Showing 9 results for Neural Network
Mostafa Langarizadeh, Esmat Khajehpour, Rahele Salari, Hassan Khajehpour,
Volume 10, Issue 5 (1-2017)
Abstract
Background and Aim: Bacterial meningitis detection is a complicated problem because of having several components in order to be diagnosed and distinguished from other types of meningitis. Fuzzy logic and neural network, frequently used in expert systems, are able to distinguish such diseases. The purpose of this paper is to compare Fuzzy logic and artificial neural networks for distinguishing bacterial meningitis from other types of meningitis.
Materials and Methods: In this study to detect and distinguish bacterial meningitis from other types of meningitis, in the first step 6 attributes were selected by infectious disease specialists. In the second step, systems were designed by Matlab software. The systems were evaluated by 26 records of meningitis patients, and results were analyzed by SPSS software.
Results: The evaluation showed that the accuracy, specificity and sensitivity of fuzzy method were 88%, 92% and 100% respectively and those of neural network methods were 92%, 94% and 88% respectively. The Kappa test result in fuzzy and neural network methods were 0.83 (p<0.001) and 0.83 (p<0.001). The areas under the ROC curves were 0.94 and 0.91 respectively.
Conclusion: The sensitivity, the Kappa test results and the areas under the ROC curve of the fuzzy logic method were better than neural network method. However the fuzzy logic method is more reliable to distinguish bacterial meningitis from other type of Meningitis, the evaluation result were obtained from 26 records of meningitis patient which were hospitalized in the same center leads to the study be still open.
Sajad Mazaheri , Maryam Ashoori, Zeynab Bechari,
Volume 11, Issue 3 (9-2017)
Abstract
Background and Aim: Nowadays heart disease is very common and is a major cause of mortality. Proper and early diagnosis of this disease is very important. Diagnostic methods and treatments of the disease are so expensive and have many side effects. Therefore, researchers are looking for cheaper ways to diagnose it with high precision. This study aimed to identify a model for the treatment of heart disease.
Materials and Methods: In this descriptive cross-sectional study, the sampling method was census. The sample consisted of data from Khatam and Ali Ibn Abi Talib Hospitals in Zahedan. The data were developed as an Excel file, and Clementine12.0 software was used for data analysis. In the present study, C5.0, C & R Tree, CHAID, and QUEST algorithms and artificial neural network were carried out on the collected data.
Results: The accuracy of 76.04 by C & R algorithm indicates the better performance of Decision Tree Algorithms than that of the Neural Network.
Conclusion: This study aimed to provide a model for the prediction of a suitable heart disease treatment to reduce treatment costs and provide better quality of services for physicians. Due to considerable implementation risks of invasive diagnostic procedures such as angiography and also obtaining successful experiences of data analysis in medicine, this study has presented a model based on data analysis techniques. The improvable point of this model is the provision of a decision support system to help physicians to increase the accuracy of diagnosis in the treatment of diseases.
Azam Orooji , Mostafa Langarizadeh , Maryam Aghazadeh, Mehran Kamkarhaghighi, Marjan Ghazisaiedi , Fateme Moghbeli,
Volume 12, Issue 4 (11-2018)
Abstract
Background and Aim: Artificial intelligence is a branch of computer science that has the ability of analyzing complex medical data. Using artificial intelligence is common in diagnosing, treating and taking care of patients. Warfarin is one of the most commonly prescribed oral anticoagulants. Determining the exact dose of warfarin needed for patients is one of the major challenges in the health system, which has attracted the attention of researchers. The purpose of this study was to determine the exact dose of warfarin needed for patients with artificial heart valves using artificial neural networks (ANN).
Materials and Methods: A total of 9 multi-layer perceptron ANNs with different structures were constructed and evaluated based on a dataset including 846 patients who had referred to the PT clinic in Tehran Heart Center in the second half of the year 2013. Finally, the best structure of ANN for warfarin dose was investigated. All simulations including data preprocessing and neural network designing were done in MATLAB environment.
Results: The effectiveness of ANNs was evaluated in terms of classification performance using 10-fold cross-validation procedure and the results showed that the best model was a network that had 7 neurons in its hidden layer with an average absolute error of 0.1, turbulence rate of 0.33, and regression of 0.87.
Conclusion: The achieved results reveal that ANNs are able to predict warfarin dose in Iranian patients with an artificial heart valve. Although no system can be guaranteed to achieve 100% accuracy, they can be effective in reducing medical errors.
Azita Yazdani, Ali Asghar Safaei, Reza Safdari, Maryam Zahmatkeshan,
Volume 13, Issue 3 (9-2019)
Abstract
Background and Aim: Breast cancer is the most common type of cancer and the main cause of death from cancer in women worldwide. Technologies such as data mining, have enabled experts in this area to improve decision making in the early diagnosis of the disease. Therefore, the purpose of this research is to develop an automatic diagnostic model for breast cancer by employing data mining methods and selecting the model with the highest accuracy of diagnosis.
Materials and Methods: In this study, 654 available patient records of Motahari breast cancer Clinic in Shiraz" were used as the sample. The number of records was reduced to 621 after the pre-processing operation. These samples had 22 features that ultimately used ten were used as effective features in the design of the model. Three types of Decision tree, Naive Bayes and Artificial neural network were used for diagnosis of breast cancer and 10-fold cross-validation method for constructing and evaluating the model on the collected data set.
Results: The results of the three techniques mentioned all three models showed promising results in detecting breast cancer. Finally, the artificial neural network accounted for the highest accuracy of 94/49%(sensitivity 96/19%, specificity 86/36%) in the diagnosis of breast cancer.
Conclusion: Based on the results of the decision tree, the risk factors such as age, weight, Age of menstruation, menopause, OCP of records duration, and the age of the first pregnancy were among the factors affecting the incidence of breast cancer in women.
Mohsen Rezaei, Nazanin Zahra Jafari, Hossein Ghaffarian, Masoud Khosravi Farmad3, Iman Zabbah, Parvaneh Dehghan,
Volume 13, Issue 5 (1-2020)
Abstract
Background and Aim: Timely diagnosis and treatment of abnormal thyroid function can reduce the mortality associated with this disease. However, lack of timely diagnosis will have irreversible complications for the patient. Using data mining techniques, the aim of this study is to determine the status of the thyroid gland in terms of normality, hyperthyroidism or hypothyroidism.
Materials and Methods: Using supervised and unsupervised methods after data preprocessing, predictive modeling was performed to classify thyroid disease. This is an analytical study and its dataset contains 215 independent records based on 5 continuous features retrieved from the UCI machine learning data reference.
Results: In supervised method, multilayer perception(MLP), learning vector quantization(LVQ), and fuzzy neural network(FNN) were used; and in unsupervised method, fuzzy clustering was employed. Besides, these precision figures(0.055, 0.274, 0.012 and 1.031) were obtained by root mean square error(RMSE) method, respectively.
Conclusion: Reducing the diagnosis error of thyroid disease was one of the goals of researchers. Using data mining techniques can help reduce this error. In this study, thyroid disease was diagnosed by different pattern recognition methods. The results show that the fuzzy neural network(FNN) has the least error rate and the highest accuracy.
Mona Sarhadi, Mohammad Amin Shayegan,
Volume 15, Issue 1 (3-2021)
Abstract
Background and Aim: For effective treatment of Alzheimer's disease (AD), it is important to accurately diagnosis of AD and its earlier stage, Mild Cognitive Impairment (MCI). One of the most important approaches of early detection of AD is to measure atrophy, which uses various kinds of brain scans, such as MRI. The main objective of the current research was to provide a computerized diagnostic system for early diagnosis of AD, using leraning machine algorithms, to help physicians. The proposed system diagnoses AD by examining the hippocampal atrophy of brain MRI images and increases the accuracy of the diagnosis.
Materials and Methods: In this study, hippocampus was segmented from the other parts of the brain by using active contour and convolutional neural network and then, three groups of “Normal Controls: NC”, AD and MCI were classified by using the SVM classifier.
Results: The proposed method has succeeded in classifying AD against NC with 98.77%, 98.74% and 97.96% in average for accuracy, sensitivity and specificity, respectively. Also in classification of MCI against NC, the mean accuracy, sensitivity and specificity of 96.14%, 96.23% and 88.21% were achieved, respectively. Compared with the nearest rival method, the proposed method showed improvement accuracy and sensitivity of classification AD from NC with 1.64% and 2.81% respectively. Also, in classification of MCI from NC it showed improvement for accuracy with 8.9% and sensitivity with 2.16%, respectively. Improving in results were due to the use of a modified ACM segmentation algorithm, the use of a combination of features extracted from hippocampal images and features already created by the ImageNet network, the removal of inappropriate features from the feature vector, and the use of deep Inception v3 network.
Concolusion: Based on the results, the combination of polygon surrounding the hippocampus features and deep network features can be useful for detection of AD and MCI.
Marsa Gholamzadeh, Seyed Mohammad Ayyoubzadeh, Hoda Zahedi, Sharareh Rostam Niakan Kalhori,
Volume 15, Issue 3 (8-2021)
Abstract
Background and Aim: Due to the important role of radiological images for identifying patients with COVID-19, creating a model based on deep learning methods was the main objective of this study.
Materials and Methods: 15,153 available chest images of normal, COVID-19, and pneumonia individuals which were in the Kaggle data repository was used as dataset of this research. Data preprocessing including normalizing images, integrating images and labeling into three categories, train, test and validation was performed. By Python language in the fastAI library based on convolution technique (CNN) and four architectures (ResNet, VGG MobileNet, AlexNet), nine models through transitional learning method were trained to recognize patients from healthy persons. Finally, the performance of these models was evaluated with indicators such as accuracy, sensitivity and specificity, and F-Measure.
Results: Of the nine generated models, the ResNet101 model has the highest ability to distinguish COVID-19 cases from other cases with 95.29% sensitivity. Other applied models showed more than 96% accuracy in correctly diagnosis of various cases in test phase. Finally, the ResNet101 model was able to demonstrate 98.4% accuracy in distinguishing between healthy and infected cases.
Conclusion: The obtained accuracy showed the accurate performance of developed model in detecting COVID-19 cases. Therefore, by implementing an application based on the developed model, physicians can be helped in accurate and early diagnosis of cases. an application based on the developed model, physicians can be helped in accurate and early diagnosis of infected cases.
Setareh Talayeh, Farzad Firouzi Jahantigh, Fatemeh Bahman,
Volume 17, Issue 5 (12-2023)
Abstract
Background and Aim: The tourism industry plays a very important role in the economic cycle of society. Medical tourism, as one of the types of tourism industries, has a direct result in globalizing health care. Therefore, by strengthening the supply chain in this area, a very high added value can be achieved. For this reason, the present study provides a conceptual framework for predicting the demand for medical tourism supply chain by determining the relationship between medical tourism demand and economic, medical, and welfare-service components of Zahedan city.
Materials and Methods: The present study is a descriptive-analytical and applied research. Data were collected using a questionnaire and field and library methods. The statistical population of interest was specialist doctors in Zahedan city, and 97 people were selected using simple random sampling with Morgan’s table. The validity of the questionnaire was confirmed by experts and its reliability was obtained using Cronbach’s alpha coefficient with SPSS software more than 0.7. Data analysis was performed using the tangent sigmoid neural network algorithm, linear regression criteria, and mean square error. For this purpose, SPSS software was used to examine the correlation between the data, and MATLAB software was used to design the neural network.
Results: There was anerrore in The basis for the optimality of the answers, linear regression criteria and mean square error. The results showed that the values related to regression, education, and health were more than 0.8 and were 0.9033, 0.8818, and 0.9985, respectively. The highest priorities of the respondents related to medical equipment, education, and health were 0.5657, 0.5558, and 0.20726, respectively.
Conclusion: According to the results obtained from the proposed model, the neural network has a high accuracy in predicting the demand for medical tourism supply chain in terms of education, health, and welfare. It is also predicted that the demand for medical tourism has been constant during the one-year period of research and it is expected that medical tourism in Zahedan city will decrease in future. Therefore, it is recommended that officials pay attention to the development and improvement of medical tourism to promote it.
Seyed Mohammad Sadegh Dashti, Amid Khatibi Bardsiri, Mehdi Jafari Shahbazzadeh,
Volume 18, Issue 1 (3-2024)
Abstract
Background and Aim: Medical reports and electronic health records are critically important for diagnosis, treatment, patient protection, and medical research. Correcting spelling errors in medical texts is essential to ensure accurate interpretation of information. This research was conducted to automatically correct spelling mistakes in Persian medical texts using neural networks.
Material and Methods: In this study, which was conducted in 2023, a computational model based on artificial intelligence neural networks and dual embedding techniques was developed using Python in a Windows environment. The dual embedding model was fine-tuned for correcting spelling errors in Persian sonography texts. The proposed model employs various techniques for automatic error detection, including dictionary lookup approach and contextual similarity coefficients. Furthermore, features specific to text processing, such as Edit-Distance, along with similarity coefficients, were utilized to automatically select the most appropriate substitute for a misspelled word. The training and testing data for the current model were sourced from a collection of sonography texts from the Imam Khomeini Hospital’s sonography clinic in Tehran.
Results: The proposed model which is based on artificial neural networks, leverages a novel dualembedding architecture to select the best candidate words for correcting both non-word and real-word errors. According to the evaluation results on Persian sonography text, the proposed model achieved an F-Measure accuracy of 90.5% in detecting real-word errors. Furthermore, it demonstrated an impressive 90% accuracy in automatically correcting these real-word errors. Additionally, the model exhibited a strong performance, achieving 90.8% accuracy in correcting non-word errors.
Conclusion: Based on the evaluation results, the proposed method is robust against various changes in word forms and can effectively manage a wide range of morphological and semantic errors, including replacements, transpositions, insertions, and deletions in medical texts. The integration of EditDistance with textual similarity coefficients extracted from the dual embedding model significantly enhanced the accuracy of spelling corrections in Persian sonography texts, ensuring greater validity of such documents. The authors believe that the proposed model represents a significant advancement in the detection and correction of spelling errors in Persian sonography texts.
