. A total of 31 schizophrenic inpatients were randomly assigned to receive either controlled endurance training or occupational therapy. The experimental group underwent endurance training of 20-30min each, 3 times per week for a total of up to 22 training sessions. The control group received about 90min of occupational therapy, 2-3 times per week for up to 22 sessions. Cognitive performance was assessed via an extensive neuropsychological examination before randomization and prior to discharge. Significant improvements in cognitive functions and psychopathology could be shown in both groups. For verbal memory functions (short-term memory, working memory, and learning performance), there was a significant advantage for the aerobic endurance training group. Physical exercise is a feasible, easy-to-implement add-on therapy for schizophrenic patients in a clinical routine setting with positive effects on verbal memory functions. Besides, it seems important to fill the gap between inpatient and outpatient health care, providing physical training supply for this patient group.

Cross-frequency coupling in cortico-hippocampal networks supports the maintenance of sequential auditory information in short-term memory.

It has been suggested that cross-frequency coupling in cortico-hippocampal networks enables the maintenance of multiple visuo-spatial items in working memory. However, whether this mechanism acts as a global neural code for memory retention across sensory modalities remains to be demonstrated. Intracranial EEG data were recorded while drug-resistant patients with epilepsy performed a delayed matched-to-sample task with tone sequences. We manipulated task difficulty by varying the memory load and the duration of the silent retention period between the to-be-compared sequences. We show that the strength of theta-gamma phase amplitude coupling in the superior temporal sulcus, the inferior frontal gyrus, the inferior temporal gyrus, and the hippocampus (i) supports the short-term retention of auditory sequences; (ii) decodes correct and incorrect memory trials as revealed by machine learning analysis; and (iii) is positively correlated with individual short-term memory performance. Specifically, we show that successful task performance is associated with consistent phase coupling in these regions across participants, with gamma bursts restricted to specific theta phase ranges corresponding to higher levels of neural excitability. These findings highlight the role of cortico-hippocampal activity in auditory short-term memory and expand our knowledge about the role of cross-frequency coupling as a global biological mechanism for information processing, integration, and memory in the human brain.

Optimizing echo state networks for continuous gesture recognition in mobile devices: A comparative study

Continuous gesture recognition can be used to enhance human-computer interaction. This can be accomplished by capturing human movement with the use of the Inertial Measurement Units in smartphones and using machine learning algorithms to predict the intended gestures. Echo State Networks (ESNs) consist of a fixed internal reservoir that is able to generate rich and diverse nonlinear dynamics in response to input signals that capture temporal dependencies within the signal. This makes ESNs well-suited for time series prediction tasks, such as continuous gesture recognition. However, their application has not been rigorously explored, with regard to gesture recognition. In this study, we sought to enhance the efficacy of ESN models in continuous gesture recognition by exploring diverse model structures, fine-tuning hyperparameters, and experimenting with various training approaches. We used three different training schemes that used the Leave-one-out Cross-validation (LOOCV) protocol to investigate the performance in real-world scenarios with different levels of data availability: Leaving out data from one user to use for testing (F1-score: 0.89), leaving out a fraction of data from all users to use in testing (F1-score: 0.96), and training and testing using LOOCV on a single user (F1-score: 0.99). The obtained results outperformed the Long Short-Term Memory (LSTM) performance from past research (F1-score: 0.87) while maintaining a low training time of approximately 13 seconds compared to 63 seconds for the LSTM model. Additionally, we further explored the performance of the ESN models through behaviour space analysis using memory capacity, Kernel Rank, and Generalization Rank. Our results demonstrate that ESNs can be optimized to achieve high performance on gesture recognition in mobile devices on multiple levels of data availability. These findings highlight the practical ability of ESNs to enhance human-computer interaction.

Associations of habitual sedentary time with executive functioning and short-term memory in 7th and 8th grade adolescents

BackgroundWhile there is increasing evidence for negative physical health consequences of high volumes of sedentary time and prolonged sedentary time in adolescents, the association with cognition is less clear. This study investigated the association of volumes of habitual sedentary time and prolonged sedentary time with executive functions and short-term memory in adolescents.MethodsThis study has a cross-sectional observational study design. Volumes of sedentary time and prolonged sedentary time (accumulated sedentary time spent in bouts of ≥ 30 min) were measured using the Axivity AX3 accelerometer. Six cognitive functions (spatial and verbal short-term memory; and working memory, visuospatial working memory, response inhibition and planning as executive functions) were measured using six validated cognitive assessments. Data were analysed using generalised linear models.ResultsData of 119 adolescents were analysed (49% boys, 13.4 ± 0.6 year). No evidence for an association of volumes of sedentary time and prolonged sedentary time with spatial and verbal short-term memory, working memory, and visuospatial working memory was found. Volumes of sedentary time and prolonged sedentary time were significantly related to planning. One hour more sedentary time or prolonged sedentary time per day was associated with respectively on average 17.7% (95% C.I.: 3.5–29.7%) and 12.1% (95% C.I.: 3.9–19.6%) lower scores on the planning task.ConclusionsNo evidence was found for an association of volumes of habitual sedentary time and prolonged sedentary time with short-term memory and executive functions, except for planning. Furthermore, the context of sedentary activities could be an important confounder in the association of sedentary time and prolonged sedentary time with cognition among adolescents. Future research should therefore collect data on the context of sedentary activities.Trial registrationThis study was registered at ClinicalTrials.gov in January 2020 (NCT04327414; released on March 11, 2020).

Denoising CSAMT signals in the time domain based on long short-term memory

Abstract Controlled-source audio frequency magnetotellurics (CSAMT) is an artificial-source electromagnetic technique that partially mitigates the limitations of weak natural-field signals. However, practical field surveys inevitably encounter strong interference, severely affecting signal quality. Traditional methods such as Fourier transformation, which directly computes apparent resistivity from frequency-domain information, are inadequate in this context, so we need alternative denoising approaches. However, research on CSAMT denoising is currently limited. Given the excellent performance of long short-term memory (LSTM) neural networks in the processing of magnetotelluric (MT) data, as demonstrated by previous studies, this paper proposes the use of LSTM to denoise CSAMT signals in the time domain. Unlike traditional denoising methods, we aim to directly extract the target frequency signal from the time-series data for denoising. For MT data, target frequency signals and noise are all mixed together, so noise suppression can only be achieved by identifying noise characteristics in the time series. However, unlike MT data, CSAMT data has an artificial transmitting source, and the frequency of the valid signal is fixed within a time interval. This allows for the direct extraction of target frequency signals without considering the complex characteristics of noise. In this study, we developed a neural network based on bidirectional LSTM to accomplish the task of noise suppression. After conducting both simulated and measured data tests, this method was able to, on average, improve the signal-to-noise ratio (SNR) of CSAMT data by approximately 20 dB and partially address the challenge of denoising when the data's SNR falls below 0 dB.

Electroencephalographic microstates as a potential neurophysiological marker differentiating bilateral from unilateral temporal lobe epilepsy.

Electroencephalographic (EEG) microstate abnormalities have been documented in different neurological disorders. We aimed to assess whether EEG microstates are altered also in patients with temporal epilepsy (TLE) and whether they show different activations in patients with unilateral TLE (UTLE) and bilateral TLE (BTLE). Nineteen patients with UTLE, 12 with BTLE, and 15 healthy controls were enrolled. Resting state high-density electroencephalography (128 channels) was recorded for 15 min with closed eyes. We obtained a set of stable scalp maps representing the EEG activity, named microstates, from which we acquired the following variables: global explained variance (GEV), mean duration (MD), time coverage (TC), and frequency of occurrence (FO). Two-way repeated measures analysis of variance was used to compare groups, and Spearman correlation was performed to study the maps in association with the clinical and neuropsychological data. Patients with BTLE and UTLE showed differences in most of the parameters (GEV, MD, TC, FO) of the four microstate maps (A-D) compared to controls. Patients with BTLE showed a significant increase in all parameters for the microstates in Map-A and a decrease in Map-D compared to UTLE and controls. We observed a correlation between Map-A, disease duration, and spatial short-term memory, whereas microstate Map-D was correlated with the global intelligence score and short-term memory performance. A global alteration of the neural dynamics was observed in patients with TLE compared to controls. A different pattern of EEG microstate abnormalities was identified in BTLE compared to UTLE, which might represent a distinctive biomarker.

THE EFFECT OF AGE, SLEEP DURATION, LEARNING DURATION, AND COFFEE INTAKE ON SHORT-TERM MEMORY

In the reality of life, learning requires a process of remembering. In the process of teaching and learning, memory is crucial. Sleep deprivation affects the process of remembering in learning. Memory is one of the important aspects that affect student performance in learning. In addition to sleep duration, the age factor can also affect short-term memory. This study wants to find out if age, sleep duration, and learning duration have an impact on short-term memory. This study incorporates elements of cognitive ergonomics because it measures short-term memory performance to assess how people process information and how they maintain it in their memory. Based on the results of multiple linear regression tests, the R Square value was 0.328 and the Adjusted R Square was 0.272. The age variable has a significance value of 0.001. The coefficient value on the age variable is 0.303 and is positive. This positive value indicates that increasing student age will increase the value of short-term memory. This shows that the higher the age of the student, the higher the ability in short-term memory. The sleep duration variable has no significant effect on short-term memory. The learning duration variable has no significant effect on short-term memory.

Prediction of nitrous oxide emission of a municipal wastewater treatment plant using LSTM-based deep learning models.

Accurate assessment of greenhouse gas emissions from wastewater treatment plants is crucial for mitigating climate change. N2O is a potent greenhouse gas that is emitted from wastewater treatment plants during the biological denitrification process. In this study, we developed and evaluated deep learning models for predicting N2O emissions from a WWTP in Switzerland. Six key parameters were selected to obtain the optimal LSTM model by adjusting experimental parameter conditions. The optimal parameter condition was achieved with 150 neurons, the tanh activation function, the RMSprop optimization algorithm, a learning rate of 0.001, no dropout regularization, and a batch size of 128. Under the same conditions, we compared the performance of recurrent neural networks (RNNs) and Long Short-Term Memory (LSTM) networks. We found that LSTM models outperformed RNN models in predicting N2O emissions. The optimal LSTM model achieved a 36% improvement in mean absolute error (MAE), a 19% improvement in root mean squared error (RMSE), and a 6.92% improvement in R2 score compared to the RNN model. Additionally, LSTM models demonstrated better resilience to sudden changes in the target sequence, exhibiting a 9.54% higher percentage of explained variance compared to RNNs. These results highlight the potential of LSTM models for accurate and robust prediction of N2O emissions from wastewater treatment plants, contributing to effective greenhouse gas mitigation strategies.

TEHRAN AIR POLLUTION MODELING USING LONG-SHORT TERM MEMORY ALGORITHM: AN UNCERTAINTY ANALYSIS

Abstract. Air pollution is a major environmental issue in urban areas, and accurate forecasting of particles 10 μm or smaller (PM10) level is essential for smart public health policies and environmental management in Tehran, Iran. In this study, we evaluated the performance and uncertainty of long short-term memory (LSTM) model, along with two spatial interpolation methods including ordinary kriging (OK) and inverse distance weighting (IDW) for mapping the forecasted daily air pollution in Tehran. We used root mean square error (RMSE) and mean square error (MSE) to evaluate the prediction power of the LSTM model. In addition, prediction intervals (PIs), and Mean and standard deviation (STD) were employed to assess the uncertainty of the process. For this research, the air pollution data in 19 Tehran air pollution monitoring stations and temperature, humidity, wind speed and direction as influential factors were taken into account. The results showed that the OK had better RMSE and STD in the test (32.48 ± 9.8 μg/m3) and predicted data (56.6 ± 13.3 μg/m3) compared with those of the IDW in the test (47.7 ± 22.43 μg/m3) and predicted set (62.18 ± 26.1 μg/m3). However, in PIs, IDW ([0, 0.7] μg/m3) compared with the OK ([0, 0.5] μg/m3) had better performance. The LSTM model achieved in the predicted values an RMSE of 8.6 μg/m3 and a standard deviation of 9.8 μg/m3 and PIs between [2.7 ± 4.8, 14.9 ± 15] μg/m3.