artificial neural network

Antimicrobial resistance (AMR) in clinical and environmental Acinetobacter baumannii strains has been recognized as a worldwide challenge for public health. Artificial neural networks (ANNs), an artificial intelligence (AI) algorithm, is a computational model for understanding the complex relationship between input and output data. The ANNs model can support authorities in making proper prescriptions in a significantly shorter time frame, facilitating a more accurate treatment procedure while saving budget and required medical staff.

The present study aimed to investigate whether AI can improve the detection of AMR A. baumannii isolates under experimental conditions.

Clinical and environmental A. baumannii isolates were collected from hospitalized patients and the perimeter of hospitals. The minimum inhibitory concentrations (MICs) of isolates to antibiotics and biocides effective doses were determined using the microdilution broth test according to CLSI-2021 guidelines. The ANNs model was trained using a portion of in vitro datasets (i.e., train set), taking into account different characteristics of clinical/environmental isolates recorded for each isolate in the dataset. Finally, the ANNs model was used to predict the AMR class and biocides dose class of the laboratory dataset (i.e., test set), and results were compared with existing data to determine the accuracy of ANNs model predictions.

On average, 35% of A. baumannii strains were isolated from clinical/environmental samples. The minimum sensitivity level [i.e., R class of ciprofloxacin (CIP5) and ticarcillin (TIC75)] was observed in 70% of clinical A. baumannii isolates, and the most effective dose of BZK and BZT biocides against environmental isolates was 256 µg/mL, while it was 128 µg/mL for CLX. Results showed that the prevalence of A. baumannii isolates resistant to antibiotics and biocides is high not only in hospitals but also in the environment, and the ANNs model can predict the classification of the remaining test dataset with approximately 90% accuracy.

This measure can contribute to the prevention of the overuse and incorrect use of antimicrobial agents to combat rising resistance rates of A. baumannii by focusing on a wider variety of potentially effective parameters on the required dose of antimicrobial agents. This model can be practically used in hospitals as part of treatment protocols, highlighting the cheap and fast diagnosis and prescription.

The thyroid size increases in some diseases of this gland, which is known as goiter or thyromegaly. Thyromegaly is more prevalent in women than men. Thyroid ultrasound is a non-invasive, inexpensive, and safe method for assessing thyroid disorders. Length, width (W), and height measuring of the thyroid lobes and volume is a time-consuming process.

The object of the present study was to offer a rapid artificial neural network (ANN) system for automatic detection of thyromegaly and to investigate the relationship between thyroid dimensions and its volume.

The thyroid dimensions including longitudinal (L) diameter, anterior-posterior or depth (AP) diameter, W diameter, and AP diameter of isthmus were measured among 131 individuals, assigned to thyromegaly and normal groups during the years 2018 and 2019. Proposed ANN system containing of one hidden layer, 3 neurons, and one output for representing the presence of thyromegaly was created by MATALB software.

The implementation of the neural network showed that the use of the measured parameters of the right lobe is able to detect thyromegaly with an accuracy of 94.74, and the isthmus information does not increase the accuracy of the network very much. Also, "width of the right lobe" is the most appropriate parameter and the isthmus thickness was the least important parameter for automatic detection of thyromegaly.

Artificial neural network as a decision-support system can be useful in reducing measurement error while performing ultrasound in detection of thyromegaly and also reducing time of diagnostic processes

Mammography is the most fundamental and widely used method for detecting breast abnormalities. Distinguishing malignant from benign lesions requires extracting relevant information, which can be challenging and timeconsuming for radiologists. Computer-aided diagnosis (CAD) techniques can serve as complementary diagnostic tools, assisting radiologists in the early detection and analysis of abnormalities in mammograms.

This study aimed to propose a CAD system for extracting significant features of abnormalities in breast mammograms using Curvelet transform and fractal analysis, and classifying breast tumors as malignant or benign based on the calculated features. Patients and

In this study, an efficient feature extraction method was applied, utilizing Curvelet transform and fractal analysis, on a dataset comprising 113 abnormal images from the Mammographic Image Analysis Society (MIAS) database, which included 62 benign and 51 malignant cases. The method yielded 575 features, but due to potential irrelevance or redundancy, a multi-objective optimization (MOO) approach based on a genetic algorithm (GA) for an artificial neural network (ANN), named GA-MOO-ANN, was proposed to obtain and focus on an optimal and effective feature set. As a result of this approach, a set of 17 efficient features was extracted. The proposed algorithm was implemented in MATLAB 2014a, and the performance metrics were calculated using 6-fold cross-validation.

The experimental results demonstrated exceptional performance, with an accuracy (Acc) of 98.2%, specificity (Sp) of 100%, sensitivity (Se) of 96.8%, positive predictive value (PPV) of 100%, negative predictive value (NPV) of 96.2%, and an impressive area under the curve (AUC) of 0.98, providing comparable results to other recent methods.

The current findings suggest that the proposed method could be a valuable tool for breast cancer diagnosis, potentially reducing the number of unnecessary breast biopsies. This method may lead to more efficient patient evaluation and earlier detection of breast tumors.

Breast cancer (BC) is the most common cancer in women, and it is important to identify models that can accurately predict mortality in patients with this cancer. The aim of the present study was to use the elastic net regression and artificial neural network (ANN) models in diagnosing and predicting factors affecting BC mortality.

A cross-sectional study.

The data of 2,836 people with BC during 2014-2018 were analyzed in this study. Information was registered in the cancer registration system of Kerman University of Medical Sciences. Death status was considered the dependent variable, while age, morphology, tumor differentiation, residence status, and residence place were regarded as independent variables. Sensitivity, specificity, accuracy, area under the receiver operating characteristic curve (AUC), precision, and F1-score were used to compare the models.

Based on the test set, the elastic net regression determined factors affecting BC mortality (with sensitivity of 0.631, specificity of 0.814, AUC of 0.629, accuracy of 0.792, precision of 0.318, and F1-score of 0.42) and ANN did so (with sensitivity of 0.66, specificity of 0.748, AUC of 0.704, accuracy of 0.738, precision of 0.265, and F1-score of 0.37).

The sensitivity and AUC of the ANN model were higher than those of the elastic net regression, but the specificity, accuracy, precision, and F1-score of the elastic net were higher than those of the ANN. According to the purpose of the study, two models can be used simultaneously. Based on the results of models, morphology, tumor differentiation, and age had a greater effect on death.

One of the main reasons for neonatal deaths is preterm delivery, and infants who have survived preterm birth (PB) are at risk of significant health complications. However, an effective method for reliable and accurate prediction of preterm labor has yet to be proposed.
This study proposes an artificial neural network (ANN)-based approach for early prediction of PB, and consequently can hint physicians to start the treatment earlier, reducing the chance of morbidity and mortality in the infant.
This historical cohort study proposes a feed-forward ANN with 7 hidden neurons to predict PB. Thirteen risk factors of PB were collected from 300 pregnant women (150 with preterm delivery and 150 normal) as the ANN inputs from 2018 to 2019. From each group, 70%, 15%, and 15% of the subjects were randomly selected for training, validation, and testing of the model, respectively.
The ANN achieved an accuracy of 79.03% for the classification of the subjects into two classes normal and PB. Moreover, a sensitivity of 73.45% and specificity of 84.62% were obtained. The advantage of this approach is that the risk factors used for prediction did not require any lab test and were collected in a questionnaire. 
The efficacy of the proposed approach for the early identification of pregnant women, who are at high risk of preterm delivery, leads to necessary care and clinical interventions, applied during the pregnancy.

Modeling energy demand in different energy consuming sectors is a crucial measure for effective management of the energy sector and appropriate policies to increase productivity. The rising importance of energy resources in economic development is evident. Sustainable energy use is crucial for environmental protection and social progress. Understanding the factors affecting energy consumption is essential for effective energy management. Therefore, the purpose of the current study is to investigate the impact of environmental factors on household electricity consumption in Yazd city.

In the present research, various environmental factors affecting electricity consumption, including air pollution, air temperature in homes, ground surface temperature, and green space were investigated. The effects of these factors on electricity consumption of subscribers were investigated with ANN and  apriori methods.

Among the environmental factors, the distance to the regional park, the area of the park, and the amount of vegetation at a distance of 300m have the greatest impact, respectively, and the average summer air temperature, the amount of vegetation at a radius of 500 m, the distance from the local park, and the average summer NDVI have had the smallest effect. Unlike neural network methods, apriori presents relationships between parameters affecting electricity consumption transparently in the form of rules.

It's used to identify the most frequently occurring elements and meaningful associations in a dataset. Greenspace can be a mitigation strateegy for reduction of energy consumption.

Due to the growing number of disabilities in elderly, Attention to this period of life is essential to be considered. Few studies focused on the physical, mental, disabilities, and disorders affecting the quality of life in elderly people. SA1 is related to various factors influencing the elderly’s life. So, the objective of the current study is to build an intelligent system for SA prediction through ANN2 algorithms to investigate better all factors affecting the elderly life and promote them.

This study was performed on 1156 SA and non‑SA cases. We applied statistical feature reduction method to obtain the best factors predicting the SA. Two models of ANNs with 5, 10, 15, and 20 neurons in hidden layers were used for model construction. Finally, the best ANN configuration was obtained for predicting the SA using sensitivity, specificity, accuracy, and cross‑entropy loss function.

The study showed that 25 factors correlated with SA at the statistical level of P < 0.05. Assessing all ANN structures resulted in FF‑BP3 algorithm having the configuration of 25‑15‑1 with accuracy‑train of 0.92, accuracy‑test of 0.86, and accuracy‑validation of 0.87 gaining the best performance over other ANN algorithms.

Developing the CDSS for predicting SA has crucial role to effectively inform geriatrics and health care policymakers decision making.

Gaining insight into the underlying molecular mechanisms of Alzheimer’s disease (AD) is crucial.

This study aimed to employ a systems biology approach to identify new non-invasive diagnostic biomarkers for AD.

Gene expression data series GSE122063 and microRNA (miRNA) expression data series GSE90828 were obtained from the Gene Expression Omnibus database. The Limma package under R software was used to assess differentially expressed miRNAs and differentially expressed genes (DEGs). Afterward the protein-protein interaction (PPI) network was constructed by the STRING software and evaluated with Cytoscape software. The multilayer perceptron neural network (MLP-NN), a widely used artificial neural network (ANN), was employed to classify two groups.

A total of 1388 DEGs were identified in AD patients compared to the control group, and 11 differentially expressed miRNAs were found in patients with mild cognitive impairment (MCI) in comparison to the control group. The results revealed that EGFR, identified as a hub gene, was targeted by miR-15b-3p and let-7a-5p, while TLR4, another hub gene, was targeted by miR-15b-3p. The MLP-NN constructed using both hsa-let-7a-5p and hsa-miR-15b-3p achieved a sensitivity of 0.857 and an area under the curve of 0.917 in detecting Alzheimer’s patients.

Our findings suggest that miR-15b-3p, by targeting EGFR and TLR4, and let-7a-5p, by targeting EGFR, may play a significant role in AD. Additionally, the constructed ANN utilizing the expression levels of plasma miR-15b-3p and let-7a-5p could serve as a potential non-invasive diagnostic tool with high sensitivity for AD detection.

The development of data mining techniques and the adaptive neuro-fuzzy inference system (ANFIS) in the last few decades has made it possible to achieve accurate predictions in medical fields.

The present study aimed to use the ANFIS model, artificial neural network (ANN), and logistic regression to predict thyroid patients.

This study aimed to predict thyroid disease using the UCI database, ANFIS and ANN models, and logistic regression. We only used four of its features as the input of the model and considered thyroid as a binary response (occurrence=1, non-occurrence=0) as the output of the model. Finally, three models were compared based on the accuracy and the area under the curve (AUC).

In this study, out of the extensive UCI database, which includes 3,772 samples and over 20 features, only five specific features were utilized. Data include 1,144 males and 2,485 females. The results of multiple logistic regression analysis demonstrated that free T4 index (FTI) and thyroid stimulating hormone (TSH) had a significant effect on thyroid. The ANFIS model had a higher accuracy (99%) compared to ANN (96%) and the logistic regression model (94%) in the prediction of thyroid.

As evidenced by the obtained results, the forecasting performance of ANFIS is more efficient than other models. Moreover, the use of combined methods, such as ANFIS, to diagnose and predict diseases increases the accuracy of the model. Therefore, the results of this study can be used for screening programs to identify people at risk of thyroid disease.

In this study, we analyze the optimal intervention strategies that lead to reducing the effects of the COVID-19 pandemic by artificial neural networks (ANNs). Our aim is to investigate the effects of optimal control strategies, such as the implementation of government intervention, testing, and vaccination policies during outbreaks. We utilized a controlled SIDAREV model to study the progression of the COVID-19 pandemic. Using Pontryagin's minimum principle (PMP) for the SIDAREV model, we defined an unconstrained minimization problem. Applying the Hamiltonian conditions, we approximated the obtained ordinary differential equations (ODE) using ANNs. We utilized the multilayer perceptron (MLP) to obtain the approximate solution of the states and co-states functions. We observed the effects of optimal control strategies, and to show the efficiency of the proposed method, we compared it with the Runge-Kutta method through some examples. Using a mathematical model that simulates the behavior of the Covid-19 disease, we can examine the effects of controllers such as government interventions, tests and vaccinations with the neural network method. The results show that this method is useful in solving the problem of optimal control of infectious diseases.

 The purpose of this study was to develop a robust and externally predictive in silico QSAR-neural network model for predicting plasma protein binding of drugs. This model aims to enhance drug discovery processes by reducing the need for chemical synthesis and extensive laboratory testing.

 A dataset of 277 drugs was used to develop the QSAR-neural network model. The model was constructed using a Filter method to select 55 molecular descriptors. The validation set’s external accuracy was assessed through the predictive squared correlation coefficient Q2 and the root mean squared error (RMSE).

The developed QSAR-neural network model demonstrated robustness and good applicability domain. The external accuracy of the validation set was high, with a predictive squared correlation coefficient Q2 of 0.966 and a root mean squared error (RMSE) of 0.063. Comparatively, this model outperformed previously published models in the literature.

 The study successfully developed an advanced QSAR-neural network model capable of predicting plasma protein binding in human plasma for a diverse set of 277 drugs. This model’s accuracy and robustness make it a valuable tool in drug discovery, potentially reducing the need for resource-intensive chemical synthesis and laboratory testing.

This paper presents the development of an intelligent system for managing medical oxygen consumption using oximetry and barometry in the hospitals of Mashhad University of Medical Sciences, utilizing machine learning methods. The system integrates various sensors and machine learning algorithms to enable real-time monitoring and control of the oxygen supply chain. 

The proposed approach utilizes multiple sensors to measure the purity and pressure of medical oxygen, and this data is collected and processed using machine learning algorithms. The system uses a decision tree model to classify the purity and pressure readings and identify deviations from the specified parameters. The system also utilizes an artificial neural network model to predict future oxygen consumption levels, enabling proactive supply chain management. The system consists of two main components: the hardware component and the software component. The software component includes machine learning algorithms for data processing and system management. 

The proposed system has been tested in several hospitals affiliated with Mashhad University of Medical Sciences, and the results show that it can effectively monitor and manage medical oxygen consumption with high accuracy and reliability. The machine learning algorithms used in the system have the potential to improve patient safety by identifying potential issues in the oxygen supply chain before they become critical.  

In conclusion, this paper presents an innovative and intelligent system that utilizes machine learning methods to enhance the management of medical oxygen consumption in hospitals significantly.

 Vulvovaginal candidiasis (VVC) is a common fungal infection caused by Candida species in the female genital tract.

 This study attempts to predict predisposition to VVC related to risk factors and clinical symptoms among vaginitis cases using the artificial neural network (ANN) model.

 This cross-sectional study was performed on 250 women referred to gynecology clinics in Birjand, Iran. A questionnaire was used to record participants' demographic information. Swabs were used for wet mounts and culture. Candida species were identified by morphological and physiological methods. The performance of the optimal neural network model was assessed by the sensitivity, specificity, and accuracy area under the ROC curve (AUC). Descriptive statistics were used for the statistical description of data, and chi-square test, t-test, and ANN analysis using SPSS application tools (Statistical Product and Service Solutions) version 22 software at 0.05 significant level.

 The prevalence of vulvovaginal candidiasis was 41.0%, and Candida albicans was the most frequently identified species (55.9%). The descriptive statistics (chi-square test and t-test) revealed no significant difference between the frequencies of Candida infection with demographic factors and clinical presentations. However, factors such as abortion history, number of sexual intercourse, dyspareunia, education, natural vaginal delivery (NVD), and lower abdominal pain included in our ANN model had significant differences (P < 0.05).

 The result of the ANN model revealed that using demographic factors and clinical symptoms can predict VVC infection. Therefore, this model can identify the effect of the clinical presentations and symptoms of infection.

Malaria remains a significant global health problem, with a high incidence of cases and a substantial number of deaths yearly. Early identification and accurate diagnosis play a crucial role in effective malaria treatment. However, underdiagnosis presents a significant challenge in reducing mortality rates, and traditional laboratory diagnosis methods have limitations in terms of time consumption and error susceptibility. To overcome these challenges, researchers have increasingly utilized Machine Learning techniques, specifically neural networks, which provide faster, cost-effective, and highly accurate diagnostic capabilities.

This study aimed to compare the performance of a traditional neural network (NN) with a convolutional neural network (CNN) in the diagnosis and classification of different types of malaria using blood smear images. We curated a comprehensive malaria dataset comprising 1,920 images obtained from 84 patients suspected of having various malaria strains. The dataset consisted of 624 images of Falciparum, 548 images of Vivax, 588 images of Ovale, and 160 images from suspected healthy individuals, obtained from local hospitals in Iran. To ensure precise analysis, we developed a unique segmentation model that effectively eliminated therapeutically beneficial cells from the image context, enabling accurate analysis using artificial intelligence algorithms.

The evaluation of the traditional NN and the proposed 6-layer CNN model for image classification yielded average accuracies of 95.11% and 99.59%, respectively. These results demonstrate that the CNN, as a primary algorithm of deep neural networks (DNN), outperforms the traditional NN in analyzing different classes of malaria images. The CNN model demonstrated superior diagnostic performance, delivering enhanced accuracy and reliability in the classifying of malaria cases.

This research underscores the potential of ML technologies, specifically CNNs, in improving malaria diagnosis and classification. By leveraging advanced image analysis techniques, including the developed segmentation model, CNN showcased remarkable proficiency in accurately identifying and classifying various malaria parasites from blood smear images. The adoption of machine learning-based approaches holds promise for more effective management and treatment of malaria, addressing the challenges of underdiagnosis and improving patient outcomes.