Deep Learning Model for Detection and Severity Analysis of Car Accidents

A survey on pre-training requirements for deep learning models to detect obstructive sleep apnea events

Over the last two decades, social media platforms have grown dramatically. Twitter and Facebook are the two most popular social media platforms, with millions of active users posting billions of messages daily. These platforms allow users to have freedom of expression. However, some users exploit this facility by disseminating hate speeches. Manual detection and censorship of such hate speeches are impractical; thus, an automatic detection mechanism is required to detect and counter hate speeches in a real-time environment. Most research in hate speech detection has been carried out in the English language. Still, minimal work has been explored in other languages, mainly Urdu written in Roman Urdu script. A few research have attempted machine learning, and deep learning models for Roman Urdu hate speech detection; however, due to a scarcity of Roman Urdu resources, and a large corpus with defined annotation rules, a robust hate speech detection model is still required. With this motivation, this study contributes in the following manner: we developed annotation guidelines for Roman Urdu Hate Speech. Second, we constructed a new Roman Urdu Hate Speech Dataset (RU-HSD-30K) that was annotated by a team of experts using the annotation rules. To the best of our knowledge, the Bi-LSTM model with an attention layer for Roman-Urdu Hate Speech Detection has not been explored. Therefore, we developed a context-aware Roman Urdu Hate Speech detection model based on Bi-LSTM with an attention layer and used custom word2vec for word embeddings. Finally, we examined the effect of lexical normalization of Roman Urdu words on the performance of the proposed model. Different traditional as well as deep learning models, including LSTM and CNN models, were used as baseline models. The performance of the models was assessed in terms of evaluation matrices like accuracy, precision, recall, and F1-score. The generalization of each model is also evaluated on a cross-domain dataset. Experimental results revealed that Bi-LSTM with attention outperformed the traditional machine learning models and other deep learning models with an accuracy score of 0.875 and an F-Score of 0.885. In addition, the results demonstrated that our suggested model (Bi-LSTM with Attention Layer) is more general than previous models when applied to unseen data. The results confirmed that lexical normalization of Roman Urdu words enhanced the performance of the suggested model.

Monitoring insect activity at hive entrances is essential for advancing precision beekeeping practices by enabling non-invasive, real-time assessment of the colony’s health and early detection of potential threats. This study evaluates deep learning models for detecting worker bees, pollen-bearing bees, drones, and wasps, comparing different YOLO-based architectures optimized for real-time inference on an RTX 4080 Super and Jetson AGX Orin. A new publicly available dataset with diverse environmental conditions was used for training and validation. Performance comparisons showed that modified YOLOv8 models achieved a better precision–speed trade-off relative to other YOLO-based architectures, enabling efficient deployment on embedded platforms. Results indicate that model modifications enhance detection accuracy while reducing inference time, particularly for small object classes such as pollen. The study explores the impact of different annotation strategies on classification performance and tracking consistency. The findings demonstrate the feasibility of deploying AI-powered hive monitoring systems on embedded platforms, with potential applications in precision beekeeping and pollination surveillance.

This paper proposes a deep-learning model for the detection of periodontal bone loss from dental radiographs. The process involves pre-processing using Cross Guided Bilateral Filtering (CGBiF), resizing and image enhancement to remove noise, followed by segmentation using a Region of Interest (ROI) model to isolate relevant periodontal structures. A novel Twin Attention-based Deep Red Convolutional Long Short Term Memory (TA_DRCLSTM) model is introduced to detect bone loss and is optimized with Enhanced Red Panda Optimization (ERPO). The model is evaluated and compared to existing approaches in terms of precision, recall, specificity, NPV and AUC-PR, achieving superior performance. The proposed model demonstrates effective and accurate detection of periodontal bone loss, providing a valuable tool for clinical applications. CGBiF improves image clarity for better detection by lowering noise and maintaining crucial features like bone structures. Resizing standardizes image dimensions, allowing for consistent model input. Image enhancement increases contrast, making periodontal structures more visible, resulting in a 93.8% recall, 81.2% precision and 95.4% AUC-PR in the study.

Efficient irrigation scheduling is crucial for optimizing crop yields and conserving water resources. In this paper, we present robust models for predicting soil moisture trends in cotton based on the integration of IoT, machine learning (ML), and deep learning (DL) frameworks. We access soil moisture and weather forecast data from IoT field sensors using web APIs, then process and incorporate them into ML and DL models to predict soil moisture trends. Specifically, in DL we used the Long Short-Term Memory (LSTM) with two LSTM architectures (Single Stream LSTM and Dual LSTM), while in ML we used the Gradient Boosting (GB), XGBoost, Random Forest (RF), and AdaBoost methods. Experiment results show that, Single Stream LSTM achieved MSE (0.0710) and (0.9684) R 2 outperforming its Dual counterpart which recorded MSE (0.1200) and (0.9466) R 2 , while under ML, Gradient Boosting (MSE = 0.2822, R² = 0.9909) outperforms other ML models. In general, LSTM methods achieved the lowest error while ML methods recorded the highest R 2 -score. Predictive methods presented in this paper, have shown great potential in estimating soil moisture patterns for optimal irrigation schedules in cotton, thereby enhancing water use efficiency and increasing crops yield.

Accurate real-time bus arrival information is essential for an efficient public transport network, as it significantly impacts passenger experience, system reliability, reduced waiting times, dwell times, and operational efficiency. In Sri Lanka's public transport system, current bus arrival time computations primarily rely on static data, neglecting real-time information and critical factors influencing travel times. This highlights the need to identify the unique variables affecting bus arrival times within the Sri Lankan context and to develop robust prediction models that account for these influences. While traditional methods such as Historical Average Models, Regression, Time Series Analysis, and Kalman Filtering have been used in previous research for short-term travel time predictions, Machine Learning (ML) approaches have proven to deliver superior accuracy. ML models are regarded as the most effective for heterogeneous, lane-less traffic conditions with varying traffic volumes, such as those found in Sri Lanka. ML techniques excel in processing large, high-quality datasets and provide accurate predictions by accounting for all relevant variables influencing travel times. Although research has been conducted on developing various basic ML models for travel time prediction, there is a noticeable gap in studies comparing these models to determine the most suitable one for the Sri Lankan context. A Long- Short Term Memory (LSTM) neural network is a deep learning model that is capable of handling long-term dependencies. In the context of bus travel time prediction, LSTMs can leverage historical traffic and travel data to capture temporal patterns and fluctuations that influence travel times. By evaluating LSTMs against basic machine learning models, this study seeks to explore the advantages of applying deep learning techniques to transportation forecasting, ultimately contributing to more accurate and efficient predictive systems in transit planning. The ML models selected in this study include two basic traditional models K- Nearest Neighbours (KNN) and Support Vector Regression (SVR) and four advanced models that utilize ensemble techniques and advanced optimization as Random Forest Regression (RFR), Ada Boost, XG Boost and Gradient Boosting Machine (GBM). The performance of these models was compared with the LSTM model to identify the gap in their accuracies.

Most Graph Neural Networks (GNNs) follow the message-passing scheme. Residual connection is an effective strategy to tackle GNNs' over-smoothing issue and performance reduction issue on non-homophilic networks. Unfortunately, the coarse-grained residual connection still suffers from class-imbalanced over-smoothing issue, due to the fixed and linear combination of topology and attribute in node representation learning. To make the combination flexible to capture complicated relationship, this paper reveals that the residual connection needs to be node-dependent, layer-dependent, and related to both topology and attribute. To alleviate the difficulty in specifying complicated relationship, this paper presents a novel perspective on GNNs, i.e., the representations of one node in different layers can be seen as a sequence of states. From this perspective, existing residual connections are not flexible enough for sequence modeling. Therefore, a novel node-dependent residual connection, i.e., Long Short-Term Graph Memory Network (LSTGM) is proposed to employ Long Short-Term Memory (LSTM), to model the sequence of node representation. To make the graph topology fully employed, LSTGM innovatively enhances the updated memory and three gates with graph topology. A speedup version is also proposed for effective training. Experimental evaluations on real-world datasets demonstrate their effectiveness in preventing over-smoothing issue and handling networks with heterophily.

Electrical energy production is an important function in a country that has abundant demand for society. Therefore, power supply forecasting is very significant to keep the balance of power consumption and production and should not be in short supply at any cost. Thus, Forecasting and modelling time series analysis of power production becomes important in prediction. We explore the synergy between Gradient Boosting and Long Short-Term Memory (LSTM) networks in the context of time series analysis. Specifically, we investigate how XGBoost and LightGBM, well-known gradient boosting frameworks, complement LSTM, a recurrent neural network architecture. By combining these methods, we aim to uncover deeper insights and elevate predictive potential for time series forecasting. In this paper, we examine the machine learning methods of forecasting the power generation of 33-year time series data. In the conclusion, we put forward a comparative study between the outcomes attained from the use of LSTM and Gradient-boosting tree-based algorithms XGBoost and LightGBM. The results show that the XGBoost model outclasses the other models with a low error value in forecasting power generation.

The research study is to predict human activity recognition and analyze their strength in machine learning (ML) processes to achieve precision and also to determine performance of human activities. The framework to classify the Novel Long Short Term Memory (LSTM) and Decision Tree (DT) to perform all measures. This research study used the Novel Long Short Term Memory and Decision Tree to perform the required actions to accurately identify body movements in adjacent film sequence. Recent study findings and a threshold of 0.05%, as well as 95% confidence interval and standard parameter estimates, were castoff in a study 47 samples divided into two groups calculated using the G-power of 80%. Since the novel Long Short Term Memory (LSTM) algorithm and local invariant methods have achieved an accuracy of 93.83% in predicting the activity analysis, this research needs to obtain improved accuracy to perform activity prediction by using the Decision Tree algorithm. When applied to human activity analysis, the Novel Long Short Term Memory technique was shown to have an accuracy of 90.12 percent, with a 95% confidence interval including the value 0.045 (p 0.05). Researchers found that when comparing two algorithms for analyzing human behaviour, Long Short Term Memory (LSTM) performed much better than Decision Tree (DT).

The prediction of wind energy generation is important to enhance the performance and dependability of renewable energy systems due to the rising demand for wind-generated electricity and advancements in wind energy technology competitiveness. This study leverages advanced machine learning (ML) and some other statistical and deep learning based time series forecasting models to enhance the accuracy of wind energy predictions. This comprehensive analysis includes nine ML models—Linear Regression, Random Forests (RF), Gradient Boosting Machines (GBM), Support Vector Machines (SVM), K-Nearest Neighbors (KNN), AdaBoost, XGBoost, Support Vector Regression (SVR), and Neural Networks—as well as Four time-series forecasting models—ARIMA, Temporal Convolutional Networks (TCNs), Long Short-Term Memory (LSTM) networks and GRU. Each ML model underwent rigorous cross-validation to ensure optimal performance. The assessment criteria utilized here comprised the Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and the R² Score. It was found that among the nine ML models, Random Forests, GBM and KNN consistently provided superior accuracy and robustness, making them the top choices for wind energy prediction whereas the performance of linear regression, SVM and SVR were very poor for the considered dataset. From the experiment, Random Forest, GBM, and KNN showed the best performance with low MSE values of 0.77, 1.95, and 1.51 respectively, while other models had MSEs above 7.5, with AdaBoost reaching 30. Their RMSEs (0.88, 1.40, 1.23) and MAEs (0.093, 0.73, 0.10) also indicate strong predictive accuracy compared to the rest.In this paper, time series forecasting, TCNs, LSTM and GRU networks showed strong capabilities in capturing temporal dependencies and trends within the wind energy data. Visualization techniques were employed to compare model performances comprehensively, providing clear insights into their predictive power. Therefore, this present study offers a robust framework for researchers and practitioners aiming to leverage machine learning and time series forecasting in the realm of renewable energy prediction.

Plant stress reduction research has advanced significantly with the use of Artificial Intelligence (AI) techniques, such as machine learning and deep learning. This is a significant step toward sustainable agriculture. Innovative insights into the physiological responses of plants mostly crops to drought stress have been revealed through the use of complex algorithms like gradient boosting, support vector machines (SVM), recurrent neural network (RNN), and long short-term memory (LSTM), combined with a thorough examination of the TYRKC and RBR-E3 domains in stress-associated signaling proteins across a range of crop species. Modern resources were used in this study, including the UniProt protein database for crop physiochemical properties associated with specific signaling domains and the SMART database for signaling protein domains. These insights were then applied to deep learning and machine learning techniques after careful data processing. The rigorous metric evaluations and ablation analysis that typified the study’s approach highlighted the algorithms’ effectiveness and dependability in recognizing and classifying stress events. Notably, the accuracy of SVM was 82%, while gradient boosting and RNN showed 96%, and 94%, respectively and LSTM obtained an astounding 97% accuracy. The study observed these successes but also highlights the ongoing obstacles to AI adoption in agriculture, emphasizing the need for creative thinking and interdisciplinary cooperation. In addition to its scholarly value, the collected data has significant implications for improving resource efficiency, directing precision agricultural methods, and supporting global food security programs. Notably, the gradient boosting and LSTM algorithm outperformed the others with an exceptional accuracy of 96% and 97%, demonstrating their potential for accurate stress categorization. This work highlights the revolutionary potential of AI to completely disrupt the agricultural industry while simultaneously advancing our understanding of plant stress responses.

Software vulnerabilities may lead to crucial security risks in software systems. Thus, prioritization of the vulnerabilities is an important task for security teams, and assessing how severe the vulnerabilities are would help teams during fixing and maintenance activities. We replicated a prior work which aims to predict the severity of software vulnerabilities by grouping vulnerabilities into different severity levels. We follow their approach on feature extraction using word embeddings, and on prediction model using Convolutional Neural Networks (CNNs). In addition, Long Short Term Memory (LSTM) and Extreme Gradient Boosting (XGBoost) models are used. We also extend the replicated work by aiming to predict severity scores rather than levels. We carried out two experiments for predicting severity levels and severity scores of 82,974 vulnerabilities. On predicting the severity levels, our LSTM and CNN models perform similarly with an F1 score of 0.756 F1 score and 0.752, respectively. On predicting the severity scores, LSTM, CNN and XGBoost models perform 16.14%, 17.03%, 18.91% MAPE values, respectively.

Travel planning is a process that allows users to obtain maximum benefit from their time, cost and energy. When planning a route from one place to another, it is an important option to present alternative travel areas on the route. This study proposes a travel route planning (TRP) architecture using a Long Short-Term Memory (LSTM) and Extreme Gradient Boosting (XGBoost) model to improve both travel efficiency and environmental sustainability in route selection. This model incorporates carbon emissions directly into the route planning process by unifying user preferences, location recommendations, route optimization, and multimodal vehicle selection within a comprehensive framework. By merging environmental sustainability with user-focused travel planning, it generates personalized, practical, and low-carbon travel routes. The carbon emissions observed with TRP’s artificial intelligence (AI) recommendation route are presented comparatively with those of the user-determined route. XGBoost, Random Forest (RF), Categorical Boosting (CatBoost), Light Gradient Boosting Machine (LightGBM), (Extra Trees Regressor) ETR, and Multi-Layer Perception (MLP) models are applied to the TRP model. LSTM is compared with Recurrent Neural Networks (RNNs) and Gated Recurrent Unit (GRU) models. Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Squared Error (MSE), and Normalized Root Mean Square Error (NRMSE) error measurements of these models are carried out, and the best result is obtained using XGBoost and LSTM. TRP enhances environmental responsibility awareness within travel planning by integrating sustainability-oriented parameters into the decision-making process. Unlike conventional reservation systems, this model encourages individuals and organizations to prioritize eco-friendly options by considering not only financial factors but also environmental and socio-cultural impacts. By promoting responsible travel behaviors and supporting the adoption of sustainable tourism practices, the proposed approach contributes significantly to the broader dissemination of environmentally conscious travel choices.

In this paper, we predict crude oil price using various factors that may influence its price. The factors considered are physical market, financial, and trading market factors, including seven key factors and the dollar index. Firstly, we select the main factors that may greatly influence the prices. Then, we develop a hybrid model based on a convolutional neural network (CNN) and long short-term memory (LSTM) network to predict the prices. Lastly, we compare the CNN–LSTM model with other models, namely gradient boosting (GB), decision trees (DTs), random forests (RFs), neural networks (NNs), CNN, LSTM, and bidirectional LSTM (Bi–LSTM). The empirical results show that the CNN–LSTM model outperforms these models.

Abstract: This research is driven by the challenges faced by Universitas Lancang Kuning (UNILAK) in attracting applicants amidst intense competition, especially after the government's policy opened independent pathways to State Universities (PTN) from 2022-2023, which impacted private university applicant numbers. To address this and support strategic planning, this study aims to predict the trend of prospective students applying to all study programs at UNILAK for the period 2025-2027. Two time series models were employed: ARIMA (AutoRegressive Integrated Moving Average) and LSTM (Long Short-Term Memory). Applicant data from 2019 to 2024 was used to build the model. The Augmented Dickey-Fuller (ADF) test confirmed the data's stationarity with a p-value of 0.0. ACF and PACF analyses determined the ARIMA parameters as p=1, d=1, q=1. The LSTM model was trained to capture more complex data patterns. ARIMA predictions for 2025, 2026, and 2027 are 3298.66, 3362.33, and 3371.30, respectively. LSTM predictions for the same years are 3335.64, 3476.52, and 3518.42. Evaluation using Root Mean Squared Error (RMSE) showed ARIMA (RMSE=588.72) to be more accurate than LSTM (RMSE=653.96). Nevertheless, LSTM provided a more optimistic prediction. This study concludes that ARIMA is better suited for short-term planning, while LSTM can be used for more ambitious long-term strategies. Keywords: arima; LSTM; applicants; prediction; university Abstrak: Penelitian ini didorong oleh tantangan Universitas Lancang Kuning (UNILAK) dalam menarik pendaftar di tengah persaingan ketat, khususnya setelah kebijakan pemerintah membuka jalur mandiri ke Perguruan Tinggi Negeri (PTN) sejak 2022-2023, yang menyebabkan penurunan jumlah pendaftar di universitas swasta. Untuk mendukung perencanaan strategis, studi ini bertujuan memprediksi tren jumlah calon mahasiswa yang mendaftar ke seluruh program studi di UNILAK untuk periode 2025-2027.Dua model deret waktu digunakan: ARIMA (AutoRegressive Integrated Moving Average) dan LSTM (Long Short-Term Memory). Data jumlah pendaftar dari 2019 hingga 2024 digunakan untuk membangun model. Uji Augmented Dickey-Fuller (ADF) menunjukkan data stasioner dengan p-value 0,0. Analisis ACF dan PACF menentukan parameter ARIMA sebagai p=1, d=1, q=1. Model LSTM dilatih untuk menangkap pola data yang lebih kompleks.Prediksi ARIMA untuk 2025, 2026, dan 2027 adalah 3298.66, 3362.33, dan 3371.30. Prediksi LSTM untuk tahun yang sama adalah 3335.64, 3476.52, dan 3518.42. Evaluasi menggunakan Root Mean Squared Error (RMSE) menunjukkan ARIMA (RMSE=588.72) lebih akurat daripada LSTM (RMSE=653.96). Meskipun demikian, LSTM memberikan prediksi yang lebih optimis. Studi ini menyimpulkan ARIMA lebih cocok untuk perencanaan jangka pendek, sementara LSTM dapat digunakan untuk strategi jangka panjang yang ambisius. Kata kunci: arima; LSTM; pendaftar; prediksi; universitas