Explore the vast range of titles available.

Bacteria are single-celled organisms with a propensity to survive in a wide range of environments. Most of these species can be found both in soil and in oceans whereas some of them are also present in the human body. Majority of bacterial species are harmful to humans, producing a wide range of infectious diseases like cholera, strep throat, tuberculosis, etc. Only a small minority of bacterial species are beneficial to humans. Thus, the study of bacteria is extremely important to analyze, identify benefits and to get rid of their negative effects. Microbiologist uses slide culturing process for bacteria identification, which involve microscope for examination of various species of bacteria. As a result, the shapes of the various samples differ, and to distinguish one sample from another, several characteristics, such as differences in cellular structure and cell-component divergence, are seen. This process is time-consuming and labor-intensive and significantly dependent on expensive machinery and human expertise. A variety of defects and problems can be easily remedied with the development and widespread use of machine learning-based computer assisted solutions in this area. The model developed with the help of machine learning tools and technologies is particularly effective in this field of image analysis and has shown an extraordinarily high rate of improvement in clinical microbiology research by recognizing different bacterial species. For more precise and better outcomes in the categorization of bacteria, feature extraction from digital images is crucial and incredibly vital. The method of feature extraction helps to eliminate extraneous data from a data collection, which speeds up learning and generalization throughout the entire machine learning process. In this paper, we try to attempt various machine learning methods to build an ensembled feature extractor and selector for better classification of microscopic bacterial Images. In this paper, we attempt to compare different feature extraction algorithms with various machine learning classifiers. The experiments have been performed on a novel dataset comprising microscopic images of four bacteria species i.e. Acetobacter aceti , Micrococcus luteus , Bacillus anthracis and Thermus sp. For feature extraction HOG (Histogram of Oriented Gradients), LBP (Local Binary Pattern), ResNet50 and VGG16 techniques have been employed. Using these features performance of five classification algorithms i.e., SVM (Support Vector Machine), RF (Random Forest), Naïve Bayes, Decision Tree and KNN (K-Nearest Neighbor) has been compared. Moreover, comparison has also been performed among feature extraction techniques also. The experimental results show that the combination of SVM and the deep features extracted using VGG16 outperformed other techniques in terms of different classification performance measures, and achieved an accuracy of 99.89%.

Natural Language Processing (NLP) is presently among the hottest scientific fields with an enormous growth rate of the relevant research. Sentiment analysis is a popular NLP problem that aims at the automatic identification of the polarity in user reviews, tweets, blog posts, comments, forum discussions and so on. Unfortunately, the natural sparseness of text, along with its intimate high dimensionality renders the direct application of machine/deep learning models problematic. For this reason, the relevant literature contains a wealth of state-of-the-art dimensionality reduction methods that confront these issues. In this paper, we conduct an experimental study on the effects of dimensionality reduction in the area of sentiment classification. More specifically, we consider multiple feature selection and feature extraction techniques and we investigate their impact on the effectiveness and the efficiency of seven state-of-the-art classifiers. The experimental evaluation includes accuracy and execution time measurements on four benchmark datasets with various degrees of reduction aggressiveness. The results indicate that, in most cases, dimensionality reduction has indeed a beneficial impact on the running times, whereas the accuracy sacrifices are usually small. However, we also indicate several exceptions where this observation is not valid. These exceptions are appropriately highlighted and discussed.

A brain tumor develops when cells multiply rapidly out of control. There is a risk of death if it is not treated in the early stages. Accurate segmentation and classification are still difficult, despite many significant efforts and promising outcomes. The wide range of tumor locations, shapes and sizes causes a significant obstacle in the field of brain tumor diagnosis. The goal of this study is to provide a comprehensive analysis of brain tumor detection malignant or benign using different features of the dataset. Our proposed model focused on the application of Machine Learning Techniques using an ensemble method to develop and classify them into malignant or benign brain tumors. The overall analysis is divided into two parts: first, we extract 30 attributes related to brain tumors from MR images, where datasets are publicly available. After that, we used the ensemble method to detect the tumors from said attributes and segment them into two categories malignant or benign tumors. The outputs of our model give robustness and cross-validation revealing to the accuracy, precision, recall, and AUC as 95.26%, 95.55%, 97.21%, and 96%, respectively. This study proposed a method of dividing the brain tumor with minimal human intervention. The goal of the proposed model is to reduce identification time so that neurosurgeons can get back to saving lives. The experimental results suggest that the method is nearly as accurate as the best existing methods.

With the increase in urbanization and population, urban waste has become a big problem. Separating these wastes as recyclable or non-recyclable and separating the recyclables according to their types have become crucial. Automating recycling systems, making improvements in this regard, and reducing human impact are beneficial for human health and increase the possibility of efficient operation of the systems. Technologies such as image processing and artificial intelligence automate the systems. This study selected six new and most preferred models from deep learning methods to separate waste materials, and training was carried out using fivefold cross-validation.The overall and class-based success of the results was measured with six different metrics: accuracy, precision, recall, F1-score, MCC and ROC curve. These were the MobileNetV2 model, which gave the most successful results with 99.36% accuracy, 0.94 MCC value, 0.99 recall, and finally 0.98 F1-Score and precision. This stands out as the most successful result in this field. In addition, the one-vs-rest method has been applied in the MobileNetV2 model to comment on the classes.

The analysis of medical data is a significant opportunity worldwide for national health systems to reduce costs and at the same time improve healthcare. The utilization of these technologies is done in the context of monitoring health issues, counting health goals, as well as for recording medical data. In such a context, early detection of users at risk of lower compliance rates and patterns of use of a health monitoring application suggesting a risk of abandonment is an invaluable opportunity to implement tailored intervention strategies aimed at recovering and avoiding abandonment thoughts. This study aims to identify patterns of early dropout in users of an application for mobile intervention, having access to a database of users who have experienced the impact of a digital monitoring application to improve their quality of life for at least 6 months. At the experimental stage, many different approaches for early dropout prediction were implemented with a different set of features. Specifically, the current study proposes a methodology using the Neighborhood Cleaning Rule and a specific classification algorithm based on the Stacked Generalization learning method to predict the early abandonment of users of the health monitoring application. The results showed that the proposed algorithm was able to predict the early dropout of users from the application with an accuracy of 97.6%, making it reliable enough to be used as an early warning system.

Kidney stone, cyst and tumour categorization are the more complex task for the radiologist to diagnosis the problem because these images are similar to one another for less experienced radiologist. An experienced radiologist may categorize the image more accurately. If the diagnosis of disease is precise, then the problem can be solved in the early stage or suppressed using medication. For this purpose, deep learning techniques can be used for automatic detection and categorization of the kidney-related problem. This paper will give a clear set of information about dataset selection and pre-processing, and the architecture used in various papers are discussed. Based on the understanding of VGG -16 and AlexNet, the proposed model was used for the classification of lesions in the kidney. The model gives an accuracy of 79%.

This paper presents the development of a hybrid approach as a solution to the multiple Traveling Salesman Problem (mTSP) applied to the route scheduling for self-drive cars. First, we use k-means to generate routes that equality distribute delivery locations among the cars. Then, these routes are set as the initial population for bio-inspired algorithms, such as Genetic Algorithm (GA) and Ant Colony System (ACS), which perform an evolutionary process to find a route which minimizes the overall distance while keeping the balance of individual tours of each car. The experiments were conducted with our route scheduling system in real and virtual environments. We compared our hybrid approaches using k-means in conjunction with GA and ACS against GA, ACS, and Particle Swarm Optimization (PSO) initialized with random population. The results showed that, as the number of cars and target locations increase, the hybrid modeling approaches outperform GA, ACS, and PSO without any pre-processing.

Plant diseases cause significant damage to agriculture, leading to substantial yield losses and posing a major threat to food security. Detection, identification, quantification, and diagnosis of plant diseases are crucial parts of precision agriculture and crop protection. Modernizing agriculture and improving production efficiency are significantly affected by using computer vision technology for crop disease diagnosis. This technology is notable for its non-destructive nature, speed, real-time responsiveness, and precision. Deep learning (DL), a recent breakthrough in computer vision, has become a focal point in agricultural plant protection that can minimize the biases of manually selecting disease spot features. This study reviews the techniques and tools used for automatic disease identification, state-of-the-art DL models, and recent trends in DL-based image analysis. The techniques, performance, benefits, drawbacks, underlying frameworks, and reference datasets of more than 278 research articles were analyzed and subsequently highlighted in accordance with the architecture of computer vision and deep learning models. Key findings include the effectiveness of imaging techniques and sensors like RGB, multispectral, and hyperspectral cameras for early disease detection. Researchers also evaluated various DL architectures, such as convolutional neural networks, vision transformers, generative adversarial networks, vision language models, and foundation models. Moreover, the study connects academic research with practical agricultural applications, providing guidance on the suitability of these models for production environments. This comprehensive review offers valuable insights into the current state and future directions of deep learning in plant disease detection, making it a significant resource for researchers, academicians, and practitioners in precision agriculture.

A promising approach to improve cloud parameterizations within climate models and thus climate projections is to use deep learning in combination with training data from storm‐resolving model (SRM) simulations. The ICOsahedral Non‐hydrostatic (ICON) modeling framework permits simulations ranging from numerical weather prediction to climate projections, making it an ideal target to develop neural network (NN) based parameterizations for sub‐grid scale processes. Within the ICON framework, we train NN based cloud cover parameterizations with coarse‐grained data based on realistic regional and global ICON SRM simulations. We set up three different types of NNs that differ in the degree of vertical locality they assume for diagnosing cloud cover from coarse‐grained atmospheric state variables. The NNs accurately estimate sub‐grid scale cloud cover from coarse‐grained data that has similar geographical characteristics as their training data. Additionally, globally trained NNs can reproduce sub‐grid scale cloud cover of the regional SRM simulation. Using the game‐theory based interpretability library SHapley Additive exPlanations, we identify an overemphasis on specific humidity and cloud ice as the reason why our column‐based NN cannot perfectly generalize from the global to the regional coarse‐grained SRM data. The interpretability tool also helps visualize similarities and differences in feature importance between regionally and globally trained column‐based NNs, and reveals a local relationship between their cloud cover predictions and the thermodynamic environment. Our results show the potential of deep learning to derive accurate yet interpretable cloud cover parameterizations from global SRMs, and suggest that neighborhood‐based models may be a good compromise between accuracy and generalizability. Plain Language Summary Climate models, such as the ICOsahedral Non‐hydrostatic climate model, operate on low‐resolution grids, making it computationally feasible to use them for climate projections. However, physical processes –especially those associated with clouds– that happen on a sub‐grid scale (inside a grid box) cannot be resolved, yet they are critical for the climate. In this study, we train neural networks that return the cloudy fraction of a grid box knowing only low‐resolution grid‐box averaged variables (such as temperature, pressure, etc.) as the climate model sees them. We find that the neural networks can reproduce the sub‐grid scale cloud fraction on data sets similar to the one they were trained on. The networks trained on global data also prove to be applicable on regional data coming from a model simulation with an entirely different setup. Since neural networks are often described as black boxes that are therefore difficult to trust, we peek inside the black box to reveal what input features the neural networks have learned to focus on and in what respect the networks differ. Overall, the neural networks prove to be accurate methods of reproducing sub‐grid scale cloudiness and could improve climate model projections when implemented in a climate model. Key Points Neural networks can accurately learn sub‐grid scale cloud cover from realistic regional and global storm‐resolving simulations Three neural network types account for different degrees of vertical locality and differentiate between cloud volume and cloud area fraction Using a game theory based library we find that the neural networks tend to learn local mappings and are able to explain model errors

Machine learning is increasingly recognized as a promising technology in the biological, biomedical, and behavioral sciences. There can be no argument that this technique is incredibly successful in image recognition with immediate applications in diagnostics including electrophysiology, radiology, or pathology, where we have access to massive amounts of annotated data. However, machine learning often performs poorly in prognosis, especially when dealing with sparse data. This is a field where classical physics-based simulation seems to remain irreplaceable. In this review, we identify areas in the biomedical sciences where machine learning and multiscale modeling can mutually benefit from one another: Machine learning can integrate physics-based knowledge in the form of governing equations, boundary conditions, or constraints to manage ill-posted problems and robustly handle sparse and noisy data; multiscale modeling can integrate machine learning to create surrogate models, identify system dynamics and parameters, analyze sensitivities, and quantify uncertainty to bridge the scales and understand the emergence of function. With a view towards applications in the life sciences, we discuss the state of the art of combining machine learning and multiscale modeling, identify applications and opportunities, raise open questions, and address potential challenges and limitations. We anticipate that it will stimulate discussion within the community of computational mechanics and reach out to other disciplines including mathematics, statistics, computer science, artificial intelligence, biomedicine, systems biology, and precision medicine to join forces towards creating robust and efficient models for biological systems.