Articles

Social media platforms like X (formerly Twitter) generate vast volumes of user-generated content that provide real-time insights into public sentiment. Despite the widespread use of traditional machine learning methods, their limitations in capturing contextual nuances in noisy social media text remain a challenge. This study leverages the Sentiment140 dataset, comprising 1.6 million labeled tweets, and develops predictive models for binary sentiment classification using Naive Bayes, Logistic Regression, and the transformer-based BERT model. Experiments were conducted on a balanced subset of 12,000 tweets after comprehensive NLP preprocessing. Evaluation using accuracy, F1-score, and confusion matrices revealed that BERT significantly outperforms traditional models, achieving an accuracy of 89.5% and an F1-score of 0.89 by effectively modeling contextual and semantic nuances. In contrast, Naive Bayes and Logistic Regression demonstrated reasonable but consistently lower performance. To support practical deployment, we introduce SentiFeel, an interactive tool enabling real-time sentiment analysis. While resource constraints limited the dataset size and training epochs, future work will explore full corpus utilization and the inclusion of neutral sentiment classes. These findings underscore the potential of transformer models for enhanced public opinion monitoring, marketing analytics, and policy forecasting.

Over the past few decades, cardiovascular disease and related complications have surpassed all others as the important causes of death on a universal scale. At the moment, they are the important cause of mortality universal, including in India. It is important to know how to find cardiovascular problems early so that patients get better care and prices go down. This project utilizes the UCI Heart Disease Dataset to develop ML and DL models capable of detecting cardiac diseases. Heart illness was categorized using Convolutional Neural Network (CNN) models, which are able to detect intricate patterns in supplied data. A confusion matrix rating, an F1-score, a ROC curve, accuracy, precision, and recall were some of the measures used to grade the model. It did much better than the Neural Network, Deep Neural Network (DNN), and Gradient Boosted Trees (GBT) models, with 91.71% accuracy, 88.88% precision, 82.75% memory, and 85.70% F1-score. Comparative study showed that CNN was the most accurate model. Other models had different balances between accuracy and recall. The experiment results show that the optional CNN model is a decent way to identify cardiovascular disease. This means that it could be used in healthcare systems to find diseases earlier and treat patients better.

Maintenance of large-scale software is difficult due to large size and high complexity of code.80% of software development is on maintenance and the other 60% is on trying to understand the code. The severity of the code smells must be measured as well as fairness on it because it will help the developers especially in large scale source code projects. Code smell is not a bug in the system as it doesn’t prevent the program from functioning but it may increase the risk of software failure or performance slowdown. Therefore, this paper seeks to help developers with early prediction of severity of code smells and test the level of fairness on the predictions especially in large scale source code projects. Data is the collection of facts and observations in terms of events, it is continuously growing, getting denser and more varied by the minute across different disciplines or fields. Hence, Big Data emerged and is evolving rapidly, the various types of data being processed are huge, but no one has ever thought of where this data resides, we therefore noticed this data resides in software’s and the codebases of the software’s are increasingly growing that is the size of the modules, functionalities, the size of the classes etc. Since data is growing so rapidly it also mean the codebases of software’s or code are also growing as well. Therefore, this paper seeks to discuss the 5V’s of big data in the context of software code and how to optimize or manage the big code. When we talk of "Big Code for Big Software's," we are referring to the specific challenges and considerations involved in developing, managing, and maintaining of code in large-scale software systems.

The stock market is very volatile and hard to predict accurately due to the uncertainties affecting stock prices. However, investors and stock traders can only benefit from such models by making informed decisions about buying, holding, or investing in stocks. Also, financial institutions can use such models to manage risk and optimize their customers' investment portfolios. In this paper, we use the Long Short-Term Memory (LSTM-RNN) Recurrent Neural Networks (RNN) to predict the daily closing price of the Amazon Inc. stock (ticker symbol: AMZN). We study the influence of various hyperparameters in the model to see what factors the predictive power of the model. The root mean squared error (RMSE) on the training was 2.51 with a mean absolute percentage error (MAPE) of 1.84%.

Financial markets have become more and more complex, so has been the number of data sources. Stock price prediction has hence become a tough but important task. The time dependencies in stock price movements tend to escape from traditional models. In this work, a hybrid ARIMA-LSTM model is suggested to enhance accuracy of stock price forecasts. Based on time series indicators like adjusted closing prices of S&P 500 stocks over a decade (2010–2019), the ARIMA-LSTM model combines influences of both autoregressive time series forecasting with the substantial sequence learning property of LSTM. Data preprocessing in all aspects including missing values interpolation, outlier’s detection and data scaling – Min-Max guarantees data quality. The model is trained on 90/10 training/testing split and met with main performance metrics: MaE, MSE & RMSE. As indicated in the results, the proposed ARIMA-LSTM model gives a MAE value and MSE value of 0.248 and 0.101 respectively and RMSE of 0.319, a measure high accuracy on stock price prediction. Coupled comparative analysis with other Artificial Neural Networks (ANN) and BP Neural Networks (BPNN) are examples of machine learning reference models, further illustrates the suitability and superiority of ARIMA-LSTM approach as compared to the underlying models with the least MAE and strong predictive capability. This work demonstrates the efficiency of integrating the classical time series models with deep learning methods for financial forecasting.

Advances in web-centric cloud computing have facilitated the establishment of an integrated cloud environment connecting a wide variety of clinical trial stakeholders. A web-centric cloud framework is proposed for real-time monitoring and risk prediction during clinical trials. The framework focuses on identifying relevant datasets, developing a data-management interface, and implementing machine-learning algorithms for data analysis. Detailed descriptions of the data-management interface and the machine-learning processes are provided, targeting active clinical trials with therapeutic uses in cancer. Demonstrations utilize publicly available clinical-trial data from the ClinicalTrials.gov repository. The real-time monitoring and risk prediction systems were assessed by developing five supervised-classification-machine-learning models for trial-status prediction and six unsupervised models for patient-safety-profile assessment, each representing a different phase of the clinical-trial process. All supervised models yielded high accuracy and area-under-the-curve values at the testing stage, while the unsupervised models demonstrated practical applicability. The results underscore the advantages of using the trial-status algorithm, the patient-safety-profile model, and the proposed framework for performing real-time monitoring and risk prediction of clinical trials.

Artificial intelligence systems have been previously used to predict post-operative complications in small studies and single institutions. Here we developed a robust artificial intelligence model that predicts the risk of having cardiac, pulmonary, thromboembolic, or septic complications after elective, non-cardiac, non-ambulatory surgery. We combined structured and unstructured electronic health record data from 3.5 million surgical encounters from 25 medical centers between 2009 and 2017. Our neural network model predicted postoperative comorbidities 15 to 80 times faster than classical models. As such, our model can be used to assess the risk of having a specific complication postoperatively in a fraction of a second. With our model, we believe clinicians will be able to identify high-risk surgical patients and use their good judgment to mitigate upcoming risks, ultimately improving patient outcomes [1].

The growing complexity and variability in healthcare delivery and costs within Medicare Advantage (MA) and Medicare Supplement (Medigap) plans present significant challenges for improving health outcomes and managing expenditures. Neural networks, a subset of artificial intelligence (AI), have shown considerable promise in optimizing healthcare processes, particularly in predictive modeling, personalized treatment recommendations, and risk stratification. This paper explores the application of neural networks in enhancing health outcomes within the context of Medicare Advantage and Supplement plans. We review how deep learning models can be leveraged to predict patient risk, optimize resource allocation, and identify at-risk populations for preventive interventions. Additionally, we discuss the potential for neural networks to improve claims processing, reduce fraud, and streamline administrative burdens. By integrating various data sources, including medical records, claims data, and demographic information, neural networks enable more accurate and efficient decision-making processes. Ultimately, this approach can lead to better patient care, reduced healthcare costs, and improved satisfaction for beneficiaries of these programs. The paper concludes by highlighting the current limitations, ethical considerations, and future directions for AI adoption in the Medicare Advantage and Supplement sectors.

The growth in demand for rail transportation systems within cities, together with high-speed and long-distance transportation running on a rail network, raises the issues of both rail safety and security. If an accident or an attack occurs, its consequences can be extremely severe. To mitigate the impact of these events, the real-time monitoring of a rail system is required. In that case, the improvements in monitoring can be achieved using artificial intelligence algorithms such as neural networks. Neural networks have been used to achieve real-time incident identification in monitoring the track quality in terms of classifying the graphical outputs of an ultrasonic system working with the rails and track bed, to predict incidents on the rail infrastructure due to transmission channels becoming blocked, and also to attempt scheduling preemptive and preventative maintenance. In terms of forecasting incidents and accidents on board the trains, neural networks have been used to model passenger behavior and optimize responses during a train station evacuation. In tackling the incidents and accidents occurring on rail transport, we contribute with two methodologies to detect anomalies in real-time and identify the level of security risk: at the maintenance level with personnel operating along the railways, and onboard passenger trains. These methodologies were evaluated on real-world datasets and shown to be able to achieve a high accuracy in the results. The results generated from these case studies also reveal the potential for network-wide applications, which could enhance security and safety on railway networks by offering the possibility of better managing network disruptions and more rapidly identifying security issues. The speed and coverage of the information generated through the implementation of these methodologies have implications in utilizing prediction for decision support and enhancing safety and security on board the rail network.

In this research work, we have applied “Lipinski’s RO5” for pharmacokinetics (PK) study and to predict the activity of peptidic HIV-1 protease inhibitors. Peptidic HIV-1-PRIs have been taken from literature with their observed biological activities (OBAs) in term of IC50. The logarithms of the inverse of IC50 have been used as biological end point o(log1/C) in the study. For calculation of physicochemical parameters, the molecular modeling and geometry optimization of all the derivatives have been carried out with CAChe Pro software using semiempirical PM3 method. Prediction of the biological activity of the inhibitors has shown that the best QSAR model is constructed from pharmacokinetic properties, molecular weight and hydrogen bond acceptor. This also proved that these properties play important role to describe the PKs of the drugs. On the basis of the derived models one can build up a theoretical basis to access the biological activity of the compounds of the same series.

Carboxylates and their antimony(III) complexes experimentally scanned earlier for anti-leishmanial activity (IC₅₀) against five leishmanial strains viz., <i>L. major, L. major (Pak),</i><i> </i><i>L. tropica, L. mex mex, </i>and<i> L. donovani</i>. These activities have been theoretically predicted by DFT method along with quantitative structure-activity relationship (QSAR) study. Molecular modeling and geometry optimization of the all the eight compounds have been performed on workspace program of CAChe Pro software of Fujitsu by opting B88-PW91 (Becke '88; Perdew &#x26; Wang '91) GGA (generalized-gradient approximation) energy functional with DZVP (double-zeta valence polarized ) basis set in DFT (Density Functional Theory). For QSAR, multiple linear regression (MLR) analysis has been performed on Project Leader Program associated with CAChe. The reliability of correlation between experimental activities and predicted activities are r²= 0.826, r²_CV= 0.426 (<i>L. major</i>); r²= 0.905, r²_CV= 0.507 (<i>L. major (Pak)</i>); r²= 0.980, r²_CV= 0.932 (<i>L. tropica</i>); r²= 0.781, r²_CV= 0.580 (<i>L. mex mex</i>) and r²= 0.634, r²_CV= 0.376 (<i>L. donovani</i>), and a comparison of the experimental values and the values obtained by theoretical calculations has been presented pictorially that shows close resemblance.

For years, risk assessment and financial calculations have been based on mathematical, statistical, and actuarial studies of existing and historical data. The manual process of building datasets, processing data, deriving trends, identifying periodicities, and analyzing diagnostics is extremely expensive and time-consuming. With the automation and evolution of data science technologies, organizations are now bringing in niche data, such as unstructured data, which contain more disruptive and precise signals for decision-making—thereby making predictions and derivative valuations more robust. This discussion highlights how investment decision-making and financial ecosystem activities are set to be transformed with the power of technical automation, data, and artificial intelligence. A noted trend in the financial investment sector is that financial valuations are highly predictive and highly non-linear in long-term occurrences. To understand these robust evolving signals and execute profitable strategies upon them, the investment management process needs to be very dynamic, open, smart, and technically deep. However, with current manual processes, reaching a high-end asset prediction still seems like a shot in the dark. In parallel, open and democratically developed financial ecosystems query relatively riskless premium opportunities in high-finance valuation and perception. The process of evolving financial ecosystems or the use of automated tools and data to move to unique frontiers could make high-yield profiting opportunities very safe and entirely riskless. Financial economic theories and realistic approximation models support this.

Crime is a serious and widespread problem in their society, thus preventing it is essential. Assignment. A significant number of crimes are committed every day. One tool for dealing with model crime is data mining. Crimes are costly to society in many ways, and they are also a major source of frustration for its members. A major area of machine learning research is crime detection. This paper analyzes crime prediction and classification using data mining techniques on a crime dataset spanning 2006 to 2016. This approach begins with cleaning and extracting features from raw data for data preparation. Then, machine learning and deep learning models, including RNN-LSTM, ARIMA, and Linear Regression, are applied. The performance of these models is evaluated using metrics like Root Mean Squared Error (RMSE) and Mean Absolute Percentage Error (MAPE). The RNN-LSTM model achieved the lowest RMSE of 18.42, demonstrating superior predictive accuracy among the evaluated models. Data visualization techniques further unveiled crime patterns, offering actionable insights to prevent crime.

In this study, oil was extracted from Moringa seed using mechanical and solvent methods. To transesterify the oil into biodiesel, factorial design of experiment of 2<SUP>4</SUP> was used to obtain different combination factors at different level of reaction temperature, catalyst amount, reaction time and alcohol to oil ratio, giving rise to 48 experimental runs. The oil sample was transesterified in 48 experimental runs, in each case the biodiesel yield was recorded in percentage. The biodiesel was then characterized according to ASTM test protocol. Factorial design model was developed using Design Expert 7.0, the model generated R of 0.987 and Mean Square Error (MSE) of 5.0453 and was used to predict and optimize biodiesel yield. Artificial Neural Network (ANN) model from MATLAB R2016a was developed using 4 input variables and 30 runs, the remaining 18 runs were tested with the ANN model to predict and compare the biodiesel yield with the experimental biodiesel yield, the model generated R value of 0.99687 and MSE of 3.50804. It was found that solvent method yielded more oil than mechanical method, the biodiesel has good thermo-physical property, optimum biodiesel yield of 91.45 % was obtained at 5:1 alcohol/ oil molar ratio, 18.89 wt% catalyst amounts, 45 minutes reaction time and at 45 reaction temperature. The experimental validation yielded 88.33 % biodiesel. The ANN model adequately predicted the remaining 18 runs with R<SUP>2</SUP> value of 0.99649 and MSE of 4.914243. Both models proved adequate enough to predict biodiesel yield but ANN model proved more adequate.

First, we propose the nonlinear whole seismology and its three basic laws. Next, based on the nonlinear equations of fluid dynamics in Earth’s crust, we obtain a chaos equation, in which chaos corresponds to the earthquake, and shows complexity on seismology. But, combining the Carlson-Langer model and the Gutenberg-Richter relation, a simplified nonlinear solution and corresponding magnitude-period formula of earthquakes may be derived approximately. Further, we research the topological seismology. From these theories some predictions can be calculated quantitatively and are already tested. Combining the Lorenz nonlinear model, we may discuss the earthquake migration to and fro. Finally, if various modern scientific instruments, different scientific theories and some paranormal ways for earthquake are combined each other, the accuracy of multilevel prediction will be increased.

Cell division that is not controlled leads to cancer, an incurable condition. An early diagnosis has the potential to lower death rates from breast cancer, the most frequent disease in women worldwide. Imaging studies of the breast may help doctors find the disease and diagnose it. This study explores an effectiveness of DL and ML models in a classification of mammography images for breast cancer detection, utilizing the publicly available CBIS-DDSM dataset, which comprises 5,000 images evenly divided between benign and malignant cases. To improve diagnostic accuracy, models such as Gaussian Naïve Bayes (GNB), CNNs, KNN, and MobileNetV2 were assessed employing performance measures including F1-score, recall, accuracy, and precision. The methodology involved data preprocessing techniques, including transfer learning and feature extraction, followed by data splitting for robust model training and evaluation. Findings indicate that MobileNetV2 achieved a highest accuracy99.4%, significantly outperforming GNB (87.2%), CNN (96.7%), and KNN (91.2%). The outstanding capacity of MobileNetV2 to identify between benign and malignant instances was shown by the investigation, which also made use of confusion matrices and ROC curves to evaluate model performance.

Predictive modeling, supported by machine learning technology, aims to analyze data in order to guide decision-making towards actions generating desired values in the future. It encompasses the set of techniques used to build models that estimate the value of a certain variable predicting a forthcoming event from the past or current values of relevant attributes. In predictive healthcare modeling, the built models represent the relationship among the data concerning customer, provider, production, and other aspects of the healthcare environment in order to assist the decision processes in the prevention of diseases and in the planning of preventive actions by detection of high-risk patients. Contrary to trend analysis, whose goal is to describe past events, predictive models aim to provide useful indications regarding future events and changes. Predictive healthcare modeling supports actions that try to prevent the manifestation of diseases in healthy individuals or try to diagnose as early as possible the incidence of a disease in patients at risk. A sound predictive analysis encompasses not only the model-training task, but also the aspects of data quality, preprocessing, and fusion during its entire implementation lifecycle to ensure appropriate input data preparation. The robustness of the predictive model and its results depends highly on data quality. Due to the variety of data sources in healthcare environments, it becomes essential to use preprocessing in order to remove noise and inconsistencies. The increasing number of endorsable data exchange standards makes each data exchange achievable, but it demands the implementation of a data-governance program. In addition, the influence of the hospital-database architect on the architecture of an early-diagnosis model is important to guarantee appropriate input-formatting modularity.